JOURNAL OF PSYCHOPHARMACOLOGY

SUPPLEMENT TO VOLUME 30, ISSUE 8, AUGUST 2016

These papers were presented at the Summer Meeting of the BRITISH ASSOCIATION FOR PSYCHOPHARMACOLOGY

17 – 20 July, Brighton, UK

Indemnity The scientific material presented at this meeting reflects the opinions of the contributing authors and speakers. The British Association for Psychopharmacology accepts no responsibility for the contents of the verbal or any published proceedings of this meeting.

All contributors completed a Declaration of Interests form when submitting their abstract

BAP Office 36 Cambridge Place Hills Road Cambridge CB2 1NS

www.bap.org.uk Aii CONTENTS

Abstract Book 2016 Abstracts begin on page:

SYMPOSIUM 1 New paradigms and treatment approaches in (S01–S04) A1

SYMPOSIUM 2 New perspectives on functions of 5-HT sub-systems (S05–S08) A3

SYMPOSIUM 3 Towards a gender specific psychopharmacology: The impact of , gonadal steroids and oxytocin (S09–S12) A5

SYMPOSIUM 4

Bipolar disorder: Progress on many fronts (S13–S16) A7

SYMPOSIUM 5

Ketamine and addiction: Promising treatment or probable cause? (S17–S20) A9

SYMPOSIUM 6

Neuronal stem cells in mental health: Will neurogenesis and iPS cells give us the antidepressants of the future? (S21–S24) A11

SYMPOSIUM 7

Anhedonia, apathy, amotivation, anergia? Disrupted reward processing as a trans-diagnostic construct in mental illness (S25–S28) A13

SYMPOSIUM 8

Dysfunctional neuro-immune system interactions in psychiatric disorders and their relevance for novel treatment strategies (S29–S32) A15

SYMPOSIUM 9

Innovation in the treatment of obesity and binge eating (S33–S36) A18 CONTENTS Aiii

POSTERS

Group A: Schizophrenia/Psychosis (A01–A35) A20

Group B: Substance Use (B01–B31) A43

Group C: Affective Disorders (C01–C52) A64

Group D: Anxiety (D01–D11) A99

Group E: Developmental Disorders and Neurodegeneration (E01–E09) A106

Group F: Cognition (F01–F07) A111

Group G: Eating Disorders (G01–G04) A116

Group H: Sleep/Circadian Rhythms (H01–H04) A119

GUEST LECTURE Parsing the serotonin system into distinct functional units (GL1) A122

POSTDOCTORAL SYMPOSIUM

Translational models and biomarkers for neuropsychiatric disorders (PD01–PD04) A122

SHORT ORAL PRESENTATIONS

Short Orals 1: Substance Use: Neural and behavioural effects See abstracts B09, B11, B16, B24

Short Orals 2: Analysing Anhedonia: From mechanism to treatment See abstracts C12, C15, C30, C46

Short Orals 3: Neurochemical mechanisms of psychosis See abstracts A02, A03, A17, A21

2016 PSYCHOPHARMACOLOGY AWARDS (PW1–PW4) A125

ABSTRACTS A1

S01 PET IMAGING STUDIES OF MICROGLIA, A MARKER FOR NEUROINFLAMMATION IN SCHIZOPHRENIA PATIENTS: POTENTIAL AVENUES FOR NEW TREATMENT APPROACHES Talbot PS, Wolfson Molecular Imaging Centre, Univ of Manchester, 27 Palatine Road, Withington, Manchester M20 3LJ [email protected] Introduction: Positron emission tomography (PET) imaging of the 18kDa translocator protein (TSPO) has been used to investigate whether microglial activation, an indication of neuroinflammation, is evident in the brain of adults with schizophrenia. To date the results have been mixed, with early studies finding evidence for microglial activation and the majority of more recent studies refuting this. The discrepant findings may reflect differences in methodology, radiotracers, and aspects of the patient populations including severity and duration of illness. However, importantly, all of the schizophrenia patients in these previous PET studies were taking medication at the time of scanning, introducing a potential confound due to modulatory effects of antipsychotic medication on microglial activity. We have therefore compared TSPO availability between antipsychotic-free patients with schizophrenia, medicated patients, and matched healthy controls. Methods: we used [11C](R)-PK11195 PET to compare TSPO availability in a predominantly antipsychotic-naïve group of moderate-to-severely symptomatic unmedicated patients (n=8), similarly symptomatic medicated patients with schizophrenia taking risperidone or paliperidone by regular intramuscular injection (n=8), and healthy comparison subjects (n=16). Binding potential (BPND) was calculated using a grey matter cerebellum reference input function and the simplified reference tissue model. Results: we found no evidence for increased TSPO availability in antipsychotic- free patients compared to healthy controls (mean difference 4%, p=0.981). However, TSPO availability was significantly elevated in medicated patients (mean increase 88%, p=0.032) across dorsolateral prefrontal, ventrolateral prefrontal, orbitofrontal, anterior cingulate and parietal cortical regions. In the patients with schizophrenia, TSPO availability was also strongly correlated with negative symptoms measured using the Positive and Negative Syndrome Scale across all the brain regions investigated (r=0.651–0.741). Conclusions: we conclude that the pathophysiology of schizophrenia is not associated with microglial activation in the 2–6 year period following diagnosis, but that antipsychotic treatment with risperidone or paliperidone may induce microglial activation. It remains to be determined whether this activation represents a pro- or anti-inflammatory state, its association with primary negative symptoms and treatment resistance, and whether there are significant differences between . Implications for novel treatment strategies will be discussed.

S02 EVIDENCE THAT PARVALBUMIN PATHOLOGY IN SCHIZOPHRENIA CAN BE RESCUED BY NOVEL DRUG TARGETS Harte MK, Manchester Pharmacy School, Univ of Manchester, Stopford Building, Oxford Road, Manchester M13 9PL [email protected] Cognitive deficits remain an unmet clinical need in a number of neuropsychiatric and neurodegenerative disorders. With an improved understanding of the pathology related to these deficits comes the hope of novel treatment strategies. Current evidence from both clinical and preclinical studies suggests a strong involvement of GABAergic interneurons (particularly the parvalbumin (PV) subset) in regulating cognitive function in vivo. Alongside this are numerous reports of dysfunction of these GABAergic neurons in schizophrenia (Lewis et al., Trends Neurosci. 2012; 35(1): 57–67), with a number of relevant preclinical models recapitulating this deficit (e.g. sub-chronic phencyclidine model (Neill et al., 2010; 128(3):419–32)). This emerging understanding of the pathophysiology of schizophrenia offers us the potential to identify novel drug targets through which it may be possible to treat the underlying neural dysfunction and potentially alleviate the cognitive deficits. One such target is the Kv3.1 voltage-gated potassium channel. These channels are expressed on PV interneurons, where they confer the ability of the neurons to fire rapidly and accurately, allowing synchronisation of cortical circuits. A recent post- mortem study found reduction in Kv3.1 protein in un-medicated schizophrenia patients (Yanagi et al. Mol Psychiatry 2014; 19: 573–579). Modulation of this channel may therefore provide a novel target A2 ABSTRACTS for restoration of cognitive function in schizophrenia patients. The presentation will cover data from behavioural, immunohistochemical, electrophysiological and imaging studies in validated preclinical models aimed at investigating the efficacy of this approach for treating cognitive deficits in schizophrenia. Briefly, we demonstrate that treatment (acute and/or chronic) with: - AUT00206, a novel Kv3.1 modulator (Autifony Therapeutics Limited) significantly attenuated the sub-chronic phencyclidine (scPCP) induced social behaviour deficits. - AUT00206 significantly attenuated scPCP induced deficits in different cognitive domains. - AUT00206 significantly enhanced the power of fast network oscillations in the prelimbic cortex from scPCP treated rats. - AUT00206 significantly reversed the scPCP induced PV deficit, an effect dependent on the presence of AUT00206. - AUT00206 significantly reduced induced BOLD signal changes in cortical and subcortical regions of the rat brain Taken together these results suggest that the modulation of Kv3 channels on PV neurons by AUT00206 could be an important novel approach for improving cognitive deficits and function in schizophrenia patients. Financial Sponsorship: These studies were funded by Innovate UK.

S03 UTILITY OF COGNITIVE AND BOLD FMRI BIOMARKERS IN INDIVIDUALS WITH HIGH SCHIZOTYPY TO ASSESS NEW THERAPIES FOR SCHIZOPHRENIA Dourish CT, P1vital, Manor House, Howbery Park, Wallingford, Oxfordshire OX10 8BA [email protected] Schizophrenia is a debilitating, life-long, psychiatric disorder affecting approximately 1% of the world’s population. Antipsychotic drug therapies do not treat all the symptoms of the illness (including positive, negative and cognitive symptoms) and cause significant side-effects including dyskinesia, obesity and diabetes. Therefore, there is a significant unmet need to develop new treatments for schizophrenia that have improved efficacy and fewer side-effects. However, few new compounds have emerged and often promising pre-clinical data have not translated to efficacy in clinical trials. Use of experimental medicine efficacy biomarkers offers a potential way forward for antipsychotic drug development and this presentation reviews a series of recent schizophrenia efficacy biomarker validation studies. Two potential models were examined, an “intermediate phenotype” model and a drug challenge model. The intermediate phenotype was individuals categorised as high schizotypes using a Schizotypal Personality Questionnaire (SPQ) and having personality traits similar to those of patients with schizophrenia. The drug challenge was the NMDA antagonist ketamine which has been reported to induce both positive and negative symptoms of schizophrenia and to disrupt cognitive function. In the schizotypy study, high schizotypes exhibited impaired working memory performance relative to controls in tasks that require swift information processing but had normal executive function in paradigms with longer encoding periods. The effects of the cognition enhancer nicotine were enhanced in high schizotypes on eye movement and N back tasks. Amisulpride improved performance in high schizotypes and impaired performance in controls and this was significant on the N back and eye movement tasks. In a spatial working memory task high schizotypes and controls performed equally well but there were significant differences in their brain fMRI responses. The results suggest that high schizotypes have an inefficient encoding strategy and increased activation of frontal-limbic regions during memory retrieval may be used as a compensatory mechanism to enable comparable task performance to be attained. In the drug challenge study we assessed performance of healthy volunteers infused with ketamine, pre-treated with placebo or risperidone, on the same set of biomarkers used in the schizotypy study. Ketamine did not affect performance on most of the biomarker tasks whereas risperidone generally impaired performance and failed to attenuate the deficits induced by ketamine. An exception was the eye movement task in which ketamine caused some occulomotor performance deficits similar to those seen in schizophrenia. The results suggest that the high schizotypy model shows more promise than the ketamine model for efficacy studies, both because the deficits found in high schizotypes are more consistent with those observed in patients with schizophrenia and since there is evidence that some of these deficits are attenuated yb antipsychotic and / or cognition enhancer (nicotine) treatment. ABSTRACTS A3

S04 PROGRESS AND NEW DIRECTIONS IN BEHAVIOURAL ASSESSMENT OF THERAPEUTIC CANDIDATES FOR PSYCHOSIS IN RODENT MODELS Moran PM, School of , Univ of Nottingham, Univ Park, Nottingham NG7 2RD [email protected] New directions in classification of psychiatric symptoms in terms of Research Domain Criteria suggest that behavioral evidence will make an important contribution to the identification, evaluation and translation of novel biological targets into new treatments for psychosis. Rather than creating models based on clinical diagnoses, it may be useful to aim to understand the pathophysiology of that particular construct in detail. For example, rather than looking for an analogue of hallucinations, different domains that contribute to that symptom such as abnormal associative learning, negative valence system and social processes can be measured instead. These are tractable behavioural constructs whose neural circuitry in many cases has already been well established. Simpler behavioural readouts from neural circuits such as locomotor activity and stereotypy are also useful models being used to identify where compounds specifically interact. Targeting of specific neural circuits directly in the first instance bypassing behaviour is an approach also increasingly being adopted for e.g. initially determining whether compounds interact with the default mode network as opposed to initially measuring their effects on attentional behaviour. In this presentation I will provide an overview of these approaches. Using associative learning abnormality as an illustrative example I will propose a framework for how these approaches could potentially be amalgamated to simultaneously advance identification and assessment of new targets and the underlying biological basis of psychosis.

S05 5-HT AND EMOTIONALITY: NEW METHODS, NEW MECHANISMS Bannerman DM, , Univ of Oxford, 9 South Parks Road, Oxford OX1 3UD [email protected] Lima J(1), McHugh SB(1), Barkus C(2), Sengupta A(3), Sharp T(3) (1) Dept of Experimental Psychology, Univ of Oxford; (2) Dept of Expt Psychology, Univ of Oxford; (3) Dept of Pharmacology, Univ of Oxford Introduction: Serotonin (5-HT) and the serotonin transporter (SERT; which regulates the amount of at the synapse) are both heavily linked to different aspects of emotionality, as well as being implicated in mood disorders, including anxiety and depression. However, their precise contribution(s) remain(s) unclear. Recent thinking suggests that, rather than contributing directly to mood state (i.e. the idea that 5-HT is a happy chemical), 5-HT may act to facilitate certain forms of learning, thus increasing sensitivity to the environment and, as a result, altering mood state indirectly in this way. Variation in SERT levels may, therefore, influence learning about the world (particularly during early life) and thus potentially predispose individuals to altered mood. However, evidence in support of this idea is equivocal. Methods: Genetically modified mice with altered levels of the transporter, such as SERT over- expressing (SERT OE) mice and SERT KO mice, are useful tools for investigating the contribution of 5-HT and SERT to mood. In particular, they allow tight control of both genetic background and environmental conditions throughout the lifetime of the animal, in marked contrast to the difficulties encountered in human studies. We have studied associative learning in SERT OE and SERT KO mice and their respective wild-type littermate controls, as well as in mice receiving the selective serotonin inhibitor (SSRI) citalopram, using paradigms such as Pavlovian fear conditioning. Behavioural studies were combined with simultaneous amygdala recordings of local field potentials and a novel technique, tissue oxygen voltammetry (TOV) which measures the concentration of tissue oxygen in a given brain region, and thus provides a haemodynamic signal of neural activity which is broadly analogous to the BOLD signal used in fMRI. TOV therefore represents an important approach for translating between rodent studies and humans. Results/Conclusions: SERT OE mice displayed impairments in fear conditioning to an auditory cue that was associated with shock. In contrast, SERT KO mice appeared to form these associations faster than their wild-type littermate controls. Furthermore, these changes in learning were reflected in differences in signals of neural activity (tissue oxygen concentration and local field potential) recorded in the basolateral A4 ABSTRACTS amygdala. Our data suggest that variation in SERT can influence associative learning about environmental cues and thus potentially determine the effect of prior life events on present mood state. Work was supported by the Wellcome Trust.

S06 5-HT, HUMAN MOTIVATION AND LEARNING den Ouden HEM, Donders Inst for Brain, Cognition and Behaviour, Radboud Univ, Kapittelweg 29, 6525EN [email protected] The importance of neuromodulators serotonin and dopamine is evidenced by their involvement in a wealth of neuropsychiatric disorders. Pinning down their exact computational roles, however, has proven to be less than straightforward. Particularly serotonin has been a conundrum. Unlike for dopamine, there is no well-developed theoretical or formal framework to guide and interpret empirical research. Predominantly, serotonin has been associated with anticipation of punishment on the one hand, and inhibition of behaviour on the other. Recent theorising has proposed to reconcile these ideas, suggesting that serotonin plays a role in preventing active responses in the face of threat. I will present a number of human pharmacology and genetics findings that only partly support this idea, and discuss their implications for our understanding of the role fo serotonin at the interface of motivation and inhibition.

S07 SEROTONERGIC INFLUENCES ON VALENCE IN MODEL-BASED LEARNING Voon V, Univ of Cambridge, Addenbrookes Hospital, Cambridge CB2 0QQ [email protected] Daw ND(1), Worbe Y(2), Robbins TW(2), Kaasinen V(3), Joutsa J(4) (1) Princeton Univ; (2) Univ of Cambridge; (3) Univ of Turku; (4) Univ of Turku Our decisions are based on parallel interacting processes of goal-directed and habitual choices. Serotonin plays an emerging role in the modulation of this process particularly by its relationship with reward and punishment. Here I describe a series of studies assessing the role of serotonin using the two-step task which provides a relative weighting of goal-directed model-based and habitual model-free behaviours. I first describe the underlying neural correlates using volumetric, resting state functional connectivity and orientation dispersion index demonstrating parallels with rodent mapping of goal-directed and habit learning processes. I then focus on the influence of serotonin and specifically its effects on reward and loss processes using tryptophan depletion and PET imaging of serotonin transporter function in healthy controls along with investigations in populations with likely serotonergic deficits. Together these convergent findings support an influence of serotonin in humans on the shift between model-based and model-free choices as a function of valence. These studies were supported by the Wellcome Trust and NIHR Biomedical Research Council at the University of Cambridge.

S08 NEUROMODULATION OF MORAL COGNITION Crockett MJ, Dept of Experimental Psychology, Univ of Oxford, 9 South Parks Road, Oxford OX1 3UD [email protected] Concern for the suffering of others is central to morality and is disturbed in antisocial behavior at great cost to society. Understanding the psychological and neurobiological mechanisms of moral cognition is critical for identifying channels for behavior change. One set of interventions involves manipulating chemical neuromodulator systems, which play a causal role in shaping moral behavior and are impaired in many psychiatric disorders. In this talk I will present a series of studies investigating how neuromodulators influence moral judgment and decision-making. I will discuss the many challenges that arise in this line of work and introduce a novel computational framework for measuring morality in the lab with great precision and specificity. Within this framework I will present new findings describing how people evaluate the suffering of others and how the neuromodulators serotonin and dopamine shape this evaluation process. ABSTRACTS A5

S09 SEX BIASED STRESS SIGNALING: A MOLECULAR BASIS FOR SEX DIFFERENCES IN NEUROPSYCHIATRIC DISEASES Valentino RJ, Anesthesiology and Critical Care Medicine, Univ of Pennsylvania, 402D Abramson Pediatric Res Ctr, Osler Cr. Philadelphia, PA 19104 [email protected] Introduction: Elucidating the neurobiological bases for sex differences in disease prevalence is an important goal and a challenge. Many psychiatric diseases, including depression and post-traumatic stress disorder, that are more prevalent in females are associated with stress. This implicates substrates of the stress response as points at which sex differences in pathology can be expressed. Because corticotropin-releasing factor (CRF) orchestrates the stress response, we tested the hypothesis that sex differences in CRF underlie female prevalence in stress-related neuropsychiatric disorders. Methods: Neuronal sensitivity to CRF was greater in female rats. To determine whether this resulted from differences in CRF receptor (CRF1) signaling, CRF1 coupling to Gs and β-arrestin 2 was quantified using receptor immunoprecipitation. Stress-induced CRF1 trafficking in male and female rat locus coeruleus (LC) neurons was compared using immunoelectron microscopy. Because CRF is hypersecreted in stress-related psychiatric diseases, LC neuronal activity and CRF1 trafficking were also examined in CRF-overexpressing mice (CRF-OE). Finally, a global phosphoproteomic analysis comparing cortex of male and female CRF-OE and wildtype littermates tested the hypothesis that sex differences in CRF1 coupling to Gs and β-arrestin 2 translate to distinct phosphoprotein profiles that could reveal mechanisms underlying sex differences in neuropsychiatric diseases. Results: CRF1 exhibited greater coupling to Gs in unstressed females compared to males, consistent with a greater CRF-elicited neuronal activation. Following swim stress, CRF1 coupling to beta-arrestin 2 increased in males but not females and this was associated with CRF1 internalization in males only. Although LC innervation by CRF was equally dense in male and female CRF-OE mice, LC neurons were hyperactive only in female CRF-OE mice. Electron microsopic analysis suggested that this was due a lack of CRF1 internalization. The phosphoproteomic results were consistent with a female bias towards Gs-protein-dependent signaling relative to males. Thus, phosphoproteins that were significantly elevated in female CRF-OE cortex compared to male CRF-OE cortex were overrepresented in the protein kinase A signaling pathway. Additionally, they were overrepresented in an Alzheimer’s disease pathway. Conclusions: CRF1 exhibits sex-biased signaling that can account for sex differences in the prevalence of stress-related psychiatric disorders. Enhanced CRF1- Gs coupling and decreased CRF1-beta-arrestin 2 association render females more sensitive to CRF and less able to adapt to excessive CRF that occurs in stress-related psychiatric diseases. Importantly under conditions of excessive CRF, sex biased CRF1 signaling gives rise to sexually distinct phosphorylation reactions that may determine sex biases in vulnerability to certain neuropsychiatric diseases, including Alzheimer’s disease. Supported by PHS MH 040008 and MH 093981.

S10 PREMENSTRUAL SYNDROME – IS IT ALL DOWN TO METABOLISM OF ? Lovick TA, Physiology, Pharmacology and Neuroscicnce, Univ of Bristol, Medical Sciences Building, University Walk, Bristol BS8 1TD [email protected] DelBen CM(1), Guapo VG(1), Brandao ML(2), Anselmo-Franci JA(3) (1) Division of Psychiatry, Faculty of Medicine, Univ of São Paulo-Ribeirão Preto, Brazil; (2) Instituto de Neurociências e Comportamento, Ribeirão Preto, Brazil; (3) Laboratory of Neuroendocrinology, School of Dentistry, Univ of São Paulo- Ribeirão Preto, Brazil Premenstrual syndrome (PMS) has an adverse affect on daily life in around 30% of women. PMS is characterised by symptoms that include irritability, anger and aggression, which increase steadily during the luteal phase, reaching a peak in the week prior to menstruation. A co-dependent of PMS is the cyclical changes in gonadal hormone secretion during the menstrual cycle. Progesterone secretion increases significantly during the early- to mid-luteal phase, but declines during the week prior to menstruation, when symptoms of PMS reach a peak. Thus paradoxically, development of symptoms of PMS during the luteal phase appears to be associated with both an increase and a decrease in progesterone. At physiological concentrations (but not above), progesterone’s neuroactive metabolite promoted A6 ABSTRACTS aggression in mice (Miczek et al, 2003, Hormone Behav 44, 242–257). Using female rats as a model we have shown that when progesterone concentration falls rapidly during the late diestrus phase of the estrous cycle responsiveness to acute psychogenic stress increased (Devall et al, 2009 Psychoneuroendocrinol 34, 587–596). Anxiety is also evoked during withdrawal from a chronic exogenous progesterone/estrogen dosing regimen in rats, but only when hormone withdrawal was rapid; gradual withdrawal had no effect (Doornbos et al, 2009, Life Sci 84, 69–74). To try and reconcile these findings we examined in women, the relationship between daily salivary progesterone concentration during the luteal phase and the development of PMS symptoms. Methods: 28 healthy premenopausal women with normal menstrual cycles provided daily waking saliva samples and completed a daily questionnaire regarding premenstrual symptoms during the luteal phase of one menstrual cycle. Results: 62.5% of participants developed symptoms of PMS during the luteal phase; the remainder (37.5%) were asymptomatic. In the former group symptoms increased progressively throughout the luteal phase; progesterone concentration in saliva also increased but then declined sharply 3 days prior to menstruation. In contrast, in asymptomatic participants progesterone was maximal 6–7 days prior to menstruation and declined slowly thereafter. Peak progesterone concentration was similar in both groups (249.1± 36.8 vs. 227.0 ± 43.7 pg/ml). Conclusions: two factors with different underlying mechanisms appear to underlie the progressive rise in PMS symptoms during the luteal phase. Symptoms in the early part of the luteal phase may relate to the aggression-promoting effect of rising concentration of progesterone (and hence allopregnanolone) in the plasma and brain. Symptoms during the late luteal phase, when progesterone concentration declines sharply may represent an anxiogenic withdrawal effect. Financial support from ERUK/FAPESP Newton Fund and CNPq Research without Borders scheme is gratefully acknowledged.

S11 DIFFERENTIAL EFFECTS OF OXYTOCIN IN MEN AND WOMEN Hurlemann RH, Dept of Psychiatry & Division of , Univ of Bonn, Sigmund-Freud- Strasse 25, Bonn, Germany 53105 [email protected] Introduction: Current perspectives on the neurochemistry of social behavior emphasize a key role of the evolutionarily conserved hypothalamic peptide oxytocin (OT). Disruptions in OT signaling are thought to underlie deficient social skills in various disorders angingr from autism to schizophrenia. As a consequence, the OT system is emerging as a promising target for novel treatment approaches in psychiatry. However, accumulating evidence for sex-dependent effects of OT hampers prediction of potential clinical efficacy. Methods: In a series of behavioral and experiments, we characterized the modulatory influence of intranasally administered OT (IN-OT) on several social readouts, including social distance, social touch, and partner preference. These experiments were carried out independently from each other in separate samples of male and female volunteers. Results: Our results indicate that IN-OT effects on social distance vary substantially, depending on the subjects’ sex and their relationship status. However, the female participants in this experiment were using hormonal contraceptives, which may significantly alter the sensing of and responding to social stimuli. Indeed, sex differences were absent regarding IN-OT effects on the behavioral and neural responses to social touch and pictures of the partner, when female volunteers were not using hormonal contraceptives. Conclusions: Our findings suggest common OT-dependent neural mechanisms underlying a diverse repertoire of social behaviors in men and women. However, these mechanisms may be altered in women using homonal contraceptives, implicating significant interactions of IN-OT with gonadal steroids. This interplay may significantly influence, and interfere with, the clinical efficacy T.of O

S12 BIPOLAR DISORDER AND POSTPARTUM PSYCHOSIS – AN IMPORTANT CLUE TO THE AETIOLOGY OF MOOD DISORDERS Jones IR, National Centre for Mental Health, Univ, Haydn Ellis Building Maindee Rd Cardiff CF24 4HQ [email protected] Introduction: It has long been recognised that childbirth is a time of considerable risk for women with bipolar disorder, with postpartum episodes (including postpartum / puerperal psychosis) occurring ABSTRACTS A7 following up to 50% of deliveries. Childbirth is therefore a very potent trigger for bipolar episodes and highlights the importance of considering issues regarding pregnancy and childbirth in women with this condition. In this talk I will review the evidence confirming that women with bipolar disorder are at very high risk of episodes of severe postpartum affective disorder and examine what is known about the factors that influence this risk. Methods: The Bipolar Disorder Research Network study (www.BDRN.org) has recruited over 6,000 bipolar participants in the UK for clinical and genetic studies. I will review what we have learnt about postpartum triggering in bipolar disorder from this study and discuss the direction our research in this area is taking. Results: Work from our group has demonstrated that familial (genetic) factors influence vulnerability to the puerperal trigger and based on these findings we are conducting molecular genetic studies to identify the genetic variants that confer risk. Conclusions: This line of research has the potential to uncover the nature of the puerperal trigger, allow a more individualised estimation of risk for women with bipolar disorder, and provide further information relating to the aetiology of mood disorders in relation to childbirth and at other times.

S13 THE CLOCK Δ19 MOUSE – THE FIRST ANIMAL MODEL OF BIPOLAR DISORDER? McClung CA, Psychiatry, Univ of Pittsburgh School of Medicine, 450 Technology Drive, Suite 223 Pittsburgh, PA, 15219 [email protected] Bipolar disorder has been difficult to study in a pre-clinical setting since the development of an animal model of this disorder has been challenging. Here I will present evidence to suggest that mice with a mutation in the circadian regulator, Clock, have many features that are similar to those seen in people with bipolar disorder, including mood switching between manic and euthymia-like states. Nearly all people suffering from psychiatric disorders have significant disruptions in circadian rhythms and the sleep/wake cycle. Furthermore, disruptions to circadian rhythms including shift work, overseas travel, and irregular social schedules tend to precipitate or exacerbate mood and psychotic episodes. Thus I will summarize recent data from our lab using the Clock mutant mouse model to understand the molecular and cellular basis of mood regulation and the switching between mood states. I will also discuss how circadian rhythm stabilization might be therapeutic in the treatment of bipolar disorder and perhaps other psychiatric disorders. This work is funded by the NIH (NIMH, NIDA), IMHRO, The McKnight foundation and The Brain and Behavior Foundation (NARSAD).

S14 GENETIC ADVANCES IN BIPOLAR DISORDER AND WHAT THEY TELL US Craddock N, MRC Centre for Neuropsychiatric Genetics and Genomics, Cardiff Univ, Hadyn Ellis Building, Maendy Road, Cardiff , CF24 4HQ [email protected] Family and twin studies demonstrate the importance of genetic factors influencing susceptibility to mood disorders in general and bipolar disorder in particular and point to substantial genetic and phenotypic complexity. Robust and replicable genome-wide significant associations for bipolar disorder have recently been reported in genome-wide association studies at several common polymorphisms. There is strong evidence for a polygenic contribution to risk (ie. many risk alleles of small effect). A striking finding is the overlap of susceptibility between bipolar disorder and schizophrenia shown for several individual risk alleles and for the polygenic risk. In contrast, genomic structural variation seems to play a lesser role in bipolar disorder than in schizophrenia. Together, these genetic findings suggest avenues for future studies for delineating the aetiology and pathogenesis of bipolar disorder, they indicate the need to re-evaluate our diagnostic classifications and they may eventually pave the way for major improvements in clinical management. The findings to date for major depression are less clear-cut. This is consistent with its lower heritability and a greater genetic complexity and indicate the need for study of very large samples. The work presented was supported by grants from the Wellcome Trust and the Stanley Centre for Psychiatric Research. Invaluable assistance was provided by members and volunteers of the Bipolar Disorders Research Network. A8 ABSTRACTS

S15 PREDICTING BIPOLAR DISORDER: THE EDINBURGH HIGH-RISK STUDY (AKA BIPOLAR FAMILY STUDY) Lawrie SM, Psychiatry, Univ of Edinburgh, Kennedy Tower, Royal Edinburgh Hospital, Morningside, Edinburgh EH10 5HF [email protected] Introduction: Abnormalities of mood-related brain circuitry are understood to underlie symptoms of Bipolar disorder (BD) and Major Depressive Disorder (MDD). However whether these abnormalities can distinguish those at greatest risk, how they change over the course of illness development, or how they relate to other risk factors, has yet to be fully determined. Methods: The Scottish Bipolar Family Study (BFS) is a prospective longitudinal imaging study examining young individuals (16–25 years) at high familial risk of mood disorder, along with a group of healthy controls. Individuals were categorised into high-risk who remained well (HR Well), high-risk who developed a mood disorder, primarily MDD (HR MDD), and healthy controls. Here we report new findings using the most up-to-date clinical information on the BFS participants including DTI/tractography, structural VBM, emotion processing fMRI and new preliminary findings integrating imaging with methylation data. Results: Comparisons of groups using structural VBM indicated significant decreases over time in the right amygdala in the HR MDD group. Using a facial emotion processing task we found significantly reduced activation of the sub-genual anterior cingulate cortex (sgACC) in the HR MDD. Longitudinal DTI however revealed differences between the HR groups as a whole versus controls but not between HR groups. Conclusions: These results illustrate neural abnormalities in mood-related brain circuitry in individuals at familial risk in the early stages of mood disorder. The findings indicate that some neurobiological differences are associated with increased risk, and some are specifically associated with the presence of MDD or impending diagnosis. This study has been sponsored by the Health Foundation and the Sackler Foundation.

S16 THE USE OF NOVEL DEVICES AND REMOTE MONITORING FOR CLINICAL MANAGEMENT AND RESEARCH IN BIPOLAR DISORDER Harrison PJ, Dept of Psychiatry, Univ of Oxford, Neurosciences Bldg, Warneford Hospital, Oxford OX3 7JX [email protected] Geddes J(1) (1) As presenting author Assessment and follow-up of patients with psychiatric disorders has traditionally been carried out using face-to-face interviews for cross-sectional and retrospective recall of symptoms and events. Now digital technologies provide the opportunity to move beyond the inherent limitations of this approach, allowing both clinicians and researchers to conduct remote, longitudinal monitoring of illness, plus the ability to capture a range of behavioural, cognitive and biological data. I will describe our experiences using one such digital platform, True Colours, in bipolar disorder. Originally just a means to capture and track symptoms via text prompts, the capabilities of True Colours have expanded and been complemented by a range of apps, to allow collection of data regarding activity, location, sleep, and cognitive performance. Parallel research is investigating the neural and molecular correlates of the findings, and the impact of pharmacological and psychological interventions. Novel mathematical approaches to the resulting data streams are also providing new insights into aspects of pathophysiology and treatment response. I will give examples showing how True Colours has thereby enhanced clinical care, shed new light on the course of illness, aided in differential diagnosis, been used in a randomised clinical trial, and is now being rolled out to new cohorts and disorders. In summary, novel devices and remote monitoring have the potential to transform psychiatric clinical care and research, providing more meaningful and quantitative data, and allowing scaling up to much larger populations than was feasible using conventional methods. However, there remain major practical and ethical challenges to overcome, and it is important that enthusiasm for technology is tempered by its careful evaluation. Acknowledgement: Work supported by a Wellcome Trust Strategic Award, CONBRIO: Collaborative Network for Bipolar Research to Improve Outcomes (102616/Z). ABSTRACTS A9

S17 INTRANASAL ESKETAMINE IN TREATMENT RESISTANT DEPRESSION – A DOUBLE-BLIND, RANDOMIZED, EFFICACY AND DOSE-RESPONSE STUDY Singh JB, Neuroscience, Janssen, 3210 Merryfield Row, San DIego, CA, USA 92121 [email protected] Background: Esketamine and ketamine have been shown to produce rapid antidepressant action in patients with treatment-resistant depression (TRD). The aim of the current study was to assess the efficacy, safety and dose response of intranasal esketamine in patients with TRD. Methods: This was a 2-Panel, doubly-randomized, double-blind, placebo-controlled, multicenter study. Panel A was conducted in the United States and Belgium and Panel B is currently ongoing in Japan. In both panels, each subject participated in up to 4 phases: a screening phase of up to 4 weeks, a double-blind treatment phase which included two 1-week periods (Periods 1 and 2), a 9-week optional open-label treatment phase and an 8-week posttreatment follow-up phase. Only Panel A double blind phase data are available, and will be presented at this time. The primary efficacy endpoint was the change from baseline to Day 8 in each period in the Montgomery-Asberg Depression Rating Scale (MADRS) total score combined. Safety and secondary efficacy endpoints were also assessed. Results: A total of 67 subjects with TRD were randomly assigned in a 3:1:1:1 ratio to one of four treatment groups: placebo (n=33), esketamine 28 mg (n=11), esketamine 56 mg (n=11), or esketamine 84 mg (n=12) in Period 1. In Period 2, 28 placebo subjects who were eligible for re-randomization at the end of Period 1 were randomly assigned to placebo (n=6), esketamine 28 mg (n=8), esketamine 56 mg (n=9), or esketamine 84 mg (n=5) in a 1:1:1:1 ratio. Subjects were eligible for re-randomization if the patient-rated 16 item Quick Inventory of Depressive Symptomatology (QIDS-SR16) total score was ≥11 at the end of Period 1. The analysis of Period 1 and Period 2, combined using the weighted combination test, showed that the mean change in MADRS total score in all three esketamine groups was statistically superior to that obtained under placebo, based on a one-sided 0.05 significance level (p=0.021, p=0.001 and p<0.001 for esketamine 28 mg, 56 mg and 84 mg respectively). The mean differences (SE) from placebo (after one week of treatment) were -4.2(2.09) for esketamine 28 mg, -6.3(2.07) for esketamine 56 mg, and -9.0(2.13) for esketamine 84 mg. The magnitude of effect size in Period 1 increases from a low Cohen’s D effect size in the 28 mg dose group (0.43) to a high Cohen’s D effect size for the 56 (0.92) and 84 mg (1.19) dose groups. The most common TEAEs during the double-blind phase (≥10% of subjects in any group) were: dizziness, dissociation, headache, dysgeusia, nasal discomfort, nausea, hypoaesthesia oral, dissociative symptoms, tunnel vision, oropharyngeal pain, throat irritation, blurred vision, hypersomnia, feeling abnormal, insomnia, hypertension, vertigo, polyuria and sedation. No death was reported. Transient elevation in blood pressure and heart rate was also observed on dosing days. The perceptual changes and dissociative symptoms measured by the Clinician administered Dissociative Symptom Scale (CADSS), suggest onset of these symptoms occurred shortly after the start of intranasal dosing and resolved by 2 hours postdose, and with repeated dosing these symptoms reduced significantly. Conclusions: Intranasal esketamine administered in doses of 28, 56 and 84 mg across the study period showed statistically and clinically significant improvement of depressive symptoms in subjects with TRD, as demonstrated by the mean changes in the MADRS total score for the combined analysis of both periods. The doses evaluated were well tolerated and adverse events were similar to what has been observed previously with IV ketamine and eskatamine. Study was sponsored by Janssen Pharmaceuticals.

S18 KETAMINE ASSISTED THERAPY OF OPIATE DEPENDENCE Krupitsky E, Addictions, St.-Petersburg Bekhterev Research Psychoneurological Insitute, Bekhtereva Street, 3, St.-Petersburg, Russia 192019 [email protected] Introduction: Ketamine is a prescription drug used for general anesthesia. In subanesthetic doses, it induces profound dissociative state that includes psychedelic experiences and hallucinations. Methods: Seventy detoxified patients with opiate dependence were randomly assigned to one of two groups receiving ketamine (KPT) involving two different doses of ketamine. There were 35 opiate addicts (27 male and 8 female) in the experimental group, and 35 opiate addicts (28 male and 7 female) in A10 ABSTRACTS the control group. The patients of the experimental group received existentially oriented psychotherapy in combination with a hallucinogenic (“psychedelic”) dose of ketamine (2.0 mg/kg i.m.). The patients of the control group received the same psychotherapy combined with a very low, non-hallucinogenic (non- psychedelic), dose of ketamine (0.2 mg/kg i.m.). This low dose induces some pharmacological effects without inducing a peak psychedelic experience. Both the psychotherapist and patient were blind to the dose of ketamine. KPT included preparation for the ketamine session, the ketamine session itself, and the post session psychotherapy aimed to help patients to integrate insights from their ketamine session into everyday life. During the ketamine session, the psychotherapist provided emotional support for the subject and carried out psychotherapy. Results: The results of this double blind randomized clinical trial of KPT for opaite addiction showed that high dose (2.0 mg/kg) KPT elicits a full psychedelic experience in heroin addicts as assessed quantitatively by the Hallucinogen Rating Scale. On the other hand, low dose KPT (0.2 mg/kg) elicits “sub-psychedelic” experiences and functions as ketamine-facilitated guided imagery. High dose KPT produced a significantly greater rate of abstinence in opiate addicts within the first two years of follow-up than did low dose KPT. High dose KPT elicited a greater and longer-lasting reduction in craving for opiates, as well as greater positive change in nonverbal unconscious emotional attitudes. Conclusion: KPT might be an effective method of treatment of opiate dependence. More research in this direction is needed. Supported by Mutidisciplinary Association of Psychedelic Studies and Heffter Research Inst.

S19 REPEATED USE OF KETAMINE IN THE TREATMENT OF DEPRESSION AND ADDICTION: INSIGHTS FROM DRUG USERS Morgan CJ, Dept of Psychology, Univ of Exeter, Washington Singer Labs, Perry Rd, Exeter EX4 4QG [email protected] Curran HV(1) (1) Clinical Psychopharmacology Unit, UCL Ketamine has thus been hailed as the major advance in the treatment of depression of the past 50 years and now looks set to emerge as a useful treatment for addiction. The impressive antidepressant effects of ketamine have spurned a great deal of research interest, but also a growing off-label use of the drug. In addition, there is considerable recreational use in certain countries, which has led to repeated calls for tighter regulatory controls on ketamine. Some treatment regimes in development involve repeated doses of ketamine in depression and addiction. Recreational use of ketamine can provide a window on the effects of repeated ketamine use, before definitive trials are completed. In this talk, I shall consider the potential consequences of repeated use of ketamine, in the context of the current clinical uses and future promise of the drug, alongside some of the ethical concerns that have been voiced. We shall also briefly consider the potential mechanisms for the action of ketamine in addiction and describe a new study that we have just begun that aims to further investigate the use of ketamine in the reduction of relapse in dependence. Financial Support This work is supported by a grant from the Medical Research Council.

S20 NEW DIRECTIONS IN MODULATION TO TREAT SUBSTANCE USE DISORDERS Dakwar E, Columbia Univ, Coll of Physicians and Surgeons, 1051 Riverside Drive, NY 10032, USA [email protected] Repeated drug consumption may progress to problematic use by triggering neuroplastic adaptations that attenuate sensitivity to natural rewards while increasing reactivity to craving and drug cues. Converging evidence suggests a single sub-anesthetic dose of the N-methyl-D-aspartate receptor antagonist ketamine may work to correct these neuroadaptations and restore motivation for non-drug rewards. In human laboratory settings with dependent individuals, we have found that ketamine, as compared to the control, at more than 24 hours post-infusion improves motivation to quit drug use; decreases behavioral reactivity; lessens cue-induced craving; and decreases cocaine self-administration. The latter effect is the most robust reduction observed to date in human cocaine users and the first to involve mechanisms other than stimulant or dopamine effects. We also discuss our work with ketamine in clinical settings, ABSTRACTS A11 highlighting an interim analysis that indicates ketamine promotes abstinence and prevents relapse. Taken together, these findings signal new directions in medication development for substance use disorders.

S21 ADULT NEUROGENESIS IN HUMANS Bergmann O, Dept of Cell and Molecular Biology, Karolinska Inst, Berzelius väg 35 SE-171 77 Stockholm, Sweden [email protected] Adult neurogenesis is present in dentate gyrus of rodents and primates, including humans. The functional significance of adult neurogenesis is debated, because adult-generated neurons have been thought to very few in humans. We established the dynamics of hippocampal neurogenesis in adult humans determined by measuring the concentration of nuclear-bomb-test-derived 14C in genomic DNA. In humans approximately one-third of hippocampal neurons are renewed. In middle-aged humans, 700 new neurons are exchanged in each hippocampus per day, corresponding to an annual neuronal turnover of 1.75%, with a modest decline during aging. Neurons are generated throughout adulthood, suggesting that adult hippocampal neurogenesis may contribute to human brain function and play a role in psychiatric conditions.

S22 SEX AND HORMONE DIFFERENCES IN DEPRESSION: RELATIONS TO NEUROGENESIS AND ANTIDEPRESSANT TREATMENT Galea LAM, Psychology, Centre for Brain Health, The Univ of British Columbia, 2136 West Mall, Vancouver, BC Canada, V6T1Z4 [email protected] Mahmoud R(1), Wainwright S(1), Workman J(1) (1) As presenting author Introduction: Sex differences exist in the prevalence and symptoms of depression. The integrity of the hippocampus is compromised in depression perhaps differentially in men versus women. In this talk I explore these sex differences to better understand the neural manifestation of depression with an eye towards an understanding of how treatment of depression in men and women may need to be tailored. I discuss findings from post-mortem tissue, and in animal models capitalising on gonadal hormone manipulations, in both males and females to determine if there is a difference in neurogenesis and vulnerability to depressive-like endophenotypes. In addition, motherhood itself is associated with increased risk to develop anxiety and depression and I discuss how motherhood alters the response to fluoxetine in a model of postpartum depression. Methods: Gonadectomized male and female rats are given chronic unpredictable stress and tested for depressive-like behaviours and measures of neurogenesis in the hippocampus. In a separate study postpartum dams were treated with vehicle or high corticosterone (40 mg/kg) to induce a depressive-like phenotype. Within these conditions, dams either received: fluoxetine (10 mg/kg) or vehicle. Dams were tested for maternal behavior, depressive-like behaviour and neurogenesis in the postpartum. Results: Gonadectomised males and females show increased vulnerability to stress- induced endophenotypes of depression, with perhaps a greater link to decreased hippocampal neurogenesis in males compared to females. Ovarian hormones in females, and testicular hormones in males, reduced the vulnerability to develop depressive-like endophenotypes. CORT-treated dams exhibited increased depressive-like behavior, disrupted maternal care, and suppressed neurogenesis in the hippocampus. Fluoxetine reversed the disruptions in maternal care but not depressive-like behaviour or neurogenesis in the dams, whereas fluoxetine in nulliparous rats increased hippocampal neurogenesis. Fluoxetine reversed disrupted negative hypothalamic-pituitary-adrenal feedback, but not depressive-like behaviour, in sham, but not ovariectomized, rats. Conclusions: Together these studies show that fluctuations in steroid hormones increase vulnerability to develop a depressive-like phenotype and that pharmacological antidepressants abrogate some depressive-like behaviours dependent on reproductive status. These data have implications for treatment of depression during times of steroid hormone modulation. Given sex differences in hippocampal function and morphology, it is important to understand hippocampal functioning in both males and females in order to have a better understanding of disease mechanisms that target the hippocampus. Acknowledgements: Funding from CIHR (MOP 142308) and Coast Capital Depression Fund. A12 ABSTRACTS

S23 (EARLY) STRESS EFFECTS ON NEUROGENESIS AND PLASTICITY Lucassen PJ, SILS – Center for Neuroscience, Univ of Amsterdam, Science Park, 904 Room C3.272 1098 XH Amsterdam , The Netherlands [email protected] Naninck EF(1), Abbink M(1), Yam K-Y(1), Korosi A(1) (1) As presenting author; Introduction; Early-life adversity increases the vulnerability to develop psycho-pathologies and cognitive decline later in life. Yet, the key determinants within the early environment underlying these associations remain poorly understood. Interestingly, various stressors including abuse, neglect or malnutrition all appear to similarly affect brain structure and function long-term. This suggests various elements act upon a common substrate and/or molecular pathway. Methods/Results; We study the rodent hippocampus as it largely develops postnatally, is sensitive to stress and continues to produce new neurons up into adult age, a process known as adult neurogenesis that has been implicated in cognition and brain plasticity (Lucassen PJ, Korosi.A Regulation of Adult Neurogenesis. Cold Spring Harb Perspect Biol. 2015). In addition to stress and maternal care, early nutrition provides an important signal from the mother to her pups that can modify brain development with lasting consequences for behavior and cognition in later life (Lucassen PJ, Korosi A. Perinatal programming of hippocampal structure and function; emerging roles of stress, neurogenesis, epigenetics and early nutrition. Trends Neurosci 2013). Conclusions; the hippocampus is particularly sensitive to stress occurring during the early life period, and shows lasting changes in neurogenesis and cognition, that also appear sensitive to early nutritional status and often occur in a sex-dependent manner (Naninck EF, Korosi A. Chronic early life stress alters developmental and adult neurogenesis and impairs cognitive function in mice. Hippocampus. 2015 Mar;25(3):309–28; Loi M et al, Effects of early-life stress on cognitive function and hippocampal structure in female rodents. Neuroscience. 2015).

S24 ANTIDEPRESSANTS REGULATE NEUROGENESIS, SYNAPTOGENESIS AND DENDRITOGENESIS IN HUMAN STEM CELLS MODELS Pariante C, Inst of Psychiatry, Psychology and Neuroscience, KCL, G.32.01, The Maurice Wohl Clinical Neuroscience Inst, Cutcombe Road, SE5 9RT [email protected] Zunszain P(1) (1) As presenting author Introduction: We have successfully developed models of “depression in a dish” that use human neuronal cells and that we have successfully used to translate findings from-bench-to-bedside-and-back. Methods: We have used two cellular models: the conditionally immortalized, multipotent, foetal hippocampal neuronal precursor cell-line HPC03A/07; and neurones differentiated from iPS cells donated by healthy controls. Results: Using the neuronal precursor cell-line HPC03A/07: 1) we have identified that antidepressants upregulate the pro-neurogenic gene, p11, and then replicated this finding in peripheral blood mRNA of depressed patients before and after eight weeks of antidepressants Anacker C et al., Proc Natl Acad Sci U S A. 2013 May 21;110(21):8708–13); 2) we have demonstrated that interleukin-1 beta (a depressogenic insult) decreases neurogenesis, and that this reduction can be rescued by inhibiting the production of neurotoxic tryptophan metabolites as well as by direct administration of antidepressants in vitro Zunszain P et al., Neuropsychopharmacology. 2012 Mar;37(4):939–49; Horowitz m et al., Int J Neuropsychopharmacol. 2014 Oct 31;18(3)); 3) using dexamethasone (another depressogenic insult), we have shown epigenetic and gene expression changes in the stress-relevant gene, FKBP5, that are identical to those identified in peripheral blood mRNA of depressed patients (Klenget T et al., Nat Neurosci. 2013 Jan;16(1):33–41.; and 4) we have used transcriptomics of these cells treated with cortisol (yet another depressogenic insult) to identify gene expression pathways regulating human neurogenesis, and then replicated these pathways by conducting brain transcriptomics of rats exposed to stress as well as candidate gene expression analyses of peripheral blood mRNA in patients with depression (Anaker C et al., Neuropsychopharmacology. 2013 Apr;38(5):872–83). In addition to these published findings, we ABSTRACTS A13 have unpublished work using induced-pluripotent stem cells from healthy controls (iPSCs) differentiated into neurons of the prefrontal cortex, and have shown that chronic exposure of iPSCs to the pro- inflammatory and depressogenic insult, interleukin-1 beta, causes a decrease in the expression of synaptic proteins and length of dendrites. This “depressive” cellular phenotype is rescued by the antidepressant strategy, ketamine. Conclusions: Using this kind of cellular models has three unique advantages: first, they can mimic clinically-relevant conditions within a tightly-controlled experimental environment, overcoming the inevitable variability of clinical samples as well as the costs (and ethical implications) of a very large number of animals; second, they deliver data from living human brain cells, inaccessible in clinical samples; and third, they generate findings that are translatable to clinical samples. We are confident that these models will now be successfully used to generate novel molecular targets for the development of antidepressants of the future. Funded by: SLAM NIHR Biomedical Research Centre; MRC Immunopsychiatry Consortium; Johnson & Johnson.

S25 TRANSLATING MODELS OF MOTIVATIONAL DISTURBANCE IN RODENTS TO PSYCHIATRIC SYMPTOMS IN HUMANS Robinson ESJ, School of Physiology, Pharmacology and Neuroscience, Univ of Bristol, Biomedical Sciences Building, University Walk, Bristol BS8 1TD [email protected] There is a long history of using animals to understand the neurobiology of reward however, how well the approaches used in animal research relates to human psychiatric is difficult to assess given the differences in experimental approaches used. Changes in the experience of reward and anticipation of reward has been investigated in clinical populations but with the majority of studies considering self-reported or questionnaire based measures. In contrast, the majority of animal studies use primary rewards (usually food) and investigate motivational changes associated with obtaining this reward. This presentation will consider the different approaches used in pre-clinical research to assess reward- related processing deficits in animal models of psychiatric disorders. The work considered will include conventional approaches, such as the sucrose preference test used to study depression-related behaviour and anhedonia, or progressive ratio tasks where motivation for reward can be quantified. There will also be a review of recent work, including studies from our own laboratory, which have tried to evaluate how reward-related processing in rodents could be studied in a more translational manner and the validation of these approaches. For example, a recent comparison between antidepressant and pro- depressant drug treatments in the affective bias test and progressive ratio task suggests that these tasks involve different neurochemical substrates. Our work in the affective bias test also suggests that reward- related learning and memory is altered in putative animal models of depression independent of changes in sucrose preference test. Funding for this research was provided by an RCUK Academic Fellowship with additional support from the British Pharmacological Society Integrative Pharmacology Fund. Research grants were also provided by BBSRC, MRC and Pfizer.

S26 REWARD PROCESSING AND MOTIVATIONAL IMPAIRMENT IN SCHIZOPHRENIA Gold JM, School of Medicine, Univ of Maryland, 655 W. Baltimore Street, Baltimore MD 21201, USA [email protected] Albrecht MA(1), Waltz JA(1), Cavanagh JF(1), Frank MJ(1) (1) As presenting author Background: In prior studies, we have suggested that the negative symptoms of schizophrenia (SZ) are associated reinforcement learning (RL) deficits likely related to degraded representations of prospective reward values, rather than a failure to signal prediction errors (PE). Such degraded representations may also account for overestimation of the cost of effort needed to obtain rewards also seen in schizophrenia. However, the RL tasks used to date have required active responses to both reward and punishing stimuli. Pavlovian biases have been shown to affect RL performance through invigoration of action to reward and inhibition of action to punishment, and may be partially responsible for RL deficits seen in A14 ABSTRACTS

SZ. Methods: Forty-five people with schizophrenia and 30 demographically-matched controls completed a four-stimulus RL task that crossed action (“Go” or “NoGo”) and the valence of the optimal outcome (reward or punishment-avoidance), as EEG was simultaneously recorded. Results: Patients demonstrated a significant reduction in Pavlovian performance bias. However, this bias did not correlate with negative symptom severity. Patients did not differ from controls in the feedback related negativity (FRN) time window suggesting intact signaling of the difference between correct and incorrect feedback. However, patients did not show the usual association between P3a amplitude and PE that reflects the updating of value representations. These findings are consistent with earlier evidence that the reward processing impairments in SZ occur at a later stage of processing than the signaling of PEs. Conclusions: People with schizophrenia show a subtle reduction in Pavlovian biases that appear to be unrelated to the severity of negative symptoms. Patients show a normal FRN, suggesting intact initial processing of error information. Significant deficits observed in the P3a time window suggest a deficit in the ability to use error signals to update representations of stimulus value. Possible explanations for inconsistent negative symptom associations across different RL tasks will be discussed. Support: This work supported by a grant from the NIMH.

S27 APATHY IN NEUROLOGICAL DISORDERS Husain M, Nuffield Dept Clinical Neurosciences and Experimental Psychology, Univ of Oxford, West Wing, John Radcliffe Hospital, Oxford OX39UD [email protected] Manohar S(1), Muhammed K(1), LeHeron C(1) (1) As presenting author Background: Apathy is now appreciated to be an extremely common disorder of motivation affecting patients across a range of neurological disorders, from neurodegenerative conditions through to focal brain lesions. Here we focus on Parkinson’s disease (PD) as a model condition to study mechanisms underlying apathy and their modulation by dopaminergic drugs. Methods: We used two tasks: 1) a new reward for effort decision-making task; 2) a saccade for reward task. This measured both pupillary response to rewards on offer as well as saccadic velocity. PD patients ON and OFF medication, young and elderly controls were tested on both tasks. Results: On the reward for effort task, PD patients without apathy displayed subclinical deficits of reward sensitivity, engaging in less effort than healthy controls for the lowest rewards when OFF dopaminergic medication. However, when ON dopaminergic medication, they showed greater willingness to allocate effort for high rewards than controls. In healthy people, this task was associated with frontostriatal activation. Surprisingly, more apathetic people showed greater activation of medial frontal premotor regions (SMA, pre-SMA and cingulate motor regions) than more motivated individuals when making decisions to allocate effort for rewards. This suggests there is a greater physiological ‘brain cost’ to making such decisions associated with apathy. On the saccade for reward task, PD patients ON dopaminergic medication and controls demonstrated increased saccadic velocity and pupil diameter with greater monetary incentives. This reward sensitivity was significantly diminished in PD patients OFF dopaminergic medication. Crucially, apathetic PD patients exhibited significantly less pupillary reward sensitivity than more motivated patients. Pupillary modulation was independent of motor impairment and autonomic dysfunction as indexed by heart rate variability measures. Conclusions: These findings reveal that apathy may be associated with altered decision-making when effort has to be allocated for reward and that such decisions may rely on frontostriatal systems. Reward sensitivity is blunted in PD patients with apathy. Dopaminergic medication significantly alters reward sensitivity and may be an effective treatment for apathy, independent of effects on motor control. ABSTRACTS A15

S28 THE NEURAL RESPONSE TO REWARD AND AVERSION AS A BIOMARKER FOR DEPRESSION AND TARGET FOR ANTIDEPRESSANT TREATMENTS McCabe C, Psychology, Univ of Reading, Whiteknights Campus, Reading RG6 6AL [email protected] Background: Anhedonia suggests reduced brain reward function. However how adolescents with depression symptomology and anhedonia respond to reward and aversion in the brain is unclear. We have shown that Selective Serotonin Reuptake Inhibitors (SSRI) blunt the neural response to reward and aversion in healthy volunteers (McCabe et al., 2010 Biological Psychiatry, 2010 Mar 1;67(5):439–45) . Although how catecholamine antidepressants like bupropion compare with this, is unknown. Methods: In Study 1 we examined 33 adolescents (13–18yrs.) with high and low depression scores. In Study 2 we examined 17 healthy volunteers receiving 7 days of bupropion (150 mg/day) and placebo, in a double- blind crossover design. Our functional MRI task examined an anticipatory phase (pleasant or unpleasant cue), an effort phase (button presses to achieve a pleasant taste or to avoid an unpleasant taste) and a consummatory phase (pleasant or unpleasant tastes). Results: In Study 1 Adolescents with depression symptoms have blunted responses to anticipation, effort and consummation of reward/aversion in regions such as the anterior cingulate and ventral striatum (p=0.02 Family Wise Error (FWE) whole brain corrected; p=0.02 FWE small volume corrected). In Study 2, Bupropion increased responses to anticipation, effort and consummation of reward/aversion in similar regions (anterior cingulate and ventral striatum p<0.001 FWE whole brain corrected). Conclusions: Taken together we show blunted neural responses to both reward and aversion in adolescents at risk of depression and this supports the idea of anhedonia as a biomarker for depression. Second we provide mechanistic evidence for how catecholamine agents are less likely than SSRIs to cause emotional blunting and may even reverse anhedonia when given to depressed patients as a first line treatment or as an adjunct to SSRIs. Funded by the Univ of Reading Start-up for Dr C McCabe.

S29 DYSREGULATED IMMUNE MECHANISMS: A THERAPEUTIC TARGET FOR PSYCHIATRIC DISORDERS? Prinssen E, NORD Discovery & Translational Area, Roche Pharma Research & Early Development, Grenzacherstrasse 124, Basel CH 4070, Switzerland [email protected] Epidemiological studies have shown a clear association between maternal infection and schizophrenia in the progeny. This maternal immune activation (mIA) is thought to lead to immune priming in the offspring and – together with genetic and other environmental risk factors – to result in a deviation from a normal development trajectory leading to schizophrenia or other CNS disorders. Data is accumulating that immune abnormalities are present throughout the trajectory of subjects that eventually develop schizophrenia as well as during the different phases of disease. Reported abnormalities include microglia number and function, astrocyte and oligodendrocyte functions, plasma cytokine levels, serum NMDA-R autoantibody levels or pathogenicity, and brain complement component 4 expression (explaining part of schizophrenia’s strongest genetic association at a population level, i.e. variation in the MHC locus). I will review this clinical evidence in light of potential symptomatic, disease-modifying and preventive treatment strategies, and discuss how mIA animal models may contribute to the implementation of these strategies. A16 ABSTRACTS

S30 ABNORMAL TRAJECTORY OF RAT BRAIN MATURATION FOLLOWING EXPOSURE TO MATERNAL IMMUNE ACTIVATION AND THE IMPACT OF PHARMACOLOGICAL INTERVENTION IN ADOLESCENCE Vernon AC, Basic and Clinical Neuroscience, King’s College London, Inst of Psychiatry, Psychology and Neuroscience, Maurice Wohl Clinical Neuroscience Inst, 5 Cutcombe Road, London SE5 9RT [email protected] So PW(1), Lythgoe DJ(1), Crum WR(1), Williams SCR(1), Chege W(2), Natesan S(2), Kapur S(2) (1) Neuroimaging, King’s College London, Inst of Psychiatry, Psychology and Neuroscience, De Crespigny Park, London SE5 8AF; (2) Psychosis Studies, King’s College London, Inst of Psychiatry, Psychology and Neuroscience, De Crespigny Park, London SE5 8AF Introduction: Environmental or genetic disturbances of brain maturation may underlie the pathophysiology of adult-onset psychiatric disorders. We therefore followed the trajectories of brain structural and prefrontal cortex metabolite abnormalities from adolescence to adulthood in rats born to mothers exposed to the viral mimic polyriboinosinic-polyribocytidylic acid (poly-I:C) in pregnancy. A recent high-profile clinical trial suggested omega-3 fatty acid supplementation may be a beneficial early intervention in those at high-risk for psychosis (Amminger et al. Arch Gen Psychiatry. 2010; 67(2): 146–54). We therefore also investigated the effects of omega-3 fatty acid supplementation in a placebo-controlled study on brain abnormalities in poly-I:C and saline exposed offspring. Methods: Sprague-Dawley rat dams were exposed to poly-I:C (4 mg/kg, i.v. n=5) or 0.9% saline, (i.v; n=5) on gestational day (GD) 15. Male offspring from poly-I:C (n=10 from 5 litters) and saline exposed (n=10 from 5 litters) underwent longitudinal in vivo sMRI and 1H-MRS with a voxel placed in the prefrontal cortex, at post-natal day (PND) 50, 100 and 180 (Vernon et al., Eur. Neuropsychopharm. 2015; 25(12): 2210–20). Additional groups of male offspring exposed to either saline or poly-I:C in utero were administered omega-3 fatty acids (220 mg/kg p.o.; n=10 saline or poly-I:C each from 5 litters) or placebo (coconut oil p.o. n=10 saline or poly-I:C each from 5 litters) from PND35-70. Longitudinal in vivo MRI data were acquired from these animals at PND90 and 150. Longitudinal sMRI data were analysed using semi-automated segmentation and tensor based morphometry (TBM). Longitudinal 1H-MRS data were analysed using LC model. Both MRI and MRS data were analysed using linear mixed effect models with age, mIA and treatment as factors, respectively, using SPSS (v22, IBM) with alpha=0.05. Results: Exposure to mIA resulted in subtle age-dependent metabolic perturbations of the normal maturing prefrontal cortex, including decreased glutathione (p<0.05; d=1.08), taurine (p<0.01; d=1.35) and N-acetyl-aspartate (p<0.05; d=1.05) at PND180. Longitudinal sMRI analyses revealed that exposure to mIA decreased hippocampus volume at PND90 and 180 (p<0.05; d=1.07). TBM revealed widespread grey and white matter alterations across brain maturation (q=0.1 FDR corrected). Treatment with Omega-3 fatty acids partially prevented hippocampal volume decreases at PND90, but this was not sustained at PND150. Conclusion: Prenatal exposure to mIA interferes with postnatal brain maturation. This may be partially ameliorated, but not prevented, by early intervention with Omega-3 fatty acid supplementation. Financial support: This work was supported by funding from the Medical Research Council (Grant ID: G0701748 and G1002198).

S31 FROM INFLAMMATION TO PSYCHOSIS: A JOURNEY THROUGH THE BRAIN Dazzan P, Dept of Psychosis Studies, Inst of Psychiatry, Psychology and Neuroscience, PO Box 40, De Crespigny Park, London SE5 8AF [email protected] Introduction: There is increasing evidence that psychosis is associated with an activation of the immune system. In fact, patients with psychosis, including in the early illness stages, show an increase in circulating levels of inflammatory markers such as cytokynes (IL-1, IL-6 for example) and C Reactive protein. However, it remains unclear what is the relationship between these peripheral markers of inflammation and the brain structural changes frequently observed in psychosis. Methods: We have evaluated brain structure (volume and myelin content) and perfusion using MRI and inflammatory markers in 3 large cohorts of patients at their first episode of psychosis or at risk of psychosis (n=180), and in healthy controls (n=110). Results: We have found that elevated levels of these inflammatory markers are associated with ABSTRACTS A17 poor short-term outcomes such as prediction of treatment response (all p<.05). In addition, we found that levels of these markers of peripheral inflammation are associated with brain structural alterations in both grey matter volumes and myelin content, and with changes in hippocampal brain perfusion (all p<.05 corrected). Conclusions: Our findings suggest that inflammatory biomarkers at the onset of psychosis should be considered as possible predictors of treatment response, as well as potential targets for the development of novel therapeutic agents. This work was supported by the Wellcome Trust, NARSAD, and the National Inst for Health Research (NIHR) Mental Health Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London, UK.

S32 IMMUNE INVOLVEMENT IN THE PATHOGENESIS OF SCHIZOPHRENIA: A META-ANALYSIS ON POST- MORTEM BRAIN STUDIES van Kesteren CFMG, Psychiatry, Univ Medical Center Utrecht, Heidelberglaan 100 3584, CX Utrecht The Netherlands [email protected] Gremmels H(1), deWitte LD(2), Kahn RS(2), Sommer IEC(2), vanGool AR(3), Hol EM(4), Falkai PG(5) (1) 2Dept of Nephrology and Hypertension, Uppsalalaan 8, 3584 CT Utrecht, the Netherlands; (2) As presenting author; (3) Dept of Psychiatry, Yulius mental health organization, Dennenhout 1, 2994 GC Barendrecht, the Netherlands; (4) Dept of Translational Neuroscience, Universiteitsweg 100, 3584 CG Utrecht, The Netherlands; (5) Nußbaumstr. 7, D-80336 Munich, Germany Although the precise pathogenesis of schizophrenia is unknown, genetic, biomarker and imaging studies suggest involvement of the immune system. In this study we performed a systematic review and meta-analysis of studies investigating factors related to the immune system in post-mortem brains of schizophrenia patients and healthy controls. Forty studies were included, reporting on 754 patients and 725 controls. We divided these studies into those investigating histological alterations of cellular composition and those assessing molecular parameters; meta-analyses were performed on both categories. Our pooled estimate on cellular level showed a significant increase in the amount of microglia (p = 0.0028) in the brains of schizophrenia patients compared to controls, albeit with substantial heterogeneity between studies. Meta-regression on brain regions demonstrated this increase was most pronounced in the temporal cortex. Numbers of macroglia (astrocytes or oligodendrocytes) did not differ significantly between schizophrenic patients and healthy controls. The results of post-mortem histology are reflected in gene expression, where we observed an overall increase in expression of pro-inflammatory genes on transcript and protein level. The results of this meta-analysis strengthen the hypothesis that neuro- inflammation plays a role in the pathogenesis of schizophrenia, although further research is needed to substantiate differences between brain regions and to further specify alterations in cell phenotypes and gene expression. Our findings may provide direction for new prevention- and treatment strategies, like immune-modulating drugs, stress or sleep modulation, food supplements or probiotics. References: Stefansson H, Ophoff RA, Steinberg S, Andreassen OA, Cichon S, Rujescu D et al. Common variants conferring risk of schizophrenia. Nature 2009; 460: 744–747 Ripke S, O’Dushlaine C, Chambert K, Moran JL, Kähler AK, Akterin S et al. Genome-wide association analysis identifies 13 new risk loci for schizophrenia. Nat Genet 2013; 45: 1150–1159 Andreassen OA, Harbo HF, Wang Y, Thompson WK, Schork AJ, Mattingsdal M et al. Genetic pleiotropy between multiple sclerosis and schizophrenia but not bipolar disorder: differential involvement of immune-related gene loci. Molecular Psychiatry 2014. doi:10.1038/mp.2013.195. Benros ME, Nielsen PR, Nordentoft M, Eaton WW, Dalton SO, Mortensen PB. Autoimmune diseases and severe infections as risk factors for schizophrenia: a 30-year population-based register study. Am J Psychiatry 2011; 168: 1303–1310 Brown AS, Derkits EJ. Prenatal infection and schizophrenia: a review of epidemiologic and translational studies. Am J Psychiatry 2010; 167: 261–280 van Berckel BN, Bossong MG, Boellaard R, Kloet R, Schuitemaker A, Caspers E et al. Microglia activation in recent-onset schizophrenia: a quantitative (R)-[11C]PK11195 positron emission tomography study. Biological Psychiatry 2008; 64: 820– 822 Doorduin J, de Vries EFJ, Willemsen ATM, de Groot JC, Dierckx RA, Klein HC. Neuroinflammation in schizophrenia-related psychosis: a PET study. J Nucl Med 2009; 50: 1801–1807 Miller BJ, Buckley P, Seabolt W, Mellor A, Kirkpatrick B. Meta-analysis of cytokine alterations in schizophrenia: clinical status and antipsychotic effects. Biological Psychiatry 2011; 70: 663–671 Khandaker GM, Pearson RM, Zammit S, Lewis A18 ABSTRACTS

G, Jones PB. Association of serum interleukin 6 and C-reactive protein in childhood with depression and psychosis in young adult life: a population-based longitudinal study. JAMA Psychiatry 2014; 71: 1121–1128 Wium-Andersen MK, Ørsted DD, Nordestgaard BG. Elevated C-reactive protein associated with late- and very-late-onset schizophrenia in the general population: a prospective study. Schizophr Bull 2014; 40: 1117– 1127 Fan X, Pristach C, Liu EY, Freudenreich O, Henderson DC, Goff DC. Elevated serum levels of C-reactive protein are associated with more severe psychopathology in a subgroup of patients with schizophrenia. Psychiatry Res 2007; 149: 267–271.

S33 CNS BIOMARKERS IN ANTI-OBESITY DRUG DEVELOPMENT Ziauddeen H, Psychiatry, Univ of Cambridge, Box 189, Herchel Smith Building, West Forvie Site, Robinson Way, Cambridge CB2 0SZ [email protected] With anti-obesity agents, weight loss may emerge from an array of metabolic, cognitive and behavioural changes that translate into weight change over time. In early drug development, characterising these changes using cognitive and behavioural measures, and functional neuroimaging, may actually be more informative than simply measuring weight loss. Biomarkers for these mechanisms can be used to determine whether potential compounds are worth developing further by providing proof of mechanistic action and detecting early signs of neuropsychiatric adverse effects. In this talk I will review potential biomarkers for effects on metabolism and satiety, hedonics and motivation, and eating behaviour; and also review biomarkers for early detection of neuropsychiatric adverse effects. While this approach has not been the norm in this field of drug development, it offers a more mechanistically informed approach to selecting and progressing potential compounds to the necessary large scale studies that are still pre- requisite for regulatory approval, and are likely to be so for some time. Funding: HZ was funded by the Wellcome Trust- GlaxoSmithKline funded Translational Medicine and Therapeutics PhD programme and is currently supported by the Bernard Wolfe Health Neuroscience Fund. He has also received consultancy fees from GlaxoSmithKline.

S34 STRATIFICATION OF RESPONDERS TO THE SATIATING EFFECTS OF SEROTONERGIC THERAPIES FOR OBESITY USING FMRI Higgs S, School of Psychology, Univ of Birmingham, Edgbaston, Birmingham B15 2TT [email protected] Dourish CT(1), Thomas JM(2) (1) P1vital, Howbery Park, Wallingford, Oxfordshire; (2) Psychology, Aston Univ There are considerable individual differences in the response to drugs that are used to treat obesity. Some patients experience substantial reductions in appetite and body weight whereas others show little or no response. Greater understanding of these individual differences in responding will facilitate the development of stratified and more effective drug treatments. We have developed an experimental medicine model to study the effects of anti-obesity drugs and have validated this model using a 5-HT2C receptor agonist (mCPP), which has a similar mechanism of action to the recently licensed anti-obesity drug Lorcaserin. The model combines behavioural analysis of the microstructure of eating with assessment of neuronal responses to food stimuli using fMRI. To assess the potential for psychiatric side effects, emotional and cognitive responses are also measured using the P1vital® Oxford Emotional Test Battery (ETB). We have reported that mCPP decreases food intake and increases satiety in healthy volunteers and these effects on appetite are more pronounced for a highly palatable snack eaten in the absence of hunger. The effects on appetite and food intake occur in the absence of negative effects on emotional responding. In addition, mCPP decreases reward-related neural responses to the presentation of food pictures. These data suggest that mCPP may decrease food intake via a reduction in food reward that occurs as food is consumed. Furthermore, stratification of participants according to whether or not they decrease their food intake in response to mCPP revealed differences in the baseline neural responses to food pictures and response to drug, particularly in reward circuitry. It may be possible in future clinical trials of novel serotonergic compounds and for prescribing of approved drugs such as the 5-HT2C receptor ABSTRACTS A19 agonist lorcaserin to identify individuals who are more likely to respond to certain types of medication. Funding: BBSRC CASE studentship in partnership with P1vital (grant number BB/G016739/1) and the Univ of Birmingham.

S35 EATING DISORDERS: INSIGHTS FROM IMAGING AND BEHAVIORAL APPROACHES TO TREATMENT Stice E, Univ of Texas at Austin, Dept of Psychology, Texas 78712 USA [email protected] Introduction: Imaging research suggests that individuals with eating disorders show elevated valuation of and attentional bias for the thin beauty ideal and high-calorie foods, and that these neural factors predict future symptom escalation, prompting the theory that eating disorders are maintained by excessive (1) valuation of the thin ideal, which prompts caloric restriction that contributes to emaciation and risk for binge eating and compensatory weight loss behaviors, and (2) valuation of high-calorie binge foods, which maintains binge eating. We created an 8-session Counter Attitudinal Therapy (CAT), wherein women with eating disorders complete verbal, written, and behavioral activities in which they discuss costs of pursuing the thin ideal and engaging in eating disordered behaviors (e.g., fasting, binge eating), which putatively creates dissonance that reduces valuation of the thin ideal and binge foods. Method: We conducted two randomized trials with individuals with the full range of eating disorders (N=72 and 84). Results: In trial 1 CAT versus usual care participants showed greater reductions in outcomes (p<.001), producing large effects for thin-ideal internalization (d=.79), body dissatisfaction (d=1.14), and blinded interview-assessed eating disorder symptoms (d=.95), and medium effects for dissonance regarding perpetuating the thin ideal (d=.65) and negative affect (d=.55). In trial 2 CAT versus usual care group treatment participants showed greater pre-post reductions in outcomes (p=.028) with the following outcomes: thin-ideal internalization (d=.70), dissonance regarding perpetuating the thin ideal (d=.32), body dissatisfaction (d=.30), and blinded interview-assessed ED symptoms (d=.53). CAT also reduced valuation of thin models and high-calorie binge foods and the dissonance-inducing activities regarding pursuit of the thin ideal produced a reduction in fMRI-assessed reward region (caudate) response to thin models. Conclusions: The effect size for greater symptom reduction relative to usual care controls is comparable to those produced by much more intensive 20-session individual therapies. The effect size for greater symptom reduction for CAT versus standardized usual care is larger than the parallel effect from any published trial that compared a 20-session treatment (e.g., CBT) to alternative treatments. Because CAT is brief, easy to implement because of the scripted manual, and inexpensive, it should be possible to implement it broadly, which would address a major public health problem because extant evidence-based treatments have not been implemented broadly.

S36 THE DEVELOPMENT AND USE OF NOVEL ANIMAL MODELS IN THE SEARCH FOR DRUGS TO TREAT BINGE-EATING DISORDER Heal DJ, RenaSci Ltd, BioCity, Nottingham NG1 1GF [email protected] Goddard S(1), Brammer RJ(1), Vickers SP(1) (1) RenaSci Ltd, BioCity, Nottingham NG1 1GF Binge-eating disorder (BED) was classified as a discrete psychiatric condition in DSM-5. BED consists of episodes of loss of control with compulsive and perseverative bouts of excessive food consumption. Although BED is often associated with obesity, 17–30% of subjects have body weights within the normal range, and ~60% in the normal weight/overweight categories. Various anti-obesity drugs have been clinically evaluated to treat BED, but have not proved particularly effective. The probable explanation is they reduce appetite/increase satiety, which delivers weight-loss, but do not tackle the psychopathology underpinning the disorder. Lisdexamfetamine (LDX), a d-amphetamine prodrug, is the first approved treatment for BED. When designing/validating animal models of binge-eating (BE), it is not enough to measure palatable food consumption, it is also essential to recreate the psychopathology of BED. To achieve this objective, we set four criteria:: 1.Consumption of palatable food must be excessive compared with amount required to induce obesity; 2.There should be compulsive and perseverative components to BE behaviour; 3.There should be an impulsive component to BE behaviour; 4.BE behaviour should produce A20 ABSTRACTS

CNS neurochemical abnormalities The model we developed is based on an unpredictable, limited access to palatable food BE paradigm (Corwin, 2004, Appetite 42: 139–142). To induce BE, adult, female rats were given continuous access to chow and water together with intermittent 2hr access to chocolate over 28 days. These rats display intense hyperphagia during the chocolate binges, 2–3x higher intake of palatable food than obese rats. In this model, LDX selectively reduced chocolate consumption, whereas the anti-obesity drug, sibutramine, non-selectively decreased both chocolate and chow intake. In a novel food reward/ punished conflict test, BE rats showed marked compulsive and perseverative responding when given access to chocolate. Acute LDX administration abolished these aberrant behaviours. In a delay-discounting task to obtain chocolate-flavoured pellets, BE rats showed intolerance of delayed reward and enhanced discounting. This impulsive behaviour was abolished by LDX. Striatal D1 receptors were decreased without a change to the size of the dopamine neuronal pool or dopamine turnover rate indicating binge eating is associated with decreased D1 signalling. The µ- receptor number was increased in striatum, but not PFC. Both neurotransmitter systems are closely linked with reward. Our results with BE rats are consistent with the psychopathology of BED and the clinical actions of LDX (McElroy et al, 2015, Eur Eat Disord Rev, Epub; McElroy et al, 2016, Neuropsychopharmacology 41: 1251–1260).

A01 CORRELATION BETWEEN PREFRONTAL GREY MATTER VOLUME AND STRIATAL DOPAMINE SYNTHESIS CAPACITY: A [18F]-DOPA PET AND VBM STUDY D’Ambrosio E, Dept of Psychosis Studies IoPPN, King’s College London, London, SE5 8AF [email protected] Jauhar S(1), Bonoldi I(1), Hathway P(1), Pepper F(1), Kapur S(1), Kotoula V(2), Veronese M(3), Turkheimer F(3), Rogdaki M(4), Howes OD(4) (1) As presenting author; (2) Centre for Neuroimaging Sciences, King’s College London, London, SE5 8AF; (3) Dept of Neuroimaging, IoPPN, King’s College London, London, SE5 8AF; (4) Dept of Psychosis Studies, IoPPN, King’s College London, London, SE5 8AF; Psychiatric Imaging Group, Imperial College, MRC Clinical Sciences Centre, Hammersmith Hospital, London, W12 0NN Introduction: Reduced grey matter volume in frontal cortex and increased striatal dopamine synthesis capacity are considered as pathophysiological features of schizophrenia (Chan et al, 2011, Schizophr Bull 37:177–188; Howes et al, 2015, J Psychopharmacol 29:97–115). Animal studies have shown that lesions in prefrontal cortex result in increased striatal dopamine function (Pycock et al, 1980, Nature 286:74–76). We therefore wished to examine the relationship between grey matter volume and dopamine synthesis capacity in healthy adults. We hypothesized that grey matter volume in frontal cortex would be inversely correlated with striatal dopamine synthesis capacity. Methods: 19 healthy controls (12 males, 7 females; mean age 23.8 ± 3.4 years) underwent 18F-DOPA PET scans to measure dopamine synthesis capacity (indexed as the influx rate constant Kicer) and structural MRI (196 T1-weighted images acquired with a 3T scanner). Structural images were normalized using DARTEL, modulated to correct for warping effect and spatially smoothed. A multiple regression was performed in SPM12 to test the relation between grey matter volume and striatal dopamine synthesis capacity. Effects were controlled for age, gender and total intracranial volume. A whole brain correction for multiple comparisons was performed (cluster level correction, p-value < 0.05). Results: We found a significant negative correlation between striatal dopamine synthesis capacity and grey matter volume in medial prefrontal cortex (BA 9: x= 3, y= 48, z= 20; k= 600, T= 5.93, Z= 4.13, pFWE corrected= 0.03). Conclusions: The hypothesis that prefrontal grey matter volume is negatively correlated with striatal dopamine synthesis capacity was confirmed. The localisation of the structural MRI findings to the medial prefrontal cortex reflects the evidences of an anatomical connection between this area and the striatum. The finding that reduced grey matter volume was associated with increased dopamine synthesis capacity is coherent with the theory that an over-activity of the subcortical dopaminergic system could be determined by modulation of the medial prefrontal cortex (Flores et al, 1996, J Neurosci 16:7366–7375). Exploring the relationship between prefrontal cortical volume and dopaminergic striatal activity in patients with schizophrenia could be crucial in clarifying the pathophysiology of this illness. Sources of financial sponsorship: Medical Research Council (UK). ABSTRACTS A21

A02 DOES DOPAMINE SYNTHESIS CAPACITY PREDICT ANTIPSYCHOTIC TREATMENT RESPONSE IN FIRST EPISODE PSYCHOSIS; AN 18 F-DOPA PET STUDY Jauhar S, Psychosis Studies King’s College, London, PO63 Denmark Hill Campus De Crespigny Park London SE5 8AF [email protected] Veronese M(1), Rogdaki M(2), Bonoldi I(3), Stone J(3), Egerton A(3), Turkheimer F(3), McGuire P(3), Kapur S(3), Howes OD(3), Hathway P(4), Pepper F(4) (1) Centre for Neuroimaging Sciences, KCL; (2) Imperial College; (3) KCL; (4) Psychosis Studies Introduction: A significant number of people with schizophrenia show a poor response to antipsychotic treatment from illness onset. Better understanding of the neurobiology of treatment response/non- response is needed, to inform alternative treatments. Altered dopamine synthesis capacity has been linked to treatment response in chronic patients (Demjaha ete al, 2012, Am J Psych, 169;1203–10) . However, to date, no study has examined the relationship between dopamine synthesis capacity and subsequent response to treatment in first episode patients. We sought to ascertain whether baseline dopamine synthesis capacity would be associated with clinical response and functional outcome in patients with first episode psychosis, and to compare dopamine synthesis capacity at baseline in treatment responders, non-responders and control subjects. Methods: Prospective case-control study. 25 people with first episode psychosis and 10 controls were scanned with 18F-DOPA PET, assessing dopamine synthesis capacity (Kicer). Patients were subsequently assessed with clinical measures (Positive and Negative Syndrome Scale and General Assessment of Functioning) at baseline before starting antipsychotic treatment and reassessed after at least 6 weeks of antipsychotic treatment. They were then classified as responders and non-responders, based on change in the Clinical Global Impression scale. Results: There was a significant positive correlation between baseline Kicer and percentage change in PANSS positive symptoms (pearson’s correlation r=0.65, p<0.01, PANSS negative symptom change (pearson’s r=0.50, p=0.03, PANSS total symptoms (pearson’s r=0.52, p=0.03. and GAF improvement (r=0.49, p=0.04)). Treatment responders had significantly elevated Kicer compared to non- responders and control subjects (effect size, cohen’s d=1.3 p=0.01). There was no significant difference in Kicer between treatment non-responders and controls (p=0.72). Conclusions: In people with first episode psychosis, dopamine synthesis capacity predicts treatment response to antipsychotic medication. There is the suggestion that treatment resistance may have a different underlying neurobiology. This work was funded by an MRC Project Grant to Dr OD Howes.

A03 ADENOSINE A2A RECEPTOR IN SCHIZOPHRENIA: A POSITRON EMISSION TOMOGRAPHY STUDY USING [11C]SCH442416 Marques TR, Clinical Sciences Centre, Faculty of Medicine Imperial College London, Hammersmith Hospital, Du Cane Road London, W12 0NN [email protected] Natesan S(1), Howes OD(1), Kapur S(1), Gunn RN(2), Searle GE(3), Rabiner EA(3) (1) Dept of Psychosis Studies, Inst of Psychiatry, Psychology and Neuroscience, King’s College London, London, UK; (2) Division of Brain Sciences, Dept of Medicine, Imperial College London, London, UK ; (3) Imanova Ltd., Centre for Imaging Sciences, Hammersmith Hospital, London, UK Introduction: Adenosine A1 and A2A receptor (A1R and A2AR) subtypes are widely expressed in the CNS and extracellular adenosine acts as the endogenous agonist at these receptors. A2AR is mainly localized in the striatum and nucleus accumbens, where they have been found to be co-localized with dopamine D2 receptors in striatal neurons, and they mutually antagonise their respective responses (Fredholm BB et al., 2005). The distribution of A2AR (activation of which increases cAMP) and its co- localization with dopamine D2 receptors (activation of which decreases cAMP) in striatal neurons, has led to the suggestion that A2AR could play a key role in modulating dopamine signalling in the human brain (Ferry S et al., 1991; Pollack & Fink, 1995, Ferré S, 1997) Post-mortem brain studies have shown a significant increase (~35 to 55%) in A2AR specific binding in the A2AR expression in the caudate and putamen of post-mortem brains of schizophrenia patients Kurumaji & Toru, 1998). However, there are A22 ABSTRACTS no A2AR positron emission tomography (PET) imaging studies in patients with schizophrenia. In this study we have assessed for the first time the availability of A2AR in vivo in patients with Schizophrenia, using [11C]SCH442416 positron emission tomography (PET). Methods: We compared PDE10A A2A availability in the brains of 132 patients with chronic schizophrenia on antipsychotic treatment with that of 12 healthy controls. Parametric images of [11C]SCH442416 11C]IMA107 binding potential (BPND) were generated from the dynamic [11C]SCH442416 [11C]IMA107 data using the simplified reference tissue model with the cerebellum as the reference tissue for non-displaceable binding. Results: There was a significant difference in the [11C]SCH442416 [11C]IMA107 volume of distribution (VT) between schizophrenia patients and healthy controls in some of the brain regions studied: putamen (P=.040; -17%), globus pallidus (P=.067; -14%), caudate (P=0.35; -8%), accumbens (P=.044; -16%). No significant correlation was found between [11C]SCH442416 [11C]IMA107 BPND and severity of psychotic symptoms or exposure to antipsychotics. Conclusions: Patients with schizophrenia on antipsychotic medication have a significant reduction in A2AR volume of distribuition when compared with healthy controls in brain regions thought to be involved in the pathophysiology of this disorder. However, there are significant limitations regarding the study of patients chronically exposed to antipsychotic medication, as well as the use of a with poor kinetic profile.

A04 INVESTIGATING THE ASSOCIATION BETWEEN INFLAMMATION AND SYMPTOMS OF DEPRESSION IN FIRST EPISODE PSYCHOSIS Worrell C, Psychological Medicince Inst of Psychiatry, Psychology and Neuroscience, The Maurice Wohl Clinical Neuroscience Inst, G.33.65, 8 Cutcombe Road, KCL, London, SE5 9RT [email protected] Russo M(1), Falcone A(1), Bonaccorso S(1), Marques TR(1), Di Forti M(1), Gaughran F(1), Pariante CM(1), Murray R(1), Dazzan P(1), Mondelli V(1) (1) Inst of Psychiatry, Psychology and Neuroscience Over the past decade, increased levels of inflammatory biomarkers have been widely reported both in depression and psychosis. However, despite there being a prevalence of research in this area, it still remains unclear whether such increased inflammation at the onset of psychosis is associated specifically with symptoms of depression. The study aimed to investigate the relationship between levels of C-reactive protein (CRP) as a biomarker for inflammation, and depression, which was measured with the Calgary Depression Scale for Schizophrenia (CDS). Investigation was done using a sample of n=187 patients, all with first episode psychosis. The sample was made up of 126 males and 61 females between the ages of 18 to 60, with a mean age of 29±9 years; all patients were recruited from the South London and Maudsley NHS Foundation and Trust. We ran correlational analyses to test the association between CRP levels and patients’ total scores on the CDS, and we additionally investigated the association between CRP levels and the severity of each individual CDS item using one way between-subjects ANOVA followed by Bonferroni post-hoc analyses. Our findings show a trend for a positive correlation between CRP levels and total CDS scores (r=.197, p = .026), indicating the basis for a relationship between inflammation and the severity of depression within first episode psychosis. When looking at individual items of the CDS, we also found that there were significantly higher CRP levels in patients with higher scores in the following items: guilty ideas of reference (Mean ± SEM, absent: 2.8±0.9mg/dl; mild: 4.1±1.6mg/dl; moderate: 2.0±1.0mg/dl; severe: 23.0±19.4mg/dl; (f(3,124)=5.420, p = .002), and suicide (Mean ± SEM, absent: 2.8±0.8mg/dl; mild:2.7±0.8mg/dl; moderate: 16.1±13.1mg/dl; severe: 7.3±4.7mg/dl; (f(3,123)=3.413, p = .02). Similarly, further trends were identified with elevated CRP levels in participants that had rated their symptoms as severe on certain items of the CDS such as hopelessness and pathological guilt. These results serve to demonstrate an association between increased levels of inflammation and the severity of depressive symptoms in first episode psychosis; making implications for the role of increased inflammation at onset of psychosis being associated with symptoms of depression. This study was funded by the Maudsley Charitable Fund and the Dept of Health.

A05 THIS ABSTRACT HAS BEEN WITHDRAWN ABSTRACTS A23

A06 THE EFFECT OF SODIUM NITROPRUSSIDE ON PSYCHOTIC SYMPTOMS AND SPATIAL WORKING MEMORY IN PATIENTS WITH SCHIZOPHRENIA – A RANDOMIZED, DOUBLE-BLIND, PLACEBO- CONTROLLED TRIAL Stone JM, Inst of Psychiatry, Psychology and Neuroscience King’s College London, Centre for Neuroimaging Sciences, De Crespigny Park, SE5 8AF [email protected] Morrison PD(1), Gao F(1), Koychev I(1), Reilly T(1), Kolanko M(1), Mohammadinasab A(1), Kapur S(1), McGuire P(1) (1) King’s College London Background: Sodium nitroprusside (SNP) has been reported to rapidly reduce psychotic symptoms in patients with schizophrenia. This has the potential to revolutionize treatment for schizophrenia. In this study, we tested the hypothesis that SNP leads to a reduction in psychotic symptoms and an improvement in spatial working memory performance in patients with schizophrenia. Methods: This was a single- center, randomized, double-blind, placebo-controlled trial performed from 27 August 2014 to February 10 2016. Twenty patients with schizophrenia aged 18–60 with a diagnosis of schizophrenia or schizoaffective disorder were recruited from psychiatric out-patient clinics in the South London and Maudsley NHS Trust, London, UK. Baseline symptoms were measured using the Positive and Negative Syndrome Scale (PANSS) and the 18 item Brief Psychiatric Rating Scale (BPRS), and spatial working memory was assessed using the CANTAB computerized test. Particpants received either an infusion of sodium nitroprusside (0.5 micrograms per kilogram per hour for 4 hours) or placebo and were re-assessed for symptoms and spatial working memory performance immediately after the infusion, and 4 weeks later. Results: Sodium nitroprusside did not lead to any reduction in psychotic symptoms or improvement in spatial working memory performance compared to placebo. Conclusions: Although this study was negative, it is possible that the beneficial effects of sodium nitroprusside may occur in patients with a shorter history of illness, or with more acute exacerbation of symptoms.This work was supported by the NIHR Biomedical Research Centre for Mental Health at the South London and Maudsley NHS Foundation Trust and Inst of Psychiatry, King’s College London. It did not receive any external funding.

A07 A TSPO BLOCKING STUDY TO DETERMINE THE SPECIFIC BINDING OF [11C]-PBR28 IN PATIENTS WITH SCHIZOPHRENIA Reis Marques T, Clinical Sciences Centre Faculty of Medicine, Imperial College, Hammersmith Hospital Du Cane Road London, W12 0NN [email protected] Veronese M(1), Owen D(2), Gunn RN(2), Rabiner EA(3), Bloomfield P(4), Howes OD(5) (1) Center for Neuroimaging Science, Inst of Psychiatry, Psychology and Neuroscience, King’s College London, London, UK.; (2) Division of Brain Sciences, Dept of Medicine, Imperial College London, London, UK; (3) Imanova Ltd., Centre for Imaging Sciences, Hammersmith Hospital, London, UK ; (4) Imperial College London, London, UK; (5) Psychiatric Imaging Group, MRC Clinical Sciences Centre, Inst of Clinical Science, Imperial College London, London, UK INTRODUCTION Epidemiological, genetic and clinical evidence all indicates that inflammation plays a role in schizophrenia. When microglia are activated, they express high levels of the 18-kDa translocator protein (TSPO). TSPO can be measured in vivo with PET radiotracers, such as [11C]-PBR28, and an increase in [11C]-PBR28 binding ratio has been shown in patients with schizophrenia relative to healthy controls. The main outcome measure in studies using TSPO tracers is the total volume of distribution (Vt), creating some difficulties in the direct estimation of specific binding. Our research group has already shown that in controls a significant proportion of Vt is specific to TSPO, but this has never been address in schizophrenia. In this study we have used a TSPO agonist to block the binding of [11C]-PBR28 in order to estimate the TSPO specific binding in schizophrenia. METHODS Seven patients with schizophrenia were recruited from South London and Maudsley NHS Trust. All subjects were genotyped and were homozygote high-affinity binders (HABs). Subjects received a baseline bolus injection of [11C]-PBR28 followedy b a 90-minute emission scan. PET data were co-registered with whole brain structural images acquired with A24 ABSTRACTS a 3T MRI scanner. During the PET acquisition, arterial blood data were sampled via the radial artery. In a follow-up visit patients received a selective TSPO blocker (90mg of XBD173) followed 2 hours later by a repeat [11C]-PBR28 scan. Quantification of [11C]-PBR28 tissue distribution was performed using the two tissue compartmental model. RESULTS Following blockade with XBD173 all patients with schizophrenia showed a significant reduction in TSPO specific binding across different ainbr regions. Regional VT data for individual subjects were fitted to an occupancy plot. The occupancy plot assumes that VND of the radioligand is unchanged by the administration of the blocking drug, and that the fractional occupancy of the target by the blocking drug is the same across all ROI. Hence, the occupancy plot provides a measure of both target occupancy and the VND. By constraining the VND to be equal for all schizophrenia patients, the population VND was estimated to be 1.99. Using the VND we were also able to calculate the specific binding (BPND) for different ROIs. CONCLUSIONS To our knowledge, this is the first study to estimate the non-displaceable binding component (VND) of [11C]PBR28 in the brain of individuals with schizophrenia. Our results shows that in patients with schizophrenia a substantial component of [11C]PBR28 VT represents specific binding. This data supports the use of TSPO PET ligands to assess microglia activity in schizophrenia. Funding: Supported by a MRC grant (MC-A656-5QD30) and the NIHR BRC SLaM NHS Foundation Trust.

A08 EFFICACY OF LURASIDONE IN PATIENTS WITH SCHIZOPHRENIA WITH PROMINENT POSITIVE SYMPTOMS: A POOLED ANALYSIS OF SHORT-TERM PLACEBO-CONTROLLED STUDIES Potkin S, Psychiatry & Human Behavior School of Medicine Univ of California, Irvine, 5251 California Ave., Ste 240, Mail Code: 1680, Irvine, CA, 92617 [email protected] Tocco M(1), Mao Y(2), Cucchiaro J(2), Loebel A(2) (1) 84 Waterford Drive, Marlborough, MA 01752; (2) One Bridge Plaza North, Suite 510, Fort Lee, NJ 07024 INTRODUCTION: Acute worsening of psychosis in patients with schizophrenia is characterised primarily by the presence of active positive symptoms, which may increase the risk of behavioral disturbance and hospitalisation. This post hoc analysis evaluated the efficacy of lurasidone in patients with an acute exacerbation of schizophrenia with prominent positive symptoms. METHODS: Patient-level data were pooled from 5 similarly designed, multiregional, randomised, double-blind, placebo-controlled, 6-week studies of fixed-dose lurasidone (37, 74, 111, and 148 mg/d) conducted in adult patients with an acute exacerbation of schizophrenia. Patients with prominent positive symptoms were defined as those with baseline Positive and Negative Syndrome Scale (PANSS) positive subscale score > baseline PANSS negative subscale score. Changes from baseline in PANSS total and positive subscale scores were evaluated using mixed-model repeated- measures analysis (MMRM). RESULTS: This analysis included 919 patients with prominent positive symptoms (mean age, 38.5 years; male, 72.3%) and 613 patients without prominent positive symptoms (mean age, 38.3 years; male, 74.1%). Study discontinuation rates were 39.5% for lurasidone and 48.7% for placebo in patients with prominent positive symptoms, and 29.5% for lurasidone and 36.2% for placebo in patients without prominent positive symptoms. For change from baseline to week 6 in PANSS total score (MMRM), effect sizes for the lurasidone 37, 74, 111, and 148 mg/d dose groups were 0.51, 0.65, 0.44, and 1.09, respectively, for patients with prominent positive symptoms (all P<0.001) and 0.29, 0.46, 0.55, and 0.67, respectively, for patients without prominent positive symptoms (P<0.05 for 37 mg/d, all other P<0.001). For change from baseline to week 6 in PANSS positive subscale score (MMRM), effect sizes for the lurasidone 37, 74, 111, and 148 mg/d dose groups were 0.42, 0.63, 0.51, and 1.31, respectively, for patients with prominent positive symptoms (P<0.01 for 37 mg/d, all other P<0.001) and 0.24, 0.55, 0.56, and 0.65, respectively, for patients without prominent positive symptoms (P=0.082 for 37 mg/d, all other P<0.001). CONCLUSIONS: In this post hoc analysis of adult patients with schizophrenia, lurasidone therapy was associated with medium to large treatment effect sizes. Larger effect sizes were observed in patients with prominent positive symptoms compared with patients without prominent positive symptoms. These results may inform the design of future clinical trials in schizophrenia. Supported by Sunovion Pharmaceuticals Inc. ABSTRACTS A25

A09 METABOLIC SYNDROME IN PATIENTS WITH SCHIZOPHRENIA TREATED WITH LURASIDONE OR OTHER ATYPICAL ANTIPSYCHOTIC AGENTS Newcomer J, Clinical Biomedical Science Florida Atlantic Univ Charles E. Schmidt College of Medicine, 777 Glades Road, Boca Raton, FL, 33431 [email protected] Tocco M(1), Pikalov A(2), Zheng H(2), Cucchiaro J(2), Loebel A(2) (1) 84 Waterford Drive, Marlborough, MA 01752; (2) One Bridge Plaza North, Suite 510, Fort Lee, NJ 07024 INTRODUCTION: Patients with schizophrenia are at high risk for developing metabolic syndrome, which may be exacerbated by treatment with antipsychotic agents. This post hoc analysis evaluated metabolic syndrome occurrence after short-term and long-term treatment with lurasidone. METHODS: Metabolic syndrome was defined by US National Cholesterol Education Program-Adult Treatment Panel III criteria (without using drug treatment criteria). Patient-level data were pooled from three phase 3, randomised, double-blind, placebo-controlled, 6-week studies of fixed-dose lurasidone (37, 74, 111, or 148 mg/d) in the treatment of adult patients with acute schizophrenia. Two studies included an active comparator (olanzapine 15 mg/d or quetiapine XR 600 mg/d). Data from 2 long-term, active-controlled (quetiapine XR or risperidone) studies of lurasidone in adult patients with schizophrenia were analysed separately. In the long-term, quetiapine XR–controlled study, patients completing a 6-week, double-blind, placebo-controlled, fixed-dose trial of lurasidone (74 mg/d or 148 mg/d) or quetiapine XR (600 mg/d) continued on double- blind, flexibly dosed lurasidone (37–148 mg/d) or quetiapine XR (200–800 mg/d) for up to 12 months. In the long-term risperidone-controlled study, patients received double-blind, flexibly dosed lurasidone (37–111 mg/d) or risperidone (2–6 mg/d) for up to 12 months. RESULTS: In patients without metabolic syndrome at baseline in the pooled analysis of short-term studies, the proportion who met criteria for metabolic syndrome after 6 weeks of treatment (LOCF) was 10.7% (26/242) of patients receiving placebo, 9.4% (52/551) of lurasidone-treated patients, 25.3% (21/83) of olanzapine-treated patients, and 30.5% (25/82) of quetiapine XR-treated patients (P<0.05 for olanzapine vs placebo; P<0.01 for quetiapine XR vs placebo). Among patients without metabolic syndrome at baseline in the long-term, quetiapine XR–controlled study, 2.2% (2/92) of patients receiving lurasidone and 10.0% (3/30) of patients receiving quetiapine XR developed metabolic syndrome at month 12 (observed cases [OC]); risk for developing metabolic syndrome was not significantly different between treatment groups (odds ratio=0.290; 95% CI, 0.042-1.994; P=NS). Of patients without metabolic syndrome at baseline in the long-term, risperidone-controlled study, 10.3% (12/117) of patients receiving lurasidone and 23.2% (16/69) of patients receiving risperidone developed metabolic syndrome at month 12 (OC); risk for developing metabolic syndrome was significantly lower for lurasidone-treated versus risperidone-treated patients (odds ratio=0.379; 95% CI, 0.167-0.858; P=0.02). CONCLUSIONS: Patients with schizophrenia were less likely to develop metabolic syndrome after short- term and long-term treatment with lurasidone compared with other medications tested. Clinicians should consider potential medication effects in the management of risk for metabolic syndrome. Sponsored by Sunovion Pharmaceuticals Inc.

A10 DECREASED GLUTAMATERGIC RESPONSIVITY IN TREATMENT RESISTANT SCHIZOPHRENIA: PRELIMINARY 1H-MRS FINDINGS EXAMINING ANTERIOR CINGULATE CORTEX GLUTAMATE FOLLOWING A CHALLENGE Pillinger TE, Psychosis Studies Inst of Psychiatry, King’s College London, 16 De Crespigny Park London SE5 8AF [email protected] Rogdaki M(1), Howes OD(1) (1) As presenting author Introduction: Treatment resistance in schizophrenia (TRS) refers to an inadequate response to at least two adequate trials of different antipsychotics. Current antipsychotics act by blocking D2 receptors (Howes et al., 2012 Arch Gen Psychiatry 69(8): 776–786). However, TRS patients show little response, despite high D2 receptor blockade (Kapur et al., 2000 Am J Psychiatry 157(4): 514–520; Meltzer 2013, Annu Rev Med 64: A26 ABSTRACTS

393–406). Imaging findings indicate that treatment resistant patients do not show the typical dopamine dysfunction seen in schizophrenia (Demjaha et al., 2012 Am J Psychiatry 169(11): 1203–1210), which may reflect the involvement of other . Glutamatergic dysfunction has been implicated in the pathophysiology of schizophrenia (Kantrowitz and Javitt 2010 Brain Res Bull 83(3–4): 108–121). Treatment resistant schizophrenia patients have demonstrated elevated anterior cingulate cortex glutamate compared to treatment responsive individuals and healthy volunteers (Demjaha et al., 2014 Biol Psychiatry 75(5): e11–13, Mouchlianitis et al., 2015, Schizophr Bull.). Glutamate elevation in treatment resistant schizophrenia is seen despite adequate antipsychotic therapy, thus a failure of responsivity in the glutamate system may underlie treatment resistance. Examining the responsivity of the glutamate system in TRS is possible with Riluzole, a drug that reduces extracellular glutamate concentration. Methods: We have acquired proton magnetic resonance spectroscopy (1H-MRS) at 3 Tesla from 5 patients with treatment resistant schizophrenia and 2 healthy controls, at baseline and after Riluzole administration. Riluzole was given for 2 days at a dose of 50mg twice daily. The patients all had a DSM-IV diagnosis of schizophrenia. Treatment resistance was defined as a score of >4 on at least one positive symptom sub-scale item of the PANSS and having a score of <59 on the Global Assessment of Functioning despite adequate dose and duration of antipsychotic treatment. The study is ongoing, with a plan to ultimately recruit 20 patients and 20 controls, matched for age, sex, and smoking status. Results: Glutamate levels were: levels pre-riluzole in patients mean[SD] = 1.28[0.18], levels post-riluzole in patients mean[SD] = 1.27[0.10]; levels pre-riluzole in controls mean[SD] = 1.5[0.25], levels post-riluzole in controls mean[SD] =1.35[0.03]. Conclusions: This preliminary evidence suggests reduced glutamate responsivity in treatment resistant patients relative to controls. Further data are required to test this. Financial Sponsorship: Research funded by the Medical Research Council.

A11 THE ROLE OF INFLAMMATORY MARKERS IN PREDICTING METABOLIC ABNORMALITIES IN FIRST EPISODE OF PSYCHOSIS AT ONE YEAR FOLLOW UP Nettis MA, Dept of Psychological Medicine, Inst of Psychiatry, Psychology and Neuroscience, King’s College London, London SE5 8AF, UK King’s College London, The Maurice Wohl Clinical Neuroscience Inst Cutcombe Road London SE5 9RT [email protected] Kolliakou A(1), Pariante CM(1), Mondelli V(1), O’Connor J(2), Bonaccorso S(2), Gardner-Sood P(2), Dazzan P(2), Gaughran F(3), David A(4), Murray RM(4) (1) As presenting author; (2) Dept of Psychosis Studies, IoPPN, King’s College London, London SE5 8AF; (3) Dept of Psychosis Studies, IoPPN, King’s College London, London; National Psychosis Service, South London and Maudsley NHS Foundation Trust, London; (4) IoPPN, King’s College London, De Crespigny Park, London Introduction: Metabolic abnormalities have frequently been reported in both pharmacologically treated and drug naïve patients with psychosis, and may lead to later increased mortality (Parsons B et al. 2009, Schizophr Res 110(1-3):103-10, Correll et al., 2014 JAMA Psychiatry Dec 1;71(12):1350-63). Such abnormalities are associated with inflammation in the general population and recent studies have shown increased inflammation in patients with First Episode of Psychosis (FEP) (Russell et al., 2015 Brain, Behavior, and Immunity 49, 25-29). Also of note, it has recently been suggested that C-reactive protein (CRP) levels could be a valuable marker of increased body mass index (Hepgul et al, 2012, Psychol. Med. 42 (9), 1893-1901) and risk of dyslipidemia in FEP (Russell et al., 2015). The aim of this study is to investigate the utility of high sensitivity CRP (hsCRP), as an inflammatory biomarker, in predicting metabolic abnormalities at 1-year follow-up in FEP. Methods: 32 FEP patients (mean age: 30.3 ± 10.1; M/F = 23/9; 25% White British) were included in this longitudinal study; social and clinical data, fasting blood samples, anthropometric measures, lipid profile and gluco-metabolic parameters (glycated hemoglobin (HbA1c), fasting glucose) were collected at baseline (T1) and at 1-year follow-up (T2). Correlation analyses were performed between hsCRP levels and BMI at baseline and anthropometric and metabolic parameters at 1-year follow-up. For those metabolic variables which were significantly associated with hsCRP, we run linear regression analyses using T1 hsCRP as predictor of T2 metabolic parameters. Results: T1 hsCRP levels were positively associated and proved to predict the following T2 parameters: HbA1c (p= 0.05, R2= 0.122), fasting glucose (0.024, R2= 0.159) and waist circumference (p=0.042, R2= 0.130). T1 BMI was correlated only with T2 waist circumference ABSTRACTS A27

(p<0.001). When running hierarchical regression including baseline BMI and hsCRP in the model for waist circumference, T1 BMI proved to be a stronger predictor of later waist circumference (p<0.002, R2=0.36) Conclusions: These data suggest that hsCRP levels at the onset of psychosis may predict later gluco-metabolic abnormalities. Conversely, our data show baseline BMI to be a better predictor than baseline hsCRP for later waist circumference. Further studies are needed to investigate the use of inflammatory markers to early detect patients in greater need of physical health intervention. This research has been supported by the National Institute for Health Research (NIHR) Mental Health Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London.

A12 GLUTAMATE IN PSYCHOSIS: A META-ANALYSIS OF PROTON MAGNETIC RESONANCE SPECTROSCOPY (1H-MRS) STUDIES Merritt K, Psychosis Studies Inst of Psychiatry, Psychology and Neuroscience, 5th floor 16 De Crespigny Park London SE5 8AF [email protected] Egerton A(1), Kempton MJ(1), Taylor MJ(1), McGuire PK(1) (1) As presenting author Introduction: Alterations in glutamatergic neurotransmission may be fundamental to the pathophysiology of schizophrenia and the glutamatergic system may be a target for new therapeutic interventions. To investigate the nature of brain glutamate alterations in schizophrenia we present a meta-analysis of glutamate proton magnetic resonance (1H-MRS) studies. Methods: Electronic databases were searched to identify journal articles reporting 1H-MRS glutamate, its metabolite glutamine or Glx (total glutamate+glutamine) in schizophrenia patients in comparison to healthy volunteers. Effect sizes were calculated for glutamate, glutamine and Glx in brain regions reported in at least 3 studies. Secondary analysis grouped studies into those examining different illness stages (high risk, first episode psychosis or chronic schizophrenia). Results: 59 eligible studies were identified. In schizophrenia, there were significant elevations in glutamate in the basal ganglia (P=0.01,g=0.63), glutamine in the thalamus (P=0.04,g=0.56), and Glx in the basal ganglia (P=0.01,g=0.39) and medial temporal lobe (P=0.002,g=0.32). No region showed a reduction in glutamate metabolites in schizophrenia. Sufficient studies were available to show that these glutamatergic elevations were present at different illness stages; glutamate metabolite elevations in the medial frontal cortex were specific to high risk and first episode patients, whereas Glx elevations in medial temporal lobe were specific to patients with chronic schizophrenia. Conclusions: Schizophrenia is associated with elevations in glutamatergic metabolites across several brain regions. This supports the hypothesis that schizophrenia is associated with excess glutamatergic neurotransmission in several limbic areas, and further indicates that compounds that reduce glutamatergic transmission may have therapeutic potential. This work is funded by the Medical Research Council.

A13 INFLUENCE OF CHILDHOOD SEXUAL ABUSE ON DIET IN FIRST EPISODE PSYCHOSIS Hastings C, Psychological Medicine Inst of Psychiatry Psychology and Neuroscience, King’s College London, The Maurice Wohl Clinical Neuroscience Inst, G.33.69 | Cutcombe Road | King’s College London | London SE5 9RT [email protected] Gardner-Sood P(1), Bonaccorso S(1), Fisher H(1), Morgan C(1), Gaughran F(1), Murray R(1), Pariante CM(1), Dazzan P(1), Mondelli V(1) (1) Inst of Psychiatry, Psychology and Neuroscience Introduction: Patients with schizophrenia have a poor diet, mainly characterized by a high intake of saturated fat and a low consumption of fibre and fruit (Di Pasquale et al., 2013 Journal of Psychiatry Research, 47(2), 197–207). Such diet is likely to contribute to the increased prevalence of metabolic abnormalities in psychosis. The factors influencing the dietary pattern in these patients have so far only been partly investigated, with most of the studies focussing on the role of antipsychotic treatment. Stress is known to influence food intake and can increase inflammation which may in turn lead to metabolic A28 ABSTRACTS abnormalities (Black, 2003, Brain Behavior and Immunity, 17, 350–364). Our previous work has shown an association between childhood sexual abuse and increased levels of high sensitive C-reactive protein (hsCRP) and increased Body Mass Index (BMI) in patients with first-episode psychosis (FEP) (Hepgul et al., 2012, Psychological Medicine, 42(9): 1893–1901). The aim of our study was to investigate whether childhood sexual abuse is associated with poor diet in FEP patients. Methods: We recruited 50 patients with FEP. We collected information about diet using the Dietary Instrument for Nutrition Education (DINE) questionnaire. Childhood trauma was assessed using the Childhood Experience of Care and Abuse (CECA) questionnaire. BMI and hsCRP levels were also measured in a subsample of patients and controls. Independent T-tests were used to analyse differences between abused and non-abused patients and Pearson correlations were performed to test association between DINE items and hsCRP and BMI. Results: Thirteen out of 50 patients had suffered childhood sexual abuse. Patients with experience of childhood sexual abuse had significantly lower intake in fibre (p= .026), bread (p= .027), cereal (p= .008) and spread (p= .010) when compared with patients without childhood sexual abuse. We found significantly higher BMI (mean±SEM: 1.62±0.14 vs. 0.46±0.08, p<.001) and a trend for higher hsCRP levels (5.01±2.12 vs. 1.01±0.20, p= .09) in patients who were sexually abused compared with those who were not abused. We did not find any significant correlation between fibre, bread, cereal or spread intake and BMI or hsCRP levels in the two groups. Conclusions: History of childhood sexual abuse was associated with lower bread, fibre, cereal and spread intake in FEP. Our preliminary findings did not show a correlation between dietary pattern and hsCRP or BMI; future studies would need to further investigate this association in larger samples. This work has been supported by a strategic award from the Wellcome Trust (Consortium of Neuroimmunology of Mood Disorders and Alzheimer’s Disease); the South London and Maudsley (SLaM) NHS Foundation Trust & Inst of Psychiatry NIHR Biomedical Research Centre (BRC) for Mental Health.

A14 EFFECTIVENESS OF PALIPERIDONE OR ARIPIPRAZOLE LONG-ACTING INJECTIONS IN CLINICAL PRACTICE Deslandes PN, Faculty of Life Sciences, Univ of South Wales, Pontypridd, CF37 1DL [email protected] Ward E(1), Norris K(1), Sewell RDE(1) (1) School of Pharmacy and Pharmaceutical Sciences, Cardiff Univ, CF10 3NB Introduction: Paliperidone and aripiprazole long-acting injections (PLAI and ALAI) were recommended as treatment options for schizophrenia in NHS Wales by the All Wales Medicines Strategy Group (AWMSG) in 2012 and 2014 respectively. The aim of the present study was to assess the effectiveness of PLAI or ALAI in a real world setting using treatment continuation at six months as an outcome measure. Method: All patients from a single health board in Wales who were initiated on PLAI or ALAI in the nine months following the publication of respective AWMSG advice were identified from pharmacy records. Demographic factors which may have influenced outcome were analysed, and reasons for treatment discontinuation noted. Previous treatment with clozapine was considered an indicator of treatment refractory illness. Data were collected by retrospective case-note review. Results: Twenty-eight patients received PLAI, of whom two were lost to follow-up, and 26 patients received ALAI. Therefore, six month outcome data were available for 52 patients. Seven (27%) PLAI and 12 (46%) ALAI treated patients received the medication for indications other than schizophrenia (p=0.25 Fisher’s Exact Test). Twenty- two (85%) and 19 (73%) patients remained on PLAI and ALAI respectively at six months (p=0.50, Fisher’s exact test). Treatment continuation with either drug was not associated with diagnosis, age at initiation, or inpatient/outpatient status on initiation. Three patients treated with PLAI and one patient treated with ALAI had previously received clozapine. The most common reason for discontinuation of PLAI was perceived lack of effect (n=3; 75%), whilst reasons for ALAI discontinuation included perceived lack of effect (n=2), adverse effects (n=2) and patient choice (n=2). Conclusions: There were no statistically significant differences in the demographics of the patients receiving each of the two drugs. The proportion of patients remaining on each treatment was comparable to that reported at the six month time-point of a one year naturalistic study of PLAI (Attard et al. Acta Psychiatrica Scand 2014; 130:46–51). In contrast to the study of Attard et al (2014), inpatient status on initiation did not appear to influence treatment ABSTRACTS A29 continuation at six months, although patient numbers were small. Prescribing outside of the product license accounted for 37% of patients. In part, this perhaps reflects the more limited indications of these long-acting injectable preparations compared to their oral equivalents. However, it also highlights the relatively common phenomenon of off-label prescribing within the field of psychiatry. The authors declare no financial sponsorship. PND has received funding to attend scientific meetings and honoraria for speaking engagements from Janssen-Cilag Ltd. Acknowledgement: The authors thank Mrs W. Davies for the opportunity to conduct this work.

A15 LONG-TERM ANTIDEPRESSANT TREATMENT FOR NEGATIVE SYMPTOMS IN SCHIZOPHRENIA Barnes TRE, Centre for Mental Health, Imperial College London, Hammersmith Hospital Campus, Commonwealth Building, Du Cane Road, London W12 0NN [email protected] Leeson VC(1), Costelloe C(1), Crawford M(1), Simon J(2), Kiss N(2), Geddes J(3), Taylor S(4), Osborn D(5), Killaspy H(5), Lewis G(5), Lewis S(6), Craig TKJ(7), Pathak R(8), Baldwin D(9), Keown P(10), Paton C(11), Ismail S(12), Kumar M(13) (1) As presenting author; (2) Centre for Public Health, Medical Univ of Vienna; (3) Dept of Psychiatry, Univ of Oxford; (4) Derbyshire Healthcare NHS Foundation Trust, Derby; (5) Division of Psychiatry, Univ College London; (6) Inst of Brain, Behaviour & Mental Health, Univ of Manchester; (7) King’s College London, Inst of Psychiatry, Psychology & Neuroscience; (8) Lincolnshire Partnership NHS Foundation Trust; (9) Mental Health Group, Univ of Southampton; (10) Northumberland, Tyne & Wear NHS Foundation Trust; (11) Oxleas Mental Health Foundation Trust; (12) Sheffield Health & Social Care NHS Foundtion Trust; (13) South Staffordshire & Shropshire Healthcare NHS Foundation Trust Negative symptoms of schizophrenia represent deficiencies in emotional responsiveness, motivation and socialisation that tend to be persistent despite standard antipsychotic treatment. Two sub-domains are recognised: expressive deficits and avolition-amotivation for daily-life and social activities. While RCT evidence suggests a combination of antipsychotic and antidepressant medication may be effective in treating negative symptoms, the limited evidence available does not allow for any robust conclusion about its potential risks and benefits. ACTIONS was a multi-centre, double-blind, placebo-controlled, RCT. It tested the clinical value and side effects of the SSRI antidepressant, citalopram, as an adjunct to continuing antipsychotic medication, in the management of persistent negative symptoms of schizophrenia. Sixty-two eligible participants with schizophrenia, maintained on a stable regimen of antipsychotic medication, who had persistent negative symptoms at a criterion level of severity were randomised to treatment with either placebo or citalopram at 20mg a day for 48 weeks (with the clinical option at 4 weeks to increase the daily dose to citalopram 40mg per day). The trial under-recruited, partly because of concerns about citalopram cardiotoxicity raised by the MHRA in 2011. The sample size achieved fell well short of the target recruitment of 358 participants, so the power of any statistical analysis to detect clinical or statistical meaningful significant differences between the treatment groups was limited. No therapeutic advantage was detected for adjunctive citalopram over 12 weeks or at 48 weeks in the primary outcomes. Specifically, odds ratios (and confidence intervals) for patients in the citalopram and control groups meeting criterion change scores for a clinically-relevant improvement on the PANSS negative subscale and Quality of Life scale at 12 weeks were 2.12 (0.64 to 7.13) and 2.0 (0.53 to 7.6) respectively. Secondary analysis, addressing the two negative symptom domains, found a modest improvement in avolition-amotivation at 12 weeks. There was no difference between the two treatment groups in the frequency or severity of adverse effects and specifically no difference in the duration of the QTc interval. The results of the secondary analyses allow for a cautious conclusion that citalopram can have a positive effect, at least in the short-term, on avolition-amotivation, recognised as a critical barrier to psychosocial rehabilitation and better functional outcomes. Comprehensive assessment of side-effect burden did not identify any serious safety or tolerability issues for citalopram. Further investigation of the viability and risk-benefit of this treatment strategy may be warranted. Funded by the NIHR Health Technology Assessment programme (07/83/01). A30 ABSTRACTS

A16 THIS ABSTRACT HAS BEEN WITHDRAWN

A17 ACUTE EFFECTS OF ON THE EXPERIENCE OF SPEECH ILLUSIONS: A DOUBLE-BLIND PLACEBO-CONTROLLED INVESTIGATION Shaban NDC, Clinical Psychopharmacology Unit UCL, 1–19 Torrington Place, London WC1E 7HB [email protected] Mokrysz C(1), Freeman TPF(1), Hindocha C(1), Lawn W(1), Pope R(1), Bloomfield M(1), Curran HV(1), Morgan CJ(2), Wall M(3), Nutt D(4) (1) As presenting author; (2) Centre for Clinical Psychopharmacology, Univ of Exeter, EX4 4QG; (3) Division of Brain Sciences, Imperial College London; (4) Neuropsychopharmacology Unit, Imperial College London Introduction: Many studies suggest that cannabis use increases the risk of psychosis. One main symptom of psychosis is auditory hallucinations/illusions. Acute cannabis use can also lead to transient auditory illusions. Previous work with regular cannabis users has shown that certain types of cannabis are associated with these positive psychotic-like symptoms. Participants with hair samples containing delta- 9- (THC) only were found to have more positive psychotic symptoms compared to the THC + cannabidiol (CBD) group and the no cannabinoids group. The present study aimed to test CBD’s potential neuro-protective effect acutely. The ‘White Noise Task’ was used to assess participants’ variation in detecting speech illusion in white noise (i.e. neutral random signals). Higher rates of speech illusions have been previously associated with psychometric and familial vulnerability for psychosis. Methods: 16 volunteers were assessed on three occasions in a repeated measures design where they either inhaled Placebo, THC+CBD cannabis or THC only cannabis by Volcano Medic (Storz & Bickel, Germany). THC content was balanced in active doses. Participants listened to a series of auditory stimuli (white noise only, white noise plus clear speech or white noise plus unclear speech) and responded whether speech was present. General Estimating Equations (GEE) analysis was used to determine differences in responses for hearing speech in white noise only between Placebo, THC+CBD cannabis and THC only cannabis. Results: On the placebo occasion 50.0% heard speech illusion (defined as hearing speech in at least one white noise only trials) compared to 75.0% on the THC only occasion and 81.3% on the THC+CBD occasion. Preliminary analyses suggest that when comparing cannabis to placebo, participants were more likely to experience speech illusion while on cannabis. However, no differences in rates of speech illusion were apparent between the THC only and THC+CBD cannabis groups. Conclusion: Although previous research supports that CBD has antipsychotic properties in chronic cannabis users, the present study highlights that acute CBD may not be protective against auditory hallucinations induced by the psychoactive ingredient THC. This study was supported by Drug Science.

A18 ACUTE EFFECT OF SINGLE-DOSE ANTIPSYCHOTICS ON CEREBRAL BLOOD FLOW AND STRUCTURAL METRICS Hawkins PCT, Dept of Neuroimaging IoPPN, King’s College London, De Crespigny Park London, SE5 8AF [email protected] Zelaya FO(1), Wood TC(1), Mehta MA(1), Vernon AC(2) (1) As presenting author; (2) Dept of Psychosis Studies, IoPPN, Kings College London Introduction: Antipsychotics have been implicated in the structural brain changes associated with schizophrenia over time, with both drug subtype and dose level shown to be important factors (Vita et al. 2015, Biological Psychiatry, 78(6), 403–1). Recent MRI evidence have revealed apparent changes in brain volume following just a single dose of these drugs, particularly in striatal areas with a large proliferation of D2 receptors (Tost, H., et al, 2010, Nature Neuroscience 13(8), 920–922). However, it is not clear if these changes represent genuine tissue volume changes, or are merely artefacts due to transient physiological effects of the drug, which are in turn reflected in the magnetic resonance properties of ABSTRACTS A31

1H spins (such as the T1 of water in blood or tissue). In this study we therefore examined the effects of a single, clinically relevant dose of three commonly prescribed antipsychotics on both the structure and cerebral blood flow (CBF) of the healthy human brain. Methods: Forty-two males were randomly assigned to one of two parallel groups in a double-blind, placebo-controlled, randomised, three-period cross-over study design. One group received a single oral dose of either 0.5mg, 2mg of risperidone or placebo during each visit. The other received a single oral dose of either olanzapine (7.5mg), haloperidol (3mg) or placebo. MRI acquisition was conducted at the estimated peak plasma concentration of the drug and included measures of CBF (using 3D pseudo-continuous Arterial Spin Labelling), and qualitative (T1-weighted) and quantitative (relaxometry) measures of brain structure. To fully elucidate any potential structural changes due to drug exposure, three standardised methods were employed to assess the qualitative images (VBM, Longitudinal registration and SIENA). Un-biased, whole brain and a priori striatal ROIs (bilateral putamen, caudate and ventral striatum) were assessed in all modalities. Results: Both 2mg risperidone and 3mg haloperidol caused localised increases in striatal CBF compared to placebo (1702 contiguous voxels at MNI peak -13,8,9 for risperidone, 881 contiguous voxels at peak 28,-4,6 for haloperidol, FWE corrected p=0.05) as well as widespread differential changes in other areas including the cerebellum. However, all qualitative assessments of brain volume and quantitative measures of T1 remained stable across sessions, even in those areas experiencing large changes in CBF. Conclusions: These results counter concerns that acute clinical doses of antipsychotics alter MRI measures of brain volume. Our observations indicate that studies assessing the chronic effects of these drugs are not likely to be confounded by acute transient effects in tissue structure or blood flow. It also provides a methodology to assess acute effects of other compounds on structural metrics. This study was funded by a grant from F. Hoffmann-La Roche.

A19 THE EFFECTS OF LONG-TERM PSYCHOSOCIAL STRESS ON GLUTAMATE FUNCTION: A MAGNETIC RESONANCE SPECTROSCOPY STUDY AT 3 TESLA Bloomfield MAP, Division of Psychiatry Univ College London, Maple House, 149 Tottenham Court Road London, WC1T 7NF [email protected] Stone J(1), McCutcheon R(1), Marrero-Watkins C(2), Howes OD(2) (1) Dept of Psychosis Studies, Inst of Psychiatry, Pscyhology & Neuroscience, London SE5 8AF; (2) Psychiatric Imaging Group, MRC Clinical Sciences Centre, Imperial College London, Hammersmith Hospital W12 0NN Background: All major psychiatric disorders including depression, schizophrenia, anxiety and post-traumatic stress disorder are associated with exposure to psychosocial stressors. In parallel, glutamatergic dysfunction has been associated with these disorders. We therefore sought to examine whether long-term exposure to psychosocial stress alters glutamate levels in the anterior cingulate and hippocampus. Based on the preclinical literature we hypothesised that glutamate would be reduced glutamate in the cingulate. We conducted an exploratory analysis in the hippocampus. Methods: Using a case-control design, 34 healthy participants with high or low levels of exposure to psychosocial stress were recruited. Using proton magnetic resonance spectroscopy with a 3-Tesla scanner, glutamate and glutamine levels were measured in the anterior cingulate cortex and left hippocampus of both groups. Results: There was no significant group difference in cingulate glutamate levels (p<0.40). Participants with high exposure to chronic psychosocial stress showed a significant elevation of combined glutamate and glutamine in the left hippocampus (tdf=2.2324, p < .05, effect size= 0.9). Conclusion: Contrary to hypothesis, long-term psychosocial stress was not associated with altered cingulate glutamate levels. Our main finding of increased glutamate measures in the hippoocampus is consistent with findings of studies in schizophrenia. Given the role of the hippocampus in memory processing, disrupted limbic memory processing may provide a common mechanism underlying increased risk of a range of mental illnesses. A32 ABSTRACTS

A20 STATE OR TRAIT? INVESTIGATION OF DOPAMINE OF FUNCTION IN INDIVIDUALS WITH 22Q11 DELETION- PRELIMINARY RESULTS Rogdaki M, Psychiatric Imaging Group, Robert Steiner MR Unit Imperial College Clinical Sciences Centre – Hammersmith Campus Hammersmith Hospital, Du Cane Road London W12 0NN [email protected] Howes OH(1), Veronese MV(2), Jauhar SJ(3) (1) As presenting author; (2) Dept of Neuroimaging, IoPPN, King’s College London, London, UK.; (3) Dept of Psychosis Studies, Inst of Psychiatry, King’s College London, London, UK Introduction: One of the most prevailing hypotheses of schizophrenia implicates the dopaminergic system (Lieberman et al, 1990, Schizophrenia bulletin 16, 97–110). Positron Emission Tomography (PET) studies have consistently reported increased presynaptic dopamine synthesis capacity (DSC) in individuals with schizophrenia (Hietala et al, 1999, Schizophrenia research 35, 41–50; Howes et al, 2009, Archives of general psychiatry 66, 13–20). Interestingly, striatal hyperdopaminergia has also been shown in the prodromal phase of schizophrenia (Howes et al, 2011, The American journal of psychiatry 168, 1311–1317). Although current evidence suggests that dopaminergic dysregulation has at least a state component, the evidence for the trait component is still ambiguous. This highlights the need for studies in homogenous population of individuals with shared etiological genetic risk factors. Over the last fifteen years, it has been well established that 22q11 deletion is one of the most important genetic risk factors for the development of schizophrenia (Owen et al, 2016, World Psychiatry 15, 23–25). Individuals with 22q11 deletion are at increased genetic risk for psychosis, reaching 30% for psychotic disorder (Murphy et al, 1999, Archives of general psychiatry 56, 940–945). This accounts for 1% –2% of sporadic cases of schizophrenia (Karayorgou et al, 1995, Proceedings of the National Academy of Sciences of the United States of America 92, 7612– 7616). The aim of the study was to investigate dopamine function in individuals with 22q11 deletion. Methods: 5 individuals with 22q11 deletion (3 females and 2 males, mean age: 26.4± 5.1 years old) and 19 healthy volunteers (8 females and 11 males, mean age: 23.6± 3.44 years old) underwent ([18F]-DOPA) PET scans to measure dopamine synthesis capacity (indexed as the influx ater constant Kicer). PET analysis: Movement correction was conducted using the standard approach (Turkheimer at al, 1999,J Cereb Blood Flow Metab 19, 1189–1208). Standardised ROIs was defined in the striatum, including limbic, associative and sensorimotor sub-regions, and the reference region, defined according to previous study (Mawlawi et al, 2001,J Cereb Blood Flow Metab 21, 1034–1057). The ROI atlas was normalised to each individual PET dynamic image (using the PET summation image and statistical parametric mapping (SPM). A Patlak analysis was applied to calculate influx constants (Ki values) for the whole striatal OIR relative to uptake in the cerebellar reference region (Kicer [min-1]). Results: Individuals with 22q11 deletion had a higher mean Ki value (mean= 0.0143; SD=0.0017) compared to healthy volunteers (0.0129; SD= 0.0012). This finding was statistically significant (t(22)= 2.1; p=0.04)(two tailed t-test). Conclusions: These preliminary results support the hypothesis that dopaminergic dysregulation may have a trait component. Further investigation in larger sample size is necessary to confirm this hypothesis. Sources of financial sponsorship: Medical Research Council (UK).

A21 CANNABIDIOL RESTORES COGNITIVE AND SOCIAL BEHAVIOUR DEFICITS IN A VALIDATED ANIMAL MODEL FOR SCHIZOPHRENIA Podda GP, School of Pharmacy, Univ of Manchester, Stopford Building, Oxford Road Manchester, M13 9PT [email protected] Woolley-Roberts MWR(1), Neill JCN(2), Piercy CP(3), Grayson BG(3) (1) GW Research Ltd, Sovereign House Vision Park, Chivers Way, Histon, Cambridge CB24 9BZ; (2) Manchester Pharmacy School, Univ of Manchester, Stopford Building Room 2.019a, Oxford Rd, Manchester, M13 9PT; (3) Manchester Pharmacy School, Univ of Manchester, Stopford Building, Oxford Rd, Manchester, M13 9PT Introduction: Increasing evidence shows a correlation between cannabis consumption and schizophrenia (Marconi et al., 2016, Schizophrenia Bulletin, doi: 10.1093/schbul/sbw003). The ABSTRACTS A33 endocannabinoid system has been suggested as a target for pharmacotherapy. The phytocannabinoid, cannabidiol (CBD) has shown promising results for the treatment of positive symptoms of schizophrenia compared with the atypical antipsychotic, amisulpiride (Leweke et al., 2012, Translational Psychiatry, 2(3), e94). More recently, a phase 2a placebo-controlled clinical trial of CBD showed positive results in schizophrenia patients when added to existing drug treatment (GW Pharmaceutical, http://www. gwpharm.com/news_2015.aspx, 15 September 2015). The aim of this study is to evaluate the efficacy of CBD to reverse the disruption in cognition and social behaviour induced by sub-chronic (sc) PCP in our rodent model. Methods: 140 adult female Lister Hooded rats were treated for 7 days twice daily with vehicle (n=20) or PCP (n=120) (2 mg/kg) i.p. followed by 7 days wash-out. CBD (2, 10, 20, 100 mg/ kg, i.p.) or vehicle was administered 120 minutes and risperidone (0.1 mg/kg. i.p.) 30 minutes prior to testing. Rats were tested in the novel object recognition (NOR) and social interaction tests to explore efficacy of CBD to restore recognition memory and social behaviour deficits. In the social interaction test, each animal is paired with an untreated conspecific and social behaviour is analysed in some detail for 10 minutes. Exploration data were analysed by student’s t-test and social behaviour were analysed by a one-way ANOVA, followed by a post-hoc LSD test. Results: As consistently demonstrated by us and others, scPCP significantly impaired object recognition memory in the NOR test, such that animals could not discriminate between the novel and familiar object. Risperidone and CBD at 2, 20 and 100 mg/kg (but not at 10 mg/kg (p=0.07) significantly reversed the scPCP-induced deficit (p<0.01). Again, as consistently demonstrated by us, scPCP significantly reduced sniffing behaviour (p<0.05) in the social interaction test. This effect was improved by CBD at 100 mg/kg only, an effect that just failed to reach statistical significance (p=0.056).We observed a significant increase in the number of avoidances in the PCP-treated group (P<0.001) compared with vehicle. This increase was significantly (P<0.01-P<0.001) improved by CBD at all doses tested (2–100 mg/kg) and by risperidone (P<0.001). Conclusion: CBD given acutely reversed both the cognitive and social behaviour deficits induced by scPCP. These results support the hypothesis that CBD could have efficacy for the treatment of negative and cognitive symptoms of schizophrenia. Declaration of interest: Jo Neill has received expenses to attend conferences and fees for lecturing and consultancy work (including attending advisory boards) from the manufacturers of various antipsychotic drugs. Sources of financial sponsorship: This work is supported by GW Pharmaceuticals and b-neuro at the Univ of Manchester School Of Pharmacy.

A22 PDE4 AND PDE10 INHIBITORS HAVE DIFFERENT EFFECTS ON AUDITORY PROCESSING IN MICE Scott L, Neuroscience and Pain Research Unit Pfizer Inc, 610 Main Street, Cambridge, MA, USA, 02139 [email protected] Buzby CY(1), Graf G(1), Hughes ZA(1) (1) As presenting author Introduction: Sensory processing abnormalities have been observed in numerous psychiatric diseases, including schizophrenia. Auditory-evoked potentials (AEPs) and pre-pulse inhibition (PPI) have been used to assess these sensory processing deficits in schizophrenia patients as well as in rodent models of schizophrenia. We compared the effects of the novel PDE4 inhibitor, ABI-4, along with those of our positive control clinically prescribed antipsychotics, risperidone and haloperidol, and our negative control, PDE10 inhibitor PF-02545920 in two assays assessing sensory processing in mice. Methods: Adult male C57BL/6J mice were implanted with recording electrodes over the frontal cortex, with ground and reference electrodes over the cerebellum, and allowed to recover for 2 weeks prior to habituation and drug testing. AEPs were recorded using a paired pulse paradigm, 75dB 10ms white noise pairs were presented with 0.5s inter-stimulus interval and 10s inter-trial interval. Drug testing utilised a randomised cross-over design, where each animal received each treatment once during the study, with 1 week wash-out between test sessions. Sensory motor gating was assessed in a separate cohort of adult male C57BL/6J mice, using an established PPI of startle paradigm. Startle response to a 120dB tone was assessed either with or without the presentation of sub-startle threshold pre-pulses of 4–16dB above background noise. Results: The PDE4 inhibitor ABI-4 (0.01–1.0 mg/kg, s.c.) caused a dose-dependent increase in the AEP in response to the first tone of each pair presented (S1) while also decreasing the A34 ABSTRACTS response to the second tone presented (S2), resulting in enhanced auditory gating (vehicle S2/S1= 0.41 ± 0.04 ; ABI-4 (0.1 mg/kg, s.c.) S2/S1 = 0.26 ± 0.02; P<0.001, ANOVA followed by Dunnett’s post hoc test). The resting EEG was also altered dose-dependently with EEG power decreased across all frequency bands. ABI-4 also increased the PPI response in a dose-dependent manner, with the minimum effective dose of 0.1mg/kg, s.c.; P<0.05 vs vehicle, ANOVA followed by Dunnett’s post hoc test). Our positive control treatments, risperidone and haloperidol, were also shown to increase auditory gating in a dose- dependent manner, however this increase in gating was mediated through an increase the S1 AEP while the S2 response was unchanged (vehicle S2/S1= 0.43 ± 0.02 ; risperidone (0.3 mg/kg, s.c.) S2/S1 = 0.34 ± 0.02, P<0.05, ANOVA followed by Dunnett’s post hoc test; haloperidol (3.0 mg/kg, s.c.) S2/S1 = 0.29 ± 0.02, P<0.001, ANOVA followed by Dunnett’s post hoc test). Risperidone and haloperidol also dose-dependently affected the resting EEG decreasing delta (1–4Hz) and theta (4–12Hz) power, but these compounds had no effect on gamma (30–100Hz) power. Our negative control, PF-02545920 had no effect on the AEP produced following S1 or S2 or on EEG power across any frequency band, as well as showing no effect on PPI. Conclusion: The PDE4 inhibitor ABI-4 dose-dependently improves sensory gating, as do risperidone and haloperidol, whilst the PDE10 inhibitor PF-02545920 had no effect. These studies support the use of AEPs as an electrophysiological end-point in clinical proof of mechanism and potentially proof of concept studies, as well as increasing confidence that this assay may prove useful in supporting other pre-clinical programs targeting sensory processing deficits.

A23 ADJUNCTIVE LURASIDONE SUPPRESSES FOOD INTAKE AND WEIGHT GAIN ASSOCIATED WITH OLANZAPINE ADMINISTRATION IN RATS Reynolds GP, BMRC Sheffield Hallam Univ, Howard St Sheffield, S1 1WB [email protected] Dalton CF(1), Watrimez W(2), Jackson J(2), Harte MK(2) (1) As presenting author; (2) School of Pharmacy, Univ of Manchester Introduction. Lurasidone shows a relative lack of the weight gain common to many antipsychotic drugs. Aripiprazole and ziprasidone have a similar metabolic profile and demonstrate protective effects against olanzapine-induced food intake and weight gain in animals (Snigdha et al 2008 J Psychopharmacol 22, 567–571). We tested the hypothesis that lurasidone, in comparison with and combined with olanzapine, would also demonstrate reductions in acute food intake and short term weight gain in rats. Methods. Female Lister hooded rats were given 30 mins access to a palatable mash food preparation. Four groups of six animals received (i.p.) either 2x vehicle (saline), lurasidone (3mg/kg) and vehicle, olanzapine (1mg/ kg) and vehicle, or olanzapine and lurasidone. Animals were given access to the food for a further 60 mins. Weight of initial and final food intake was measured. An equivalent series of rats underwent a seven-day regime of once-daily administration of the above doses and free access to food and water. Prior to each dose the animals were weighed, and weight gain over the course of the study was monitored; differences were determined by t-test. Results. The initial acute studies showed increased food intake with olanzapine (p=0.031 vs vehicle) in the absence of any significant effect of lurasidone administration which remained reduced below olanzapine (p=0.019 vs olanzapine). Co-administration of lurasidone with olanzapine suppressed the increased food intake seen with the latter (p=0.043 vs olanzapine). Repeated dosing for 7 days showed an equivalent effect. A relative increase in body weight occurred after olanzapine (p=0.046 vs vehicle), with no significant effect observed with either lurasidone or lurasidone and olanzapine. Repeated administration of lurasidone with olanzapine suppressed the effect of olanzapine alone on body weight, although this did not achieve statistical significance (p=0.076 vs olanzapine). Conclusions. These findings show that, in addition to having no signficant effect on acute food intake and chronic weight gain in clear contrast to olanzapine, lurasidone demonstrates a protective effect on olanzapine-induced food intake and weight gain in rats. This indicates the drug to have an active anti-hyperphagic mechanism, rather than solely the absence of a propensity for weight gain and suggests a clinical potential for such a protective effect of adjunctive lurasidone in antipsychotic-induced weight gain, an effect already demonstrated for aripiprazole. This study was supported by Sunovion. ABSTRACTS A35

A24 CHRONIC TREATMENT WITH RO4993850, A GLYT1 INHIBITOR, ATTENUATED SOME OF THE SOCIAL DEFICITS IN ADULT RATS INDUCED BY NEONATAL-PHENCYCLIDINE WITH ISOLATION-REARING, BUT DID NOT AFFECT PARVALBUMIN EXPRESSION IN THE HIPPOCAMPUS OR PFC Kohli S, School of Life Sciences Univ of Nottingham, Univ of Nottingham Medical School QMC Nottingham NG7 2UH [email protected] King MV(1), Fone KCF(1), Alberati D(2), Ballard TM(2), Steward LJ(2) (1) As presenting author; (2) Neurosci, Roche Innovation Center, CH-4070, Basel Neonatal-phencyclidine (PCP) followed by isolation-rearing from weaning, causes neurodevelopmental behavioural changes in adult rats akin to symptoms seen in schizophrenia, as well as neuropathological changes. transporter (GlyT1) inhibitors, such as RO4993850 (an analogue of Bitopertin), regulate synaptic levels of the NMDA receptor co-agonist, glycine, and are a potential therapeutic to restore NMDA receptor hypofunction thought to occur in schizophrenia. Sixty-eight male Lister-hooded rat pups (Charles River UK) received 10mg/kg PCP or saline (2ml/kg s.c.) on post-natal days (PND) 7, 9 and 11, with subsequent housing as a littermate group (SAL-GH) or alone (PCP-SI) from weaning (PND23) and the study duration. The ‘behavioural syndrome’ was assessed by novel arena-induced locomotor activity (PND61 for 60 min) before injection of RO4993850 (3mg/kg) or vehicle 4ml/kg i.p. for 14 days (n=16–18). On day 14 social interaction behaviours were recorded between rats (paired by weight, litter and treatment) during a 10 min trial. Post-trial (90 min) rats were transcardially perfused under anaesthesia and brains processed by immunohistochemistry to count parvalbumin-positive interneurons using a fluorescence microscope in key regions of interest. In a novel arena, PCP-SI were significantly more hyperactive (mean ± SEM=2165 ± 60 total beam breaks) than SAL-GH (1879 ± 71; p<0.01; Student’s t-test) rats. PCP-SI displayed reduced pro-social interaction compared to SAL-GH rats, which was attenuated by RO4993850 such that there was a significant PCP/housing x treatment interaction (p=0.001) but no main effect of either alone. PCP-SI displayed less (p<0.0001) ano-genital sniffing, but more body sniffing (p<0.0001) and boxing/biting (p=0.042) than SAL-GH controls. RO4993850 significantly increased (p<0.05; Tukey post-hoc) body sniffing in PCP-SI rats. In the adult hippocampus PCP-SI significantly reduced the number of parvalbumin-positive neurons, with a main effect of PCP/housing in the CA1 (p=0.0001), CA2/3 (p=0.0009) and dentate gyrus (p=0.0191) hippocampal sub-regions. By contrast, there was no change in parvalbumin-immunoreactivity in any prefrontal cortex (PFC) sub-region (although a main effect of PCP/housing almost reached significance in the dorsolateral orbitofrontal cortex p=0.0655). Treatment with RO4993850 did not affect parvalbumin-immunoreactivity numbers in the PFC or hippocampus. PCP-SI reduced pro-social behaviour and the number of parvalbumin-GABAergic interneurons in the adult rat hippocampus, which may result from altered GABAergic neuronal connectivity. Chronic treatment with RO4993850 attenuated some of these social deficits, but did not influence parvalbumin expression in the hippocampus or PFC. GlyT1 inhibitors may be of use to treat social deficits seen in disorders such as schizophrenia or ASD. Funded by F.Hoffman La-Roche Ltd.

A25 EXERCISE RESCUES THE NOVEL OBJECT RECOGNITION DEFICIT IN THE SUB-CHRONIC PHENCYCLIDINE RAT MODEL FOR SCHIZOPHRENIA Heaney LM, School of Pharmacy, Univ of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT [email protected] Harte MK(1), Neill JC(1), Yung AR(2) (1) As presenting author; (2) IBBMH, Uni of Manchester, Jean McFarlane Building, Oxford Road, Manchester, M13 9PL Introduction: Cognitive impairment in schizophrenia remains an unmet clinical need that contributes to markedly reduced quality of life. Exercise has an effect to alleviate cognitive impairment in certain patients (Pajounk et al., 2010; Arch Gen Psychiatry), but the mechanisms are poorly understood. The sub-chronic phencyclidine (PCP) rat model for schizophrenia is a well validated, widely used tool in the preclinical study of the disease (Neill et al., 2010; J Pharm & Ther. 128(3): 419–32), and the novel object recognition (NOR) A36 ABSTRACTS task is widely used to test cognitive deficits in a range of disorders (Grayson et al., 2015; Behav. Brain Res. 285: 176–193). Our aim is to use this animal model to explore the mechanisms by which exercise exerts its pro-cognitive effects in the clinic. Methods: Female Lister Hooded rats (n=40), were habituated to voluntary wheel running cages, then pseudo-randomised into four groups; vehicle control; vehicle exercise; PCP control and PCP exercise (n=10 per group). Rats were either treated with saline (vehicle groups) or PCP (2mg/kg, i.p. bi-daily followed by a seven day washout). Vehicle and PCP exercise groups were allowed access to a wheel for one hour a day, five days a week, for six weeks. The vehicle and PCP control groups were placed individually in identical cages with an upturned plastic tunnel as the control condition. Rats were tested in the NOR paradigm pre- and post-exercise. Data were analysed using a t-test or two-way ANOVA. Results: Post exercise both the vehicle exercise and vehicle control groups spent significantly more time exploring the novel versus the familiar object (p<0.01). As we have demonstrated several times, the PCP control group couldn’t discriminate between the novel and familiar object. In contrast the PCP exercise group spent significantly more time exploring novel versus the familiar object (p<0.001). Conclusions: These results show that daily exercise, 5 days per week over 6 weeks, attenuates the NOR deficit in the sub-chronic PCP rat model. This corresponds with human studies reporting positive effects of exercise in the treatment of cognitive deficits in schizophrenia, providing a means to thoroughly investigate the mechanism(s) of the pro-cognitive effects of exercise. LM Heaney is supported by a Medical Research Council Ph.D. Studentship Award and the Univ of Manchester President’s Doctoral Scholarship.

A26 MATERNAL POLY I:C INJECTION FOLLOWED BY POST-WEANING ISOLATION REARING – A NOVEL ‘DUAL-HIT’ MODEL FOR SCHIZOPHRENIA? Goh J-Y, School of Life Sciences, Univ of Nottingham, Queen’s Medical Centre Nottingham, NG7 2UH [email protected] Fone KCF(1), King MV(1) (1) As presenting author Maternal viral infection and early-life social adversity are recognised environmental risk factors for schizophrenia. Polyinosinic:polycytidylic acid (poly I:C) administration and post-weaning isolation rearing have both independently been widely used as rat neurodevelopmental models to replicate schizophrenic symptoms. This study evaluated their combination as a novel ‘dual-hit’ rat model for schizophrenia. Seventeen Lister-hooded dams (CRUK, 280–337g) received temperature microchips (s.c.) on gestational day (GD) 14 and i.p. saline (1ml/kg, Veh) or poly I:C (10mg/kg) on GD15. Body temperature and home cage behaviour (e.g. activity, piloerection) were monitored for 6h and body weight for 3d post-injection. Male and female pups (n=96 and 32, respectively) were housed in groups (4/cage; GH) or individually (SI) from weaning (post-natal day 22). Locomotor activity (LMA) was evaluated in both genders (n=8/combination/gender) two weeks post weaning, immediately followed by brain and blood collection. Separate male littermates (n=16/combination) underwent consecutive assessment of LMA, novel object/location discrimination (NOD/NOL), social interaction, pre-pulse inhibition of acoustic startle (PPI) and conditioned fear response (CFR) at one-week intervals starting five weeks post-weaning. Data (mean±SEM) were analysed by 2-way ANOVA with Sidak/Tukey post-hoc. Poly I:C-treated dams showed reduced body temperature (peak 2h post-injection, -1.3±0.2°C, p<0.001), increased sickness behaviour (peak 3–4h, p<0.0001) and weight loss compared to vehicle at 24h and 72h (p<0.05). At 2 weeks post-weaning male Veh-SI and poly I:C-SI were both hyperactive but this effect was less robust in females. Male Veh-SI were hyperactive 5 weeks post-weaning but there were no differences in NOD/ NOL. In the social interaction test, poly I:C-SI tended to reduce total social behaviour (136±14s) compared to other groups (Veh-GH=156±17s, Poly I:C-GH=158±13s,Veh-SI=157±12s). Furthermore, Veh-SI and poly I:C-SI showed increased body sniffing and reduced anogenital sniffing compared to Veh-GH (p<0.001, p<0.01, respectively) and Poly I:C-GH (p<0.0001, p<0.01, respectively). In poly I:C-SI this change occurred without increased aggression (boxing and biting; 7±2s) seen in Veh-SI (18±4s; p<0.01). There were no PPI impairments. In the CFR paradigm, only Veh-SI (p<0.05) showed reduced freezing to the light and tone cue previously paired with foot shock at 24h (131±23s, Veh-GH=225±20s, Poly I:C-GH=189±15s, Poly I:C- SI=222±26s). Exposure of pregnant dams to a viral mimetic induced sickness behaviour and weight loss ABSTRACTS A37 and appeared to attenuate several components of the behavioural syndrome produced by subsequent isolation rearing of the offspring. Future pharmacological work will examine whether this ‘dual-hit’ model is a predictive marker for modelling negative symptoms of schizophrenia. Funded by Univ of Nottingham and Monash Univ.

A27 ACUTE HALOPERIDOL CHALLENGE ALTERS THE PROTEIN EXPRESSION OF THE TRANSLOCATOR PROTEIN (TSPO) IN A REGION-SPECIFIC MANNER IN THE NAÏVE ADULT RAT BRAIN Cotel MC, Basic and Clinical Neurosciences, IOPPN, The Maurice Wohl Clinical Neuroscience Inst 5 Cutcombe Road London SE5 9RT [email protected] Williams SCR(1), Kapur S(2), Tilkitzi K(3), Cooper JD(3), Vernon AC(3) (1) Center for Neuroimaging Sciences, De Crespigny Park, London SE5 8AF; (2) Inst of Psychiatry, De Crespigny Park, London SE5 8AF; (3) Maurice Wohl Clinical Neurosciences Inst, 5 Cutcombe road, London SE5 9RT Introduction: Positron emission tomography using TSPO specific radio-ligands is a common in vivo measure of putative microglial activation in the brain. Patients with chronic schizophrenia (SZ) show significantly elevated TSPO radio-ligand binding (Bloomfield et al., Am J Psych., 2015). However, these data are not equivocal (Kenk et al., Mol. Psych., 2015). Notably, the effects of antipsychotic medication on TSPO are unclear, making it difficult to dissect drug from disease effects in chronically ill SZ patients.e W therefore investigated whether a single dose of haloperidol leads to measurable changes in TSPO protein expression in the naïve rat brain. Methods: Adult male Sprague-Dawley rats (250–300 g) were injected with either drug vehicle (n=6) or a clinically-relevant dose of haloperidol (0.05 mg/kg s.c.; Kapur et al. JPET, 2003). At peak plasma drug concentration (120 min post-dose), rats were culled and anterior cingulate cortex (ACC), corpus striatum (STR) and cerebellum (CB) dissected and flash frozen. Whole tissue lysates were probed for TSPO protein expression by western blot, relative to β-actin. Immunoreactive proteins of 18, 36, 54 and 108 kDa were present in each brain region, suggesting the presence of TSPO polymers. The optical densities of each band were quantified from digital images of the films using ImageJ software. Group differences were investigated using 2-way ANOVA with main effects of haloperidol and region, and post-hoc Bonferroni test if p(ANOVA) <0.05, using GraphPad (Prism Software, USA). Results: There were no significant effects of haloperidol on expression of the 18 kDa TSPO monomer (F1, 30 = 1.78; p>0.05). orF the 36, 54 and 106 kDa polymers, there were significant region x treatment interactions (36 kDa: F2, 30 = 3.98; p<0.05; 54 kDa: F2, 30 = 14.26; p<0.001; 108 kDa: F2, 30 = 5.08; p<0.05). Post-hoc testing revealed decreased expression of the 36 and 54 kDa TSPO polymers in the CB of haloperidol treated rats relative to vehicle (p<0.05 and 0.01 respectively). In contrast, expression of both the 54 and 108 kDa TSPO polymers were increased in the STR of haloperidol-treated rats (both p<0.05). Conclusions: These data provide preliminary evidence that a single, clinically relevant dose of haloperidol leads to region-specific measurable changes in TSPO protein expression. Furthermore, we extend prior work in Leydig cells (Delavoie et al., Biochemistry, 2003) to suggest TSPO polymers exist in the mammalian brain. Further studies are required to determine the relationship to TSPO radio-ligand binding following acute and chronic antipsychotic exposure. This study was generously supported by a Medical Research Council (MRC) and a King’s Health Partners (KHP) grants.

A28 KETAMINE BLOCKS THE EXTINCTION OF CONTEXTUAL FEAR MEMORY Clifton NE, NMHRI, Cardiff Univ, Hadyn Ellis Building, Maindy Road, Cardiff CF24 4HQ [email protected] Thomas KL(1), Hall J(1) (1) As presenting author Introduction: Recent large-scale studies of copy number variants in schizophrenia have linked the disorder to processes of associative learning. Some evidence suggests that ketamine, an NMDA antagonist and psychomimetic, may impair associative learning, but it is currently unclear which phases A38 ABSTRACTS of the learning process are affected. Here, we used contextual fear conditioning in rats to investigate the effect of ketamine on the consolidation, recall and extinction of fear learning. Methods: Adult male Lister Hooded rats were placed into a novel context for 2 min before receiving a 0.5mA foot shock. Freezing behaviour was quantified in the 1 min following foot shock and for 2 min (recall) or 10 min (extinction) in the same context 48 hours later. Ketamine (8 or 25 mg/kg) or vehicle were administered 30 min prior, or immediately after, conditioning or extinction. Memory of CS-US pairing or extinction was tested in further recall trials (2 min) with or without the administration of ketamine. Results: Ketamine (25 mg/kg) administration prior to conditioning induced a state-dependent contextual fear memory. Both 8 mg/kg and 25 mg/kg ketamine pre-treatment blocked the extinction of contextual fear memory (25 mg/kg F(2, 15) = 11.59, P < 0.001, 8 mg/kg F(3, 20) = 3.471, P < 0.05, two-way repeated measures ANOVA). Administration immediately after consolidation or extinction had no effect on the contextual fear memory. Conclusions: These results suggest that ketamine specifically impairs the extinction of contextual fear memory, but not its consolidation or recall. This effect only occurs when ketamine is present during the acquisition of new learning, not by influencing the molecular processes that occur subsequently. The findings have important implications for the role of inhibitory learning processes in schizophrenia and related psychiatric disorders. Financial Sponsorship: This work was funded by a PhD grant to NEC from the Wellcome Trust.

A29 DISRUPTION OF LATENT INHIBITION BY SUB-CHRONIC PHENCYCLIDINE PRE-TREATMENT IN RATS Al-Ali A, Psychology, Leicester Univ, Lancaster Road, Leicester LE1 9HN, UK, LE1 9HN [email protected] Young AMJ(1) (1) As presenting author The NMDA-type glutamate receptor antagonist, phencyclidine (PCP) evokes schizophrenia-like symptoms in normal people, and exacerbates symptoms in schizophrenia patients. In rats, sub-chronic PCP treatment induces behavioural changes resembling negative and cognitive schizophrenia symptoms (Neill et al., 2010, Pharmacology & Therapeutics, 128(3), 419–432). However, little is known about effects of sub-chronic PCP on behavioural changes related to positive symptoms. Latent inhibition (LI: repeated pre-exposure (PE) of a stimulus impairs subsequent conditioning to that stimulus) models aspects of behaviour which are dysfunctional in positively symptomatic schizophrenia sufferers (Gray et al., 1995, Behavioural Brain Research, 71, 19–31). This study investigated the effect of sub-chronic PCP pre-treatment on LI in rats. Female Lister-hooded rats (c250g) were randomly assigned to test groups and pre-treated with PCP (2mg/ kg, i.p.) or saline (1ml/kg, i.p.) twice/day for 5 days, then left for 10 days before LI testing. They received restricted water availability (8 hours/day), and were trained to lick for water in the test chamber. For PE, animals were placed in the chamber, and presented with 40 tones (5sec) at 1min intervals: non- pre-exposed (NPE) animals were in the chamber for an equivalent time, but without tone presentation. Next day all animals underwent conditioning with two pairings of tone with mild footshock (1sec) at 5min intervals. The following day, animals were returned to the chamber, with free access to water, and conditioned suppression of licking was measured during presentation of the tone. The time for 10 licks prior to tone onset (TimeA) and for 10 licks during tone presentation (TimeB) were measured and a suppression ratio (SR) was calculated (TimeA/(TimeA+TimeB)): SR close to zero indicated good learning, and SR close to 0.5 indicated poor learning. Statistical analysis was by ANOVA with appropriate post hoc testing. All procedures had appropriate project (PPL6004390) and personal licence authority under the Animals [Scientific Procedures] Act, 1986. Saline pre-treated animals showed LI, as expected, with significantly reduced learning to the PE stimulus (SR[NPE]=0.055±0.019; n=8: SR[PE]=0.416±0.077; n=8: p<.001). However, animals pre-treated with PCP showed no LI, due to attenuation of the PE effect (SR[NPE]=0.025±0.005; n=8: SR[PE]=0.148±0.056; n=8: p>.05). These experiments demonstrate that sub- chronic PCP pre-treatment disrupts LI. Importantly, testing occurred >10 days after PCP pre-treatment, indicating that in LI, as in other behavioural tasks (Neill et al., 2010, ibid) the effect of PCP pre-treatment endures beyond the time the drug is present in the brain. Financial sponsorship: Iraq Ministry of Higher Education and Scientific Research. ABSTRACTS A39

A30 AUT00206, A NOVEL AND SELECTIVE KV3 CHANNEL MODULATOR RESTORES COGNITIVE FUNCTION IN THE SUB-CHRONIC PHENCYCLIDINE RAT MODEL FOR SCHIZOPHRENIA IN THE PRESENCE OF CHRONIC TREATMENT WITH ANTIPSYCHOTIC DRUGS Cadinu D, Manchester Pharmacy School, Univ of Manchester, Stopford Building, Oxford Rd, Manchester UK M13 9PT [email protected] Podda G(1), Doostdar N(1), Harte MK(1), Neill JC(1), Alvaro G(2), Large C(3) (1) As presenting author; (2) Autifony S.r.l., Verona, Italy; (3) Autifony Therapeutics Ltd, Imperial College Incubator, Bessemer Building, Imperial College London, SW7 2AZ, UK Introduction. Although antipsychotic drugs alleviate psychotic symptoms of schizophrenia, cognitive symptoms remain an unmet clinical need (Keefe et al. 2007. Arch Gen Psychiatry. 64: 633–647). We have previously demonstrated efficacy of a novel Kv3.1/Kv3.2 channel modulator, AUT00206, to reverse sub-chronic PCP (scPCP) induced cognitive deficits in antipsychotic naive rats (Leger et al. 2015. J Psychopharmacology Suppl. 29(8), A66). Although scPCP is a well validated model for chronic schizophrenia, in order to fully mimic the clinical condition, the animal model must include chronic antipsychotic treatment prior to treatment with any novel drug. Here we test the efficacy of AUT00206 to restore cognitive function in two tests of different domains affected in the illness in scPCP treated rats that also received chronic treatment with the typical drug, haloperidol or the atypical antipsychotic, olanzapine at doses providing >70% D2 occupancy. Methods. Two separate cohorts of adult female Lister Hooded rats, received vehicle (n=20) or scPCP (2 mg/kg; n=85) i.p. twice daily for 7 days, followed by 7-days washout. scPCP-treated rats then received 14 days treatment with haloperidol (0.1 mg/ kg, i.p. once per day) or olanzapine (1.2 mg/kg, i.p. once per day) followed by acute AUT00206 (novel object recognition) or 21 days treatment with haloperidol with AUT00206 at 60 mg/kg given for 14 days from day 7 onwards (attentional set shifting task). We chose 60 mg/kg, a dose previously shown to be effective in the scPCP model. Data were analysed using ANOVA followed by post-hoc LSD t-test or using Student’s t-test. Results. Acute AUT00206 treatment restored NOR deficits induced by scPCP alone and after chronic treatment with haloperidol (P<0.05) and olanzapine (P<0.05). scPCP treatment produced a significant and selective deficit in the EDS phase of the attentional set shifting task (P<0.001) which was significantly attenuated by chronic treatment with AUT00206 at 60 mg/kg (P<0.001) as well as by the combination of AUT00206 and haloperidol (P<0.001). Conclusion. Studies in animals (Dean, 2006, Prog Neuropsychopharmacol Biol Psych, 30, 174–189) and in chronic schizophrenia patients (Navari and Dazzan, 2009, Psychol Med, 39, 1763–1777) suggest that antipsychotics can adversely affect neuronal structure and function which may further impair cognition. Our data demonstrate the efficacy of a novel Kv3.1/Kv3.2 channel modulator in the scPCP model of cognitive deficits of schizophrenia, in the presence of antipsychotic drugs. These results provide support for the hypothesis that modulation of Kv3.1 channels can restore cognitive function in 2 different domains affected in schizophrenia in the presence of chronic antipsychotic treatment. Sources of financial sponsorship: the work is supported by Innovate UK and Autifony Therapeutics Limited. Declaration of interest: Jo Neill has received expenses to attend conferences and fees for lecturing and consultancy work (including attending advisory boards) from the manufacturers of various antipsychotic drugs.

A31 ALTERED EXPRESSION OF PLASTICITY GENES ARC AND BDNF IN THE CYFIP1+/- KO MOUSE AS A MODEL OF GENETIC RISK FOR PSYCHIATRIC ILLNESS Trent S, Neuroscience & Mental Health Research Inst, Cardiff Univ, 3.34 Hadyn Ellis Building, Maindy Road, Cardiff CF24 4HQ [email protected] Best C(1), Thomas KL(1), Hall J(1) (1) NMHRI, Cardiff Univ, Cardiff, CF24 4HQ Introduction: Neuropsychiatric illnesses such as schizophrenia and Autism Spectrum Disorder are characterised by clinically diagnosed psychiatric symptoms, often with severe cognitive deficits that remain under-researched and poorly treated. Recent genomic studies implicate a convergent biological A40 ABSTRACTS mechanism at the glutamatergic synapse, including Cyfip1, FMRP and ARC proteins, which may explain the genetic risk for psychiatric illness and accompanying cognitive deficits (Hall et al., 2014, Biological Psychiatry, 77:52–58). CYFIP1 protein acts in concert with FMRP, to repress the protein translation of ARC, a BDNF-activated plasticity-related protein that is vital in memory (Trent et al., 2015, Nature Communications, 6:7897). Here we utilised a Cyfip1 +/- heterozygous KO mouse model, analogous to the reduced dosage observed in humans, to explore dysregulation of Arc and BDNF gene expression in key brain areas for cognition and sites of dysregulation in psychiatric illness. Methods: Cyfip1+/- heterozygous knockout mice (MRC Harwell, Taconic 6NTac background) were crossed with a CR JAX (6J) line producing a hybrid 6N/6J line. Whole brains from naïve adult males and females were extracted (2–3 months old), as were wild-type littermate controls (n=14/genotype for males, n=15/genotype for females). Prefrontal cortical and hippocampal regions were dissected and prepared for quantitative real-time PCR analysis of Arc and BDNF exon IX mRNA (communal to all BDNF transcripts). Gapdh and Hprt acted as housekeeper genes and fold change of gene expression was performed via the ΔΔCt method. Statistical comparisons were made using a 1-way ANOVA (normality assessed via the Shapiro- Wilk test). Results: qPCR revealed a significant 2.5-fold increase of Arc mRNA in the prefrontal cortex of female Cyfip1 knockout mice compared with female wild-type littermate controls (F(1,29)=16.851, P=0.0001). However, male Cyfip1 KO mice did not exhibit altered Arc expression in the prefrontal cortex (F(1,27)=0.636, P=0.433, vs male wild-type littermates), whilst no changes were seen for hippocampal Arc mRNA expression in male and female Cyfip1 KO mice compared to wild-types. Meanwhile, there was widespread reduction of BDNF exon IX mRNA, in female and male Cyfip1 KO mice in both brain regions (40% decrease in male Cyfip1 KO hippocampus, F(1,27)=4.410,P=0.046). Conclusion: In an animal model of reduced Cyfip1 dosage, we observed altered expression of plasticity-relevant genes, Arc and BDNF. The sexually dimorphic dysregulation of Arc requires further investigation, with complementary protein assays currently ongoing. Meanwhile, BDNF is widely reduced and not only may it drive increased risk for psychiatric illness and cognitive impairments, but provide a tractable target for therapeutic rescue. Sources of financial sponsorship: Wellcome Trust, Brain & Behavior Research Foundation (NARSAD grant).

A32 GENETIC KNOCKDOWN OF THE SCHIZOPHRENIA RISK GENE CACNA1C AFFECTS SPECIFIC COMPONENTS OF ASSOCIATIVE LEARNING AND THE EXPRESSION OF RELEVANT DOWNSTREAM GENES Sykes LH, NMHRI, Cardiff Uni, Hadyn Ellis Building, Maindy Road, Cardiff CF24 4HQ [email protected] O’Donovan M(1), Thomas KL(1), Hall J(1) (1) As presenting author CACNA1C, encoding the alpha-1 subunit of Cav1.2 L-type voltage-gated calcium channels (LVGCC), shows genome-wide significant association with schizophrenia and other psychiatric illnesses (Smoller et al., 2013, Lancet, 381, 1371–9). Risk variants have been shown to be putative cis-expression quantitative trait loci for CACNA1C, with reduced expression associated with increased risk (Gershon et al., 2013, Mol. Psychiatry, 1–5). Calcium influx and gene expression via LVGCCs is necessary for forms of synaptic plasticity and learning, and impairments in associative memory are consistently found in psychiatric disorders (Pohlack et al., 2011, 16, 1072–3). However, the role of CACNA1C and related signalling in cognitive processes affected in schizophrenia is not yet fully understood. We investigated whether global genetic knockdown of Cacna1c in a novel rat model altered associative learning behaviour and the expression of downstream target genes. Methods: Heterozygous Cacna1c knockout rats were created using -Finger-Nuclease (ZFN) technology on a Sprague Dawley background (Sage). Cohorts of adult male rats (heterozygous and wild-type littermate controls) underwent contextual fear conditioning in a naïve context, to assess the effect of mutation on memory acquisition, consolidation, extinction and latent inhibition. In an additional cohort, regional basal expression of Cacna1c and Brain-Derived Neurotrophic Factor (BDNF), a target gene regulated by calcium influx through LVGCCs and a key regulator of learning and memory (Zheng et al., 2011, PLoS One, 6, 1–14), was measured using ABSTRACTS A41

RT-qPCR. Results: There was no effect of genotype on the acquisition or consolidation of contextual fear memory. However, there were selective impairments in heterozygous rats in extinction (Repeated measures ANOVA, Genotype: F(1,12) = 5.103 P=0.043) and latent inhibition (F(1,7) = 14.695, P=0.006). In addition, we observed a brain region specific effect of genotype on the expression of individual transcripts of BDNF (Student’s t-test: t(7) = 2.566, P=0.037). Conclusions: Reduced expression of Cacna1c results in specific deficits in inhibitory learning, but not consolidation of contextual fear memory. Furthermore, the reduction in Cacna1c is associated with changes in expression of BDNF variants. Reduction in peripheral BDNF has been reported in patients with schizophrenia (Green et al., 2011, Mol. Psychiatry, 16, 960–972). The genetic risk of the CACNA1C variant in schizophrenia may be directly linked to dysregulated gene expression. The identity of BDNF as part of the molecular mechanism for the associated cognitive deficits would provide a target for novel personalised treatments for schizophrenia and other psychiatric illnesses. Funded by Wellcome Trust.

A33 METABOTROPIC GLUTAMATE 3 RECEPTOR (MGLU3): ANTIBODY CHARACTERISATION García-Bea A, Psychiatry, Univ of Oxford, Neurosciences Building, Warneford Hospital, Oxford OX3 7JX [email protected] Walker MA(1), Lane TA(1), Harrison PJ(2), Kleinman JE(3), Hyde TM(4) (1) As presenting author; (2) Dept of Psychiatry, Univ of Oxford, Oxford, UK and Oxford Health NHS Foundation Trust, Warneford Hospital, Oxford, UK; (3) Lieber Inst for Brain Development, Baltimore, USA; (4) Lieber Inst for Brain Development, Baltimore, USA and Depts of Neurology and Psychiatry & Behavioral Sciences, John Hopkins School of Medicine, Baltimore, USA Introduction. Group II metabotropic glutamate receptors comprised of mGlu2 and mGlu3 are glutamatergic presynaptic inhibitory autoreceptors. GRM3, which encodes mGlu3, shows genetic association with schizophrenia, and mGlu3 is a possible treatment target. However, attempts to determine the distribution and abundance of mGlu3 in schizophrenia have been hampered by the use of poorly characterised antibodies. Here, we address this limitation by characterizing six commercially available mGlu3 antibodies for specificity, subtype selectivity, and suitability for use in post mortem human brain. Methods. We carried out western blot experiments with each antibody on membrane protein isolated from brains of mice lacking mGlu3 (grm3-/-) or both mGlu2 and mGlu3 (grm2/-/3-/-), as well as on protein extracted from HEK293T/17 cells overexpressing mGlu3, and on post-mortem human brain tissue. The most successful antibody was also tested in immunohistochemistry and immunocytochemistry assays. Results. One antibody, ab166608, produced two bands at the predicted molecular weight (~90–100 kDa for the monomer and ~200 kDa for the dimer) in wild-type mouse brain which were not seen in the knockout mice. A similar banding pattern was seen in overexpressing HEK293T/17 cells and in human brain. A second antibody, ab188750, also detected the ~200kDa band (but not the smaller band) in wild type mouse brain, HEK293T/17 cells and human brain; this antibody also produced two smaller, non-specific bands. Three antibodies (sc-47137, sc-47139 and ab10309) did not produce bands at the expected molecular weight but they did produce several smaller bands in both wild-type and double knock-out mice. The final antibody, sc-47138, produced a band at ~90 kDa, as well as one at ~50 kDa, both of which were also present in double knock-out mice. Antibody ab166608 proved to be useful for immunocytochemistry in overexpressing cells, but not in brain sections. Conclusion. Out of six commercially available antibodies we tested, only one (ab166608) detected both monomeric and dimeric mGlu3 and did not show additional non-specific bands. Four of the antibodies did not detect mGlu3 at all. These results highlight the value of the knock-out mouse tissue for antibody validation, and the need for careful characterisation to avoid false negative and false positive findings when antibodies are used as investigational tools in neuroscience and psychopharmacology. Funding. Aintzane García-Bea is funded by a Fellowship from the Fundación Alfonso Martín Escudero. This work is supported by Wellcome Trust Strategic Award (102616/Z) and the Medical Research Council (G0801747). A42 ABSTRACTS

A34 METABOTROPIC GLUTAMATE 3 RECEPTOR (MGLU3) IN SCHIZOPHRENIA: A SEMI-QUANTITATIVE WESTERN BLOT STUDY García-Bea A, Psychiatry, Univ of Oxford, Neurosciences Building Warneford Hospital, Oxford OX3 7JX [email protected] Lane TA(1), Walker MA(1), Harrison PJ(2), Kleinman JE(3), Hyde TM(4) (1) As presenting author; (2) Dept of Psychiatry, Univ of Oxford, Oxford, UK and Oxford Health NHS Foundation Trust, Warneford Hospital, Oxford, UK; (3) Lieber Inst for Brain Development, Baltimore, USA; (4) Lieber Inst for Brain Development, Baltimore, USA and Depts of Neurology and Psychiatry & Behavioral Sciences, John Hopkins School of Medicine, Baltimore, USA Introduction. Metabotropic glutamate receptor 3 (mGlu3), encoded by GRM3, is a risk gene for schizophrenia and a possible therapeutic target for this disorder. A lack of commercially available, well characterised, subtype specific antibodies has made it difficult to determine whether mGlu3 immunoreactivity is altered in schizophrenia. Here we used a validated antibody to examine mGlu3 in a relatively large sample of subjects with schizophrenia and healthy controls. The link between GRM3 risk genotype and mGlu3 immunoreactivity was also examined. Methods. A semi-quantitative western blot study was performed using an antibody to the C-terminus of mGlu3 (ab166608). Total cellular membrane protein was isolated from the superior temporal cortex of 70 patients with schizophrenia and 87 control subjects, who were genotyped for GRM3 SNP rs10234440. The antibody recognized two bands corresponding to the monomeric and dimeric forms of the receptor (~90–100 kDa and ~200 kDa, respectively). The samples were run in quadruplicate, in parallel with a pooled standard. Beta-actin was used as a loading control. mGlu3 monomer and dimer bands were measured, and samples were normalised to the internal standard. Results. mGlu3 immunoreactivity correlated negatively with age at death and post-mortem interval (PMI) and positively with brain pH. These variables differed between groups so a sub-sample of 63 controls and 46 patients in which these variables were matched was selected. Neither mGlu3 dimer, nor monomer, nor the ratio of dimer to monomer differed, between cases and controls (p=0.11, p=0.51 and p=0.77, respectively). Beta-actin also showed no difference between groups. mGlu3 immunoreactivity was not related to rs10234440 genotype; neither did we find any genotype-by-diagnosis interactions. Conclusion. This study demonstrates that there is no difference in mGlu3 abundance in membranes from the superior temporal cortex of subjects with schizophrenic compared with healthy individuals. The negative conclusion is robust given the use of a validated antibody in a much larger sample than hitherto tested. However, these results do not rule out the possibility of cell-specific alterations or aberrant receptor signalling in the disorder. Our findings confirm that there is a robust decrease of mGlu3 across adulthood in human temporal cortex. Funding. Aintzane García-Bea is funded by a Fellowship from the Fundación Alfonso Martín Escudero. This work is also supported by Wellcome Trust Strategic Award (102616/Z) and the Medical Research Council (G0801747).

A35 EFFECTS OF ANTIPSYCHOTIC TREATMENT ON MACROPHAGIC ACTIVATION IN THE ADIPOSE TISSUE Calevro AC, Dept. of Psychological Medicine, Kings College London, Inst of Psychiatry, Psychology & Neuroscience, The Maurice Wohl Clinical Neuroscience Inst, Cutcombe Road, London SE5 9RT [email protected] Cotel MC(1), Kapur SK(1), Pariante CP(1), Vernon AV(1), Mondelli VM(1) (1) As presenting author Introduction: Adipose tissue is known to play an important role in the development of systemic inflammation and mitochondrial dysfunction which can lead to insulin insensitivity and metabolic dysregulation typically present in obesity (Bournat JC, Brown CW 2010, Curr Opin Endocrinol Diabetes Obes:17: 446–452). Translocator Protein, TSPO, is considered a marker of neuroinflammation and its density is elevated in activated microglia, the macrophages of the CNS. In the periphery and specifically in the adipose tissue the macrophages display several phenotypic markers and amongst these there is F4/80. In this study we looked ABSTRACTS A43 at the expression of TSPO, F4/80, and inflammatory cytokines in the rat adipose tissue following chronic antipsychotic treatment. Our aim was to look at the presence of inflammation in the periphery as the same rats previously had shown microglial activation in the brain. Methods: We used adipose tissue samples from a group of 24 male Sprague-Dawley rats treated in a previous study with vehicle (n=8), haloperidol (2 mg kg per day, n=8), or olanzapine (10 mg kg per day, n=8), using osmotic minipumps, for 8 weeks. In the current study we assessed the level of pro and anti inflammatory cytokines: tumor necrosis factor α (TNFα), interleukin-1 β (IL1β), interleukin-6 IL-6, interleukin-5 IL-5, interleukin-4 IL-4, interleukin-10 IL-10 and KC/ GRO a CXC chemokine; The assessment of the cytokines was performed with electrochemiluminescence immunoassay MSD Mesoscale. Furthermore, we performed Western Blot analysis using TSPO antibody and later F4/80 antibody, we have employed the Beta-Actin antibody for the housekeeping expression. Afterwards we calculated the density of the bands with ImageJ program. Results: Amongst the cytokines, only IL-6 was significantly up-regulated in olanzapine animals compared to vehicle animals (One-Way ANOVA with Bonferroni post hoc test p<0.05). Interestingly, there was a trend (t-test p=0.06) for higher F4/80 expression in the olanzapine group compared with the vehicle group (mean±SEM: 0.96±0.06 vs 0.80±0.04 respectively). As regards to the TSPO expression, we did not find any significant difference between each condition, either with haloperidol or olanzapine and the vehicle (t-test p=0.32 vs p=0.10, respectively). Conclusions: In this study, our preliminary results appear to suggest a possible increased inflammation and macrophagic activation in the adipose tissue following olanzapine treatment (based on the IL-6 data and on the trend of F4/80), but not following haloperidol treatment. These findings suggest that, olanzapine tends to increase/ activate inflammation in the adipose tissue with possible important consequences in terms of metabolic dysregulation. Financial Sponsorship: nothing to declare.

B01 AGE-RELATED DIFFERENCES IN APPETITIVE TRACE CONDITIONING Cassaday HJ, School of Psychology, Univ of Nottingham, Univ Park Nottingham NG72RD [email protected] Marshall HJ(1), Pezze MA(1) (1) As presenting author Introduction: We have recently shown medial prefrontal cortex dopamine D1 receptor modulation of the ability to condition over a trace interval (Pezze et al., 2015, Psychopharmacology, 232, 2669–2680). However, these results were obtained testing young adult rats and using a novel refined appetitive trace conditioning procedure. It has yet to be established whether appetitive trace conditioning shows sensitivity to age-related decline, as is known to be the case for aversive trace conditioning. We therefore examined appetitive trace conditioning in a longitudinal study over the course of ageing in untreated rats. The specific ages of interest are 12–18 months (i.e. early to late middle age, prior to neuropathological changes). Methods: Two matched cohorts, aged to 2 versus 12 months, each comprised 24 male Wistar Han rats. Rats were conditioned on two consecutive days at 2 versus 10s trace intervals, at the same 6 week timepoints, up to 8 and 18 months of age. Nosepoking during CS presentations was used to track age-related decline in the ability to condition noise CS with food US over a trace interval. We compared responding in the inter- trial-interval (ITI) and when the food US was delivered to distinguish non-specific motor and motivational effects of ageing. Within the longer (10s) trace interval, we also examined age-related changes in the distribution of responding which may reflect changes in timing ability. Results: Older rats showed initially depressed CS responding followed by a greater increase in CS responding from timepoint 1 to timepoint 5, as compared to the younger cohort, F(4,164)=3.495, p=0.009. In contrast, levels of ITI responding dropped and US responding was maintained in the older cohort. Overall conditioning improved in both cohorts as the study progressed, at the 2s but not at the 10s trace interval, F(4,164)=13.673, p<0.001. A further four-way interaction, F(4,164)=3.654, p=0.007, appears to relate to older rats’ reduced improvement from one day to the next, at the early timepoint measures of CS responding at the 2s trace interval. At later timepoints, responding within the 10s trace progressively distributed towards the end of the trace, in the younger but not the older rats, F(16,352)=1.735, p=0.039. Conclusions: Thus preliminary analyses are consistent with some age-related impairment in appetitive trace conditioning and the findings do not seem to be attributable to non-specific motor or motivational changes. However, over the course of the A44 ABSTRACTS

5 timepoints rats of both age groups showed overall improved conditioning at the 2s trace interval. This work was supported by the BBSRC (ref. BB/K004980/1).

B02 EFFECTS OF BLOCKADE OF NMDA RECEPTORS AND INHIBITION OF SOLUBLE GUANYLATE-CYCLASE ON - ABSTINENCE- INDUCED ANXIETY Padovan CM, Psychology, FFCLRP - Univ of São Paulo, Avenida Bandeirantes, 3900, Ribeirão Preto, SP, Brazil, 14040901 [email protected] Viequetin TA(1), Tirapelli CR(2) (1) As presenting author; (2) EERP-USP, Avenida Bandeirantes, 3900, Ribeirão Preto, SP, Brazil, 14040902 Chronic ethanol consumption (CEC) leads to several diseases and abstinence to ethanol withdrawal induces symptoms of anxiety and/or depression. Nitric oxide (NO) levels are increased during abstinence and the Dorsal Raphe Nucleus (DRN) has increased levels of the enzyme Nitric Oxide Synthase (NOS). Activation of glutamatergic NMDA receptors (NMDAr) increase NO production, which in turn can activate the enzyme soluble guanylate cyclase (sGC), mediating the glutamate-induced enhancement of cyclic GMP (cGMP) levels. However, in the DRN NO- synthesizing cells are mostly co-localized within serotonergic neurons. Therefore, the aim of this work was to investigate the involvement of NMDAr and the NO/cGMP pathway in the DRN on ethanol withdrawal-induced anxiety Male wistar rats with cannulas aimed at the DRN were submitted to different experimental protocols. In the first one, they received an intra-NDR injection (0.1; 0.3 or 1 nmol/0.2µL) of AP7, a NMDA antagonist dissolved in sterile saline or ODQ, a sGC inhibitor dissolved in Vehicle (DMSO 5%, 15%, or 50% in sterile saline). Five minutes after injection, rats were submitted to the EPM test. Enclosed arms entries (EAE) and percentages of entries (OAE) and time spent (TOA) in the open arms were analyzed by ONEWAY ANOVA followed by Duncan test. Significance was set at p<0.05. In another experiment, rats received a 6% (v/v) ethanol or water solution to drink for 21 days. Then, they received only water for 48hrs until test in the EPM. Five min before test, animals received an intra-NDR injection of 0.1nmol/0.2µL of AP7 (or saline), or ODQ (or vehicle). Data was analyzed by two-way ANOVA considering drink (DRINK) and intra-NDR treatment (DRUG) as factors. AP7 did not change exploratory activity in naive rats in any dose used (EAE:F3,34=1.7; p>0.05; OAE:F3,34=1.0; p>0.05; TOA:F3,34=0.7; p>0.05). However, DRUG effects were observed on OAE (F1,43=15,0; p<0.05) and TOA (F1,43=7,03; p<0.05), but not on EAE (F1,43=0.38; p>0.05). DRINK effects were observed on EAE (F1,43=5.56; p<0.05), but not on OAE (F1,43=3.2; p>0.05) nor TOA (F1,43=0.4; p>0.05). Interactions were observed only on TOA (F1,43=4.9; p<0.05). Different doses of ODQ did not change EAE (F6,61=0.1; p>0.05), OAE(F6,61=0,13; p>0.05) nor TOA (F6,61=0.36; p>0.05). However, TOA was affected in rats in abstinence to ethanol (DRINK:F1,36=4.5; p<005). No effects of DRUG nor interaction were observed on EAE, OAE or TOA (p>0.05). Our data suggest that activation of NMDAr is involved on the abstinence-induced anxiety. However, it does not seem to involve NO/cGMP through activatation of sGC. Financial Support: FAPESP (2012/01599-0).

B03 INVESTIGATION OF THE DISCRIMINATIVE AND REINFORCING PROPERTIES OF THE K-OPIOID RECEPTOR AGONIST CR845 IN RATS Smith SL, RenaSci Ltd, BioCity, Nottingham, NG1 1GF [email protected] Spencer R(1), Menzaghi F(1), Gosden J(2), Slater N(2), Holland SJ(2), Slade J(2), Heal DJ(2) (1) Cara Therapeutics, Inc., 1 Parrott Drive, Shelton CT 06484; (2) RenaSci Ltd, BioCity, Nottingham NG1 1GF CR845 is a peripherally-acting κ-opioid receptor agonist being developed to treat pain and pruritus. CR845 has no activity at µ- or δ-opioid receptors. These studies investigated whether CR845 generalised to the discriminative cue elicited by the centrally acting mixed κ/σ-receptor agonist and μ-receptor partial agonist, (-)pentazocine or substituted as a positive reinforcer in an intravenous self administration (IVSA) experiment in heroin-maintained rats. Lister-Hooded female rats were trained to discriminate (-) pentazocine (5mg/kg ip) from saline. Butorphanol, a κ/µ opioid agonist, was selected as the reference comparator. Test drugs were tested 15min after iv injection. Results are reported as mean±SD percentage ABSTRACTS A45 generalisation to the (-)pentazocine cue. Sprague-Dawley, male, rats were trained to self administer heroin (0.015mg/kg/inj) on a FR5 schedule. After saline extinction, CR845 and (-)pentazocine were evaluated in separate groups of rats in 2 hour sessions. IVSA results are reported as mean±SEM. Drug-discrimination was validated by the dose dependent generalisation of iv (-)pentazocine (0.017 0.5 mg/kg,iv) to the training cue (17.7±8.9% to 75.8±8.2% [n=6]). The reference comparator, butorphanol (0.001–0.25 mg/kg iv), dose- dependently generalised to (-)pentazocine (17.1±11.7% to 76.6±17.5% [n=6/7]). CR845 (0.05, 0.125, 0.25, 0.5 mg/kg iv) generalised to saline at the lowest dose, (23.6±10.5% [n=7]) and partially generalised to (-) pentazocine at all other doses (35.4±20.2% [n=7], 31.0±17.9% [n=7], 34.5±10.1% [n=7], respectively) with no evidence of dose-dependence. Heroin-maintained IVSA in all rats (15.83±0.85 inj/session, n=13) that was significantly greater (p<0.001) than saline (3.65±0.37 inj/session, n=13). None of the doses of CR845 (0.001, 0.005, 0.025 or 0.125mg/kg/inj) maintained rates of IVSA at a level significantly greater than saline and all doses were significantly lower than heroin. (-)Pentazocine (0.03, 0.1 or 0.245mg/kg/inj) maintained IVSA at significantly greater levels than saline. (-)Pentazocine and butorphanol both generalised fully to the cue elicited by ip (-)pentazocine validating the model for detecting drugs with µ- and κ-agonist properties. CR845 produced low-level, non-dose-dependent, partial generalisation to (-)pentazocine. This result is consistent with the fact that CR845 is a potent κ-opioid receptor agonist with poor brain penetration. CR845 did not serve as a positive reinforcer in heroin-maintained rats. The reference comparator, (-)pentazocine, substituted for heroin in the model and served as a positive reinforcer across a range of doses. If these results translate into man, they predict that CR845 is unlikely to be recreationally abused by humans. Study funded by Cara Therapeutics Inc.

B04 DETERMINATION OF THE CONCENTRATIONS OF D-AMPHETAMINE, NEUROTRANSMITTERS AND VARIOUS METABOLITES IN MICRODIALYSATES TAKEN FROM THE BRAINS OF FREELY-MOVING RATS Kulkarni RS, Microdialysis Group, RenaSci Ltd, BioCity, Nottingham, NG1 1GF [email protected] vanDam M(1), Mascher H(1), Mascher D(1), Pinder L(2), Rowley HL(2), Heal DJ(2), Cheetham SC(2) (1) pharm-analyt Labor GmbH Ferdinand-Pichler-Gasse 2, 2500 Baden; (2) RenaSci Ltd, BioCity, Nottingham NG1 1GF In a previous study, we have demonstrated that d-amphetamine-induced increases in striatal dopamine efflux measured by microdialysis in freely-moving rats were correlated with the plasma concentrations of the drug (Rowley et al, 2012, Neuropharmacology 63:1064–74). In this investigation, we have taken this technology one step further by exploring the relationship between d-amphetamine-induced changes in the efflux of dopamine and its metabolites in nucleus accumbens CB)(A and acetylcholine in the frontal cortex (FC) and the concentration of d amphetamine in the same microdialysate sample. Two 2.0mm microdialysis probes were stereotaxically implanted into the ACB (AP +2.2mm, ML ±1.5mm, DV -8.0mm relative to bregma) and FC (AP +3.2mm, ML ±2.5mm, DV 4.0mm) of anaesthetised male, Sprague Dawley rats (~300–350g). After ≥16hr recovery, 20min microdialysate samples (1.2µl/min artificial CSF [aCSF] or aCSF+1.0μM neostigmine) were taken from freely-moving rats for 3hr after d-amphetamine dosing. Dopamine, DOPAC and HVA were measured by ALEXYS™ hplc-ecd and acetylcholine by ALEXYS™ uhplc-ecd (Antec). d-Amphetamine (amphetamine-D5 internal standard) was measured in 2.0μl samples by hplc-MS/MS with ESI positive ionization (Shimadzu, AB Sciex API 5000). All results are presented as mean±SE for n=3–8. d-Amphetamine (0.5mg/kg sc) produced rapid increases in the efflux of dopamine in ACB with a peak increase at 40min of 53.8±9.7fmol/5μl (476% baseline [p<0.001]). There were concomitant decreases in DOPAC and HVA with maximum falls to 944±47fmol/5μl (41% baseline at 60min [p<0.001]) and 779±29fmol/5μl (66% baseline at 80min [p<0.001]), respectively. Concentrations of d-amphetamine in ACB microdialysates followed an identical pattern with a peak increase of 44.8±12.9ηg/ml at 40min. In the FC, d-amphetamine produced a more gradual increase of acetylcholine efflux with a peak increase at 60min of 462±54fmol/10μl (389% baseline [p<0.001]). The time-course and magnitude of d-amphetamine concentrations in FC microdialysates were very similar to those observed in ACB with a peak increase of 32.5±5.9ηg/ml at 40min. d-Amphetamine concentrations were highly correlated with both the magnitude of increases in ACB dopamine (r2=0.967) and FC acetylcholine (r2=0.879). In summary, we have demonstrated A46 ABSTRACTS that it is feasible to measure the concentration of d-amphetamine in intracerebral microdialysate samples as well as neurotransmitters and their metabolites. These microdialysis experiments have demonstrated that the actions of d-amphetamine on the efflux of dopamine in theCB A and acetylcholine in FC are highly correlated with the concentration of the drug in the extracellular fluid surrounding the sampling sites. No financial sponsorship.

B05 THE EFFECTS OF Δ9-TETRAHYDROCANNABIVARIN ON FEEDING BEHAVIOURS OF LEAN, HIGH-FAT AND HIGH-SUGAR DIET-INDUCED OBESE RATS Carr SM, School of Psychology & Clinical Language Sciences, Univ of Reading, Harry Pitt Building, Earley Gate, Reading RG6 7BE [email protected] Carr SM(1), Brierley D(1), Williams CM(1), McCabe C(1) (1) As presenting author Introduction: Development of obesity treatments is emerging from research on caloric intake and how the body regulates feeding. How the endogenous system (ECS), which is central to feeding regulation, can be modulated for treatment development needs further exploration. The ECS has been shown to induce weight loss when stimulated by antagonist cannabinoids, blocking the activation of cannabinoid type 1 (CB1) receptors which supresses appetite (hypophagia) (Farrimond, et al., 2011, Phytotherapy Research, 25 (2), 170–188). The phytocannabinoid Δ9-tetrahydrocannabivarin (Δ9-THCV) acts as a neutral antagonist at CB1 receptors, and is thought to hold anti-obesity metabolic properties including hypophagia. How it effects high sugar and high fat food intake is yet unknown. This study aimed to test the effects of Δ9-THCV on feeding behaviour of diet-induced obese (DIO) rats. Method: 24 adult, male Lister hooded rats were split into two groups. Weight gain was induced in one group through a high sugar mash (HS) plus chow diet, and the other through a high fat mash (HF) plus chow diet. Rats were fed these controlled diets for seven weeks prior to testing. During testing purified-Δ9-THCV was given at doses of 0.2, 1.0 and 5.0mg/kg, plus a vehicle control (0.0mg/kg). Each rat received each dose once over the two week testing period, using a within-subjects design. Critical meal parameters included total food intake, latency to onset, duration of feeding and time to satiety. Drugs were administered orally, followed by one hour of food deprivation. Subjects were placed individually into test cages at the start of the dark phase for 240 minutes. A CCTV camera was fitted above each and the same diet was available; HS/HF mash plus chow. After testing the remaining food mass was weighed and videos were coded for different feeding states. Results: No significant dose effects were found for any of our measures of feeding behaviour. Between- groups, weight gain was significantly greater for the HF group; t(22)=1.953, p=.011. Post-mortem analyses showed that animals were ‘overweight to obese’ (BMI scores; 0.70–0.83 g/cm², M=0.76). Findings contribute towards developing a more standardised DIO procedure and measure of weight gain in rats, showing that DIO can be induced at a relatively fast pace without a control group for comparison. Conclusion: Further studies are needed to clarify whether findings could have been observed if Δ9-THCV had been given more frequently, with analysis of its effects on activity levels (energy expenditure) in addition to feeding behaviour. Sources of Financial Sponsorship: Univ of Reading, British Association for Psychopharmacology In-vivo Award 2015.

B06 PLASTICITY MECHANISMS IN THE NUCLEUS ACCUMBENS UNDERLYING BEHAVIORAL SENSITIZATION TO PSYCHOMOTOR STIMULANTS Hussein AN, Dept of Neuroscience, Psychology and Behaviour, Univ of Leicester, Lancaster Road, Leicester LE1 9HN [email protected] Canales JJ(1), Young AM(1) (1) As presenting author Rationale: Behavioral sensitization, measured as the progressive, long lasting increase in locomotor activity, occurs with repeated administration of psychostimulant drugs such as amphetamine and ABSTRACTS A47 nicotine. This phenomenon is thought to result from neuroadaptive changes occurring in the mesolimbic dopaminergic pathway projecting from ventral tegmental area (VTA) to nucleus accumbens (NAc). The acute action of psychostimulant drugs in VTA is an important cellular event for initiation of behavioral sensitization while behavioral expression occurring after repeated drug exposure is mediated by adaptations in NAc neurons and extended circuits (Kalivas, P and Stewart, J. 1991, Brain Research Reviews, 16(3), pp. 223–244). Objective: This study investigated whether sensitization to amphetamine or nicotine was accompanied by increased activation of the immediate early-gene proteins, c-Fos and activity-regulated cytoskeleton-associated protein (Arc), and the chromatin-associated protein, methy1- CpG binding protein 2 (MECP2), in NAc shell and core, since these markers have been implicated in synaptic plasticity and addiction-related physiological adaptations Methods: Male Lister-hooded rats received five daily injection of either amphetamine (1 mg/kg, i.p.), nicotine (0.6 mg/kg, i.p.) or saline (1 ml/ kg, i.p.) and locomotor activity (distance travelled) was recorded for 60 min after drug injection. Ten days later, rats were challenged with same dose of amphetamine or nicotine and their locomotor activity was measured. Two hours later, rats’ brains were processed for immunocytochemistry with antibodies raised against c-Fos, Arc and Mecp2. Significant differences were identified using appropriate analysis of ANOVA followed by post hoc (Bonferroni, p < 0.05). Results: Rats pretreated with nicotine exhibited increased locomotor activity, which showed significant sensitization, both through the repeated daily drug treatment and when challenged ten days later with nicotine. Accompanying this behavioral sensitization, there was an increase in MECP2 protein expression in NAc shell and core, but no significant changes in either c-Fos or Arc. Although we saw a robust increase in activity following amphetamine injection, there was no significant sensitization of the response over consecutive treatment days, nor on challenge day. In addition, we observed no evidence of behavioral cross-sensitisation between the two drugs, nor any effects on c-fos, Arc or Mecp expression after repeated amphetamine treatment. Conclusions: Behavioral sensitization to repeated nicotine injection was accompanied by increases in MECP2 expression in NAc suggesting that nicotine sensitization may be linked to complex epigenetic regulation within the NAc. Surprisingly, there was no evidence of behavioral sensitization to amphetamine, nor evidence of cross-sensitisation between the two drugs. This study was supported by an Iraqi government studentship.

B07 FRACTIONATING STRATEGIES FOR REWARD-GUIDED LEARNING IN MICE Reveles Jensen K, Psychiatry, Oxford Univ, Neurosciences Bldg, Warneford Hosp, Oxford OX3 7JX [email protected] Huber A(1), Korn C(1), Tunbridge E(1), Walton M(2), Akam T(3) (1) As presenting author; (2) Dept. of Experimental Psychology, Univ of Oxford; (3) Dept. of Experimental Psychology, Univ of Oxford, Oxford, England and Champalimaud Centre for the Unknown, , Introduction: How do we learn about the consequence of our actions to select those that lead to positive outcomes, while avoiding negative outcomes? There are two evaluation strategies grounded in reinforcement learning (RL), which have distinct strengths and weaknesses regarding computational demand and statistical efficiency: model-free (low demand and high efficiency) and model-based (high demand and low efficiency). A two-step probabilistic decision-making task can distinguish these models in humans (Daw et al. 2011, Neuron, 69, 6), which we had previously adapted for use with mice using water restriction and rewards. We adapted the task for use with sucrose-reinforcement, since this is more typically used for behavioural studies in the UK. Methods: Male mice (n=7; C57BL/6) maintained at 85–90% of their free-feeding weight were tested at 11–24 weeks of age (6 days/week during the light-phase of a 12-hour light/dark-cycle). Testing was conducted in custom-made Plexiglass cages with four nose- poke holes controlled by custom Python scripts. Mice make a series of nose-pokes to receive probabilistic rewards (2–10ul of 10–20% w/v sucrose). To assess the feasibility of using sucrose reward for the paradigm, we evaluated task activity (rewards obtained per hour across sessions) using linear regression with data previously obtained on water-deprivation (n=6; C57BL/6). Results: With task modifications, mice were able to acquire the task using sucrose reinforcement, although activity remained lower than in water- reinforced mice. Linear regression of the activity during the training period for water- and sucrose reinforcement showed equal increases in activity during training of 2.2 rewards/hour/day (beta). Yet, A48 ABSTRACTS the water-reinforced mice had a significantly higher overall activity of 91 rewards/hour/day (alpha) compared to sucrose-reinforced activity levels of 19 rewards/hour/day (alpha, P<0.0001). To compensate for the decreased activity in sucrose-reinforced mice, the session time was increased to 2.5 hours (from 1.5 hours). With this modification, training was equal after 23 days on water restriction and 39 days of food restriction. Initial results show that mice on sucrose-reinforcement are less model-based then with water-reinforcement. Conclusions: We have successfully adapted the two-step probabilistic RL task for use with sucrose reward. We extended training session length and added additional training days, given the lower activity in sucrose-reinforced mice. The modified task will provide a useful tool to assess the impact of genetic, environmental and pharmacological manipulation on model-free/model-based RL in mouse models. Financial sponsorship: K. Jensen is supported by The Oticon Foundation, The Augustinus Foundation and The Medical Society in Copenhagen. M. Walton, T. Akam and C. Korn are supported by the Wellcome Trust. A. Huber is supported by the BBSRC and E. Tunbridge by The Royal Society.

B08 BRAIN VOLUMETRIC CHANGES FOLLOWING CHRONIC KETAMINE EXPOSURE IN BOTH HUMANS AND MICE: SIMILARITIES AND DIFFERENCES Chesters RA, Dept of Basic and Clinical Neuroscience, Inst of Psychiatry, Psychology and Neuroscience, Maurice Wohl Clinical Neuroscience Inst Kings College London, 5 Cutcombe Road London SE5 9RT [email protected] Cooper JD(1), Vernon AC(1), Hughes E(2), Howes O(3), Stone JM(4) (1) As presenting author; (2) Imperial College London, Imaging Sciences Unit, Robert Steiner MRI Unit, Hammersmith Hospital, London W12 0NN, UK; (3) Imperial College London, MRC Clinical Sciences Centre, Hammersmith Hospital Campus, Du Cane Road, London W12 0NN, UK; (4) King’s College London, Inst of Psychiatry, Psychology and Neuroscience, Dept of Neuroimaging, Centre for Neuroimaging Sciences, De Crespigny Park, London SE5 8AF, UK Introduction: Ketamine is an increasingly popular drug of abuse amongst young adults, particularly in the UK and China. Chronic users experience deficits in cognition and spatial memory, alongside sub threshold psychotic symptoms [Morgan, C.J.A, Curran H.V., 2012. Ketamine use: A review. Addiction 107, 27–38]. The mechanisms driving these effects are unclear. Structural magnetic resonance imaging (MRI) studies suggest that chronic ketamine use induces brain atrophy, particularly in the frontal lobe [Liao, Y., et al, 2011. Reduced dorsal prefrontal grey matter after chronic ketamine use. Biological Psychiatry 69, 42–48]. However, these data are confounded by poor matching between subjects, particularly on measures of drug use, making it difficult to identify the specific effects of ketamine. In contrast, rodent MRI studies offer precise control of such variables and permit invasive post-mortem measurements to link neuroimaging to neuropathology. Purpose of the study: To compare structural MRI differences between human ketamine users and poly- drug controls and compare these data to structural MRI data from two mouse models of chronic ketamine administration, to identify the specific effects of ketamine on brain macrostructure. Methods: T1-weighted 3D-MPRAGE MR images were acquired from chronic ketamine users (n=14) and poly-drug control subjects (n=13). Group-level differences in grey matter volume were assessed using Freesurfer Image Analysis Suite (www.freesurfer.net) and voxel-based morphometry (VBM) using FSL (www.fmrib.ox.ac.uk/fsl). In parallel, male C57/Bl6 mice (aged 11–12 weeks) were exposed to ketamine (20 mg/kg, i.p.) (n=15) or saline (n=15) intermittently (once per day every 3 days for 30 days) or daily (once daily for 14 days). 3D T2-weighted ex vivo MR images were acquired and group-level differences in brain morphometry analysed using brain- wide tensor-based morphometry (TBM). Results: Freesurfer analysis identified a 17.5% reduction in grey matter volume in the right paracentral lobule of ketamine users (q<0.05; FDR-corrected), as well as 10–20% volumetric reductions in regions of the frontal, temporal and parietal cortices. FSLVBM analysis confirmed volumetric reductions in the right paracentral lobule, as well as regions of the frontal, temporal and parietal cortices (q<0.05; FDR-corrected). In mice administered ketamine daily, TBM analysis revealed a trend-level decrease in volume in the somatosensory cortex (p≤0.05). TBM analysis of mice administered ketamine every three days revealed significant volume increases in the prefrontal cortex, somatosensory cortex and hippocampus, coupled with decreased volume of white matter structures including the corpus callosum, anterior cingulate, and external capsule (q<0.05; FDR-corrected). Conclusion: These data provide preliminary ABSTRACTS A49 evidence to suggest there are both similarities and differences in the effects of ketamine on brain structure in humans and mice, depending on the frequency of exposure. Studies are now underway to map the cellular correlates of these changes in the mouse models. Funding: We would like to acknowledge the Medical Research Council, NIHR Maudsley Biomedical Research Centre, King’s Health Partners and Guy’s and St Thomas’ Charitable Trust for helping fund this study.

B09 THE ROLE OF THE SEROTONIN 2A RECEPTOR IN THE FABRIC AND MODULATION OF PERSONAL MEANING IN LSD-INDUCED STATES Preller KH, Dept of Psychiatry, Psychotherapy and Psychosomatics, Univ Hospital for Psychiatry Zurich, Lenggstr. 31 8032 Zurich Switzerland 8032 [email protected] Herdener M(1), Pokorny T(1), Planzer A(1), Kraehenmann R(1), Staempfli P(1), Seifritz E(1), Vollenweider FX(1) (1) As presenting author Introduction: Abnormalities in the attribution of personal relevance to stimuli are critical features of many psychiatric disorders such as schizophrenia, addiction, and mood disorders. However, the neuronal substrates enabling meaningful and personally relevant experiences are largely unknown. Lysergic acid diethylamide (LSD) is a prototypical hallucinogen which has high affinity at serotonin (5-HT)-2A/C, -1A/B, -6, and -7, and dopamine D2/D1 receptors (R). Animal studies suggest that LSD produces its psychedelic effects primarily via agonist action at 5-HT2AR although D2R may also be implicated. A core feature of LSD-induced psychedelic states is an increased sense of meaning of objects. Ketanserin is a selective 5-HT2AR antagonist which may in combination with LSD and functional magnetic resonance imaging (fMRI) offer the opportunity to determine the role of the 5-HT2AR system in personal meaning processing. Methods: This double-blind, randomized, counterbalanced, cross-over study assessed the neural and behavioral response to personally meaningful, neutral, and personally meaningless music using fMRI and behavioral ratings. Twenty-two healthy participants received either 1) placebo+placebo (Pla condition) 2) placebo + LSD (100 µg po, LSD condition), or 3) ketanserin (40 mg po) +LSD (100 µg po, Ket+LSD condition) at three different occasions. Results: Meaningfulness ratings were increased for meaningless and neutral music under LSD compared to both Pla and Ket+LSD (all p<0.05). Furthermore, listening to meaningless music compared to neutral and personally meaningful music was associated with increased BOLD signal in medial and lateral frontal brain areas in the LSD condition compared to both Pla and Ket+LSD conditions. (all p<0.05, FWE corrected). Furthermore, subjective LSD effects were fully blocked by ketanserin. Conclusions: LSD increased the attribution of personal relevance to previously meaningless stimuli. This effect appears to be attributable to 5-HT2AR stimulation, since these alterations are normalized after pretreatment with ketanserin. Furthermore, surprisingly virtually all LSD-induced psychological effects were blocked by 5-HT2AR antagonism, despite LSD’s additional action on dopamine receptors. The current results emphasize the pivotal role of the 5-HT2AR in the generation of personal meaning in medial and lateral frontal brains structures implicated in self-referential processes. These findings may be relevant for increasing our understanding of the biochemical underpinnings of personal meaning processing and may reveal prospective targets in the treatment of psychiatric illnesses characterized by alterations in meaning attribution. Sources of financial sponsorship This study was financially supported by the Swiss Neuromatrix Foundation, the Heffter Research Inst, and the Swiss National Science Foundation.

B10 MISMATCH NEGATIVITY PROCESSING UNDER THE LSD STATE: FEED-FORWARD AND BACKWARD CONNECTIVITY Timmermann CB, Neuropsychopharmacology, Imperial College London, Burlington Danes Bldg, Hammersmith Hospital Campus London W12 0NN [email protected] Fielding A(1), Kaelen M(2), Roseman L(2), Carhart-Harris R(2), Nutt D(2), Muthukumaraswamy S(3) (1) Beckley Foundation; (2) Dept Neurpsychopharmacology, ICL; (3) Dept Psychology, Uni Auckland The Mismatch Negativity (MMN) and other event-related potentials/fields (ERP/ERF) have been used as diagnostic tools to assess consciousness. Previous research has provided intriguing results on how A50 ABSTRACTS top-down connectivity in MMN paradigms may provide a more reliable marker to index the normal conscious state, while being impaired in conditions of reduced awareness. LSD is known for altering consciousness in a significant manner, without reducing wakefulness or awareness at moderate doses. In this study, a placebo-controlled design was used to study the effects LSD has on top-down and feed- forward processes as indexed by the MMN. A balanced order, within subject design was used for the study. Healthy volunteers underwent MEG recordings under placebo and LSD conditions at least 2 weeks apart. Participants were presented with auditory stimuli consisting of oddball and standard tones. Following preprocessing, data was averaged and the resulting event-related fields were converted into scalp-map images corresponding to each of the four conditions: auditory stimuli of standards under LSD (1) and placebo (2) and deviants under LSD (3) and placebo (4). The ERFs were entered into an within-subject analysis of variance with 2 factors: drug (LSD and placebo) and expectation (standard and deviant tones). Results (corrected for multiple comparisons using FWE P< 0.05) show a main effect of expectation at temporal sensors at 53–110ms (p<0.001) and in frontal sensors at 170–217ms (p<0.001). An interaction effect was found in right frontal sensors at a later component at 210ms (p<0.001) These results indicate that while no significant differences were found regarding early components of the MMN, later ones did differ under LSD as compared to placebo. These results may reflect modulation of prediction error under the LSD state, thereby providing insights on the mechanisms underlying the characteristic subjective effects elicited by psychedelic drugs. This research received financial and intellectual support from the Beckley Foundation and was conducted as part of a wider Beckley-Imperial research programme. C.T. is funded by Comisión Nacional de Investigación Cientifica y Tecnológica (CONICYT), Chile.

B11 ACUTE MDMA ADMINISTRATION MODULATES NEURAL AND BEHAVIOURAL MEASURES OF SELF- REFERENTIAL PROCESSING Wall MB, Imanova Ltd,, Burlington Danes Bldg, Imperial College London, Hammersmith Hospital, Du Cane Road, London W12 0NN [email protected] Demetriou L(1), Carhart-Harris R(2), Erritzøe D(2), Nutt DJ(2), Stewart L(3), Curran HV(3), Morgan CJA(4), Ferguson B(5) (1) As presenting author; (2) Centre for Neuropsychopharmacology, Division of Brain Sciences, Imperial College London; (3) Clinical Psychopharmacology Unit, Univ College London; (4) Dept of Psychology, Univ of Exeter; (5) Univ Medical Center, Utrecht Introduction: +3,4-methylenedioxymethamphetamine (MDMA, or ‘ecstasy’) is a serotonergic drug with well- described effects on social behaviour, promoting interaction, openness, and positive emotion. Previous work has shown that self-referential processing can be modulated by serotonergic anti-depressants (Di Simplicio et al., 2012, Molecular Psychiatry 17, 503–510) and this may be mediated by the medial pre-frontal cortex (mPFC). We investigated the effect of MDMA on self- and other-referential processing. Methods: Twenty-three participants completed a self vs. other referential task in two treatment sessions (placebo and MDMA), using a randomised, double-blind, placebo-controlled design. Brain activity was monitored using functional MRI (Siemens 3T scanner, standard fMRI acquisition parameters). The task presented positive (e.g. “caring”) or negative (e.g. “miserable”) adjectives in one of three conditions: 1. Self-referent (“Are you? ADJECTIVE“), 2. Other-referent (“Is Wayne Rooney? ADJECTIVE”), 3. Control (“More than 2 syllables? ADJECTIVE”). The control task was included as a non-social/non-evaluative baseline condition. Blocks of three trials of the same type were presented, each lasting five seconds, and positive and negative valences alternated within each block (counter-balanced across blocks). Blocks were pseudo-randomly ordered to avoid repetition of the same condition, and interspersed with an equal number of rest blocks. There were 48 blocks for a total time of 12 minutes. Imaging data analysis used a standard pre-processing and analysis pipeline, including motion correction, spatial smoothing (6mm) temporal filtering (100s) and analysis with a general linear model. The six trial conditions (positive/negative words, combined with self/other/control tasks) were separately modelled, with temporal derivatives and head-motion parameters also included. Group-level analyses used a statistical threshold of Z=2.3, p<0.05 (cluster-corrected for multiple comparisons). Results: Task effects (regardless of drug treatment) showed robust activation for the evaluative conditions (relative to control) in mPFC, posterior cingulate and lateral parietal regions, consonant with previous studies. Analysis of drug-treatment effects ABSTRACTS A51 showed a robust and specific effect in the mPFC (plus the left insula and hippocampus), showing that MDMA significantly reduces self-referential processing in these areas. ROI analysis of the mPFC confirmed this effect is driven by reductions on negative (t[22]=2.992, p=0.007) rather than positive (t[22]=0.934, p=0.361) self-evaluation trials. Drug effects on other-evaluation trials were negligible. Conclusions: MDMA specifically reduces brain activity related to negative self-referential processing in the mPFC. This suggests a potential novel clinical use as an adjunct in the treatment of disorders that have a significant negative self-referential component. This study was funded by the Channel 4 Television Corporation.

B12 DRINKING TO REMEMBER: THE RETROGRADE FACILITATION OF MEMORY EFFECT IN A NATURALISTIC SETTING McAndrew A, Psychology, Univ of Exeter, College of Life and Environmental Sciences, Washington Singer Building, Perry Road, Exeter UK EX4 4QG [email protected] Polkinghorne M(1), Carlyle M(1), Bush G(1), Roberts K(1), Stevens T(1), Morgan C(1) (1) As presenting author Background: Much research has shown the alcohol impairs memory if it is consumed during or before a learning phase. But it has also been found that memory, and in particular recall, can be improved if alcohol is consumed after a learning phase (Parker et al 1980; Mueller, Lisman & Spear, 1983) in what is termed the retrograde facilitation effect. However no research has tested this effect naturalistically or related this to the acute memory impairment. Method: 97 social drinkers were recruited, n= 47 to the “Sober” condition and n=50 to the “Alcohol” condition. Participants were tested in their own homes prior to prinking (pre-drinking) and before a night out on a psycholinguistic based memory consolidation task (Dumay & Gaskell, 2007). The memory consolidation task involves a learning phase, then cued recall component which participants underwent prior to drinking and then ~ 16 hours post drinking, the following day. Participants also completed the Mneomic Separation Task (MST: Stark et al., 2012) when intoxicated and the following day to examine acute memory impairment. Volunteers were breathalysed at several time points across the evening and again the following morning when researchers returned to test memory again. Results: A significant Day X Alcohol Condition interaction was observed (p=0.029) whereby participants in the alcohol group remembered more of the words that they encoded before drinking alcohol on day 2 than they had straight after encoding, whereas this effect was absent in the placebo group. There was a positive correlation between memory performance on the MST and day 2 performance in the sober group (0.52, p<0.001) but this correlation was absent in the alcohol group. Discussion: This study provides the first demonstration of the retrograde facilitation effect in a naturalistic setting with better performance in the group who self-administered alcohol following memory encoding. Results are discussed in relation to the opportunistic theories of hippocampal memory consolidation. This project was funded by the Univ of Exeter.

B13 STIMULI THAT PREDICT REWARD CAPTURE ATTENTION IN THE ABSENCE OF CONSCIOUS AWARENESS BUT CORRELATE WITH INTEROCEPTION SENSIBILITY Leganes M, CBS Psychology, Sussex Univ, Pevensey Building, Falmer BN1 9QH [email protected] Scott R(1), Duka T(1) (1) As presenting author The existence of unaware conditioning is still an ongoing debate. Studying the mechanisms by which implicit conditioning is learnt and becomes observable can help understand addictive behaviours. According to the Incentive Sensitization Theory of addiction, rewards elicit a discernible “Liking”, an emotional reaction, and “Wanting”, a motivational drive, towards them (Robinson & Berridge, 2001. Addiction, 96(1),103–114). Interoception is a focus of interest in understanding implicit mechanisms underlying affective behaviours, and its role in drug addiction is becoming apparent (Paulus & Stewart, 2014 Neuropharmacology, 76PtB, 342–50). The purpose of this experiment was to study the emotional A52 ABSTRACTS value acquired by CS through implicit emotional and attentional processes, and if Interoceptive Awareness can explain individual differences. Using a similar procedure to the one designed by Yokoyama et al. (2015, Frontiers in Psychology, 6,269) with 27 participants (21 females), stimuli belonging to two different categories (Houses or Buildings) were conditioned with High rewards (80% probability of earning 10p) or Low rewards (20% probability of earning) using a task irrelevant conditioning; participants had to respond with the appropriate key to the colour of a square overlaid to the picture. Expectancy of the reward outcome was measured randomly in 25% of the trials via Likert scale. Pleasantness was measured after the conditioning task. Most of the participants were contingency-unaware (85%) and did not develop a subjective emotional response towards CS+ (p>0.05). On a subsequent Emotional Attentional Blink task using a Rapid-Serial-Visual-Presentation of stimuli, aversive distractors decreased accuracy on the detection of Low reward stimuli but not of High reward stimuli (F(1, 25)= 11.764, p=0.002), even for participants unaware of the contingencies; t(22)=5.01, p<0.001. This bias did not correlate with pleasantness or expectancy scores. Interoception was measured via Heart-beat Tracking and Discrimination tasks. For unaware participants, Interoceptive Awareness correlated positively with pleasantness ratings for High over Low reward stimuli; r=0.45, n=23, p=0.031. This experiment shows that positively CS develop attentional salience irrespective of their subjective pleasantness or expectancy allocation, demonstrating that implicit CS are able to capture attention. Correlates of Interoceptive Awareness with expectancy ratings open a new path to comprehend how individual characteristics determine differences in conditioning processes, bringing new insights towards the understanding of risk factors in drug addiction. This research is funded by a grant provided by Sussex Univ.

B14 MULTI-VOXEL PATTERN ANALYSIS REVEALS SPECIFIC EFFECTS OF TWO DIFFERENT STRAINS OF CANNABIS ON THE BRAIN’S RESTING-STATE NETWORKS Kowalczyk OS, Dept of Psychology, Royal Holloway, Univ of London, Egham Surrey TW20 0EX [email protected] Pope R(1), Freeman T(1), Mokrysz C(1), Hindocha C(1), Lawn W(1), Moss A(1), Curran HV(1), Nutt DJ(2), Bloomfield M(3), Wall MB(4), Demetriou L(4) (1) Clinical Psychopharmacology Unit Research Dept of Clinical, Educational and Univ College London Gower St London WC1E 6BT; (2) Division of Brain Sciences, Imperial College London, London, UK. ; (3) Division of Psychiatry, UCL, London, UK. ; (4) Imanova Ltd. Burlington Danes Building Imperial College London Hammersmith Hospital Du Cane Road London W12 0NN Introduction: Two primary psychoactive constituents in cannabis are delta-9-tetrohydrocannabinol (THC) and cannabidiol (CBD). High-THC strains of cannabis may be associated with negative effects (e.g. addiction, psychosis) while strains with a more balanced content of constituents may be relatively safer. Methods: Two strains of cannabis were tested in a double-blind, placebo-controlled, three-way crossover design (N=17). One strain had 13% THC and <0.1% CBD, the second had 6.5% THC and 8% CBD; double the quantity of the second was administered so that total THC content was similar for both conditions. Effects on the brain were monitored using resting-state fMRI. A comprehensive evaluation of effects across a broad range of networks was conducted by using 1) seed-based analyses to define three major resting-state networks (the default-mode network, the executive control network, and the salience network), and 2) An additional set of 10 standard networks derived from Smith et al. (2009, PNAS 31, 13040–5). Effects of the treatments on these networks were assessed using Multi-Voxel Pattern Analysis (MVPA); a highly sensitive multivariate ‘decoding’ analysis method that uses machine-learning algorithms to discriminate patterns of brain activity between experimental conditions. The PIPR software (Doyle & Joules, Kings College London) was used, with a two-class Gaussian Process Classifier (GPC), and statistical significance was determined by 1000 iterations of a permutation test. Results: In the seed-based networks, the classifier reliably discriminated between both strains of cannabis, compared with placebo (p < 0.04 for the THC+CBD strain, and p < 0.02 for the pure THC strain) in the default-mode network. Good discrimination (82%, p < 0.01) was also seen in the salience network, but only for the pure-THC strain vs. placebo. In the additional set of 10 networks derived from Smith et al. (2009), above- chance classifier performance was seen in this version of the default-mode network for the THC+CBD ABSTRACTS A53 strain (88%, p < 0.02), but also in another network, centred on visual area V5/MT (71%, p < 0.05 for the THC+CBD strain, and 76%, p < 0.03 for the pure-THC strain; both compared to placebo). Conclusions: The neurological effects of cannabis are rather specific, and are focussed on the default-mode and salience networks, with some additional effects on lateral visual areas. Further work on the brain effects of cannabis should focus on these networks and the psychological functions that rely on them. This study was supported by DrugScience.

B15 THE ACUTE EFFECTS OF A DOPAMINE D3 RECEPTOR PREFERRING AGONIST ON MOTIVATION FOR CIGARETTES IN DEPENDENT AND OCCASIONAL CIGARETTE SMOKERS Lawn W, Clinical Psychopharmacology Unit, Univ College London, 1–19 Torrington Place, London WC1E 7HB [email protected] Freeman TP(1), East K(1), Gaule A(1), Das RK(1), Curran HV(1), Aston E(2), Morgan CJA(3), Bloomfield MAP(4) (1) As presenting author; (2) Behavioural and Social Sciences, Brown Univ, Providence; (3) Dept of Psychology, Univ of Exeter, Exeter; (4) Psychiatric Imaging Group, Medical Research Council Clinical Sciences Centre, Hammersmith Hospital, London Introduction: Dopaminergic functioning is thought to play critical roles in both motivational processing and addiction. There is preliminary evidence that dopamine may reduce the motivation for cigarettes in smokers. However, the effects of pramipexole, a dopamine D3-preferring agonist, have not yet been investigated. We aimed to examine the effects of an acute dose of pramipexole on the motivation to earn cigarettes and non-drug rewards in occasional and dependent cigarette smokers. Methods: A double-blind placebo-controlled crossover design assessed the effects of pramipexole (0.5mg oral) in dependent (n=20) and occasional (n=20) cigarette smokers following at least 1h 45m of nicotine deprivation. Motivation for cigarettes and consummatory non-drug rewards was measured using the DReaM-Choice task, in which participants can earn, and then later consume, cigarettes, music and chocolate. Demand for cigarettes was measured using the Cigarette Purchase Task (CPT). Self-reported craving, withdrawal and drug effects were also recorded. Results: Dependent smokers made more choices for (p<0.001) and pressed more (p<0.001) for cigarettes than occasional smokers. Moreover, dependent smokers made fewer choices for chocolate (p<0.001) than occasional smokers. There were null effects of pramipexole on choices for and pressing for all rewards available. Similarly, the dependent smokers had greater demand for cigarettes than occasional smokers across all outcomes of the CPT (ps<0.021), apart from elasticity. Pramipexole had null effects on all outcomes of the CPT. In terms of subjective effects, pramipexole produced greater ‘feel drug’ and ‘dislike drug’ effects than placebo, but pramipexole did not reduce craving and withdrawal symptoms. Conclusions: Dependent and occasional cigarette smokers differed in their motivation for cigarettes, but not consummatory non-drug rewards. There was no evidence to suggest that a low dose of pramipexole acutely altered motivation for cigarettes. These results question the role of dopamine D2/3/4 receptors in nicotine dependence in humans. This research was funded by a BBSRC PhD studentship.

B16 DISSOCIABLE EFFECTS OF CANNABINOIDS ON ANTICIPATORY AND CONSUMMATORY REWARD PROCESSING Freeman TP, Clinical Psychopharmacology Unit, Univ College London, Research Dept of Clinical, Educational and Health Psychology, 1–19 Torrington Place London WC1E 7HB [email protected] Curran HV(1), Pope RA(1), Hindocha C(1), Mokrysz C(1), Lawn W(1), Moss A(1), Nutt DJ(2), Morgan CJA(3), Bloomfield MAP(4), Wall MB(5), Bisby JA(6), Luijten M(7) (1) As presenting author; (2) Centre for Neuropsychopharmacology, Imperial College London, UK; (3) Dept of Psychology, Univ of Exeter, UK; (4) Div of Psychiatry, Univ College London, UK; (5) Imanova Centre for Imaging Sciences, Imperial College London, Hammersmith Hospital, London, UK; (6) Inst of Cognitive A54 ABSTRACTS

Neuroscience, Univ College London, UK; (7) Radboud Univ Nijmegen, Nijmegen, The Netherlands Introduction: Reward processing can be parsed into dissociable components of anticipation (e.g. wanting) and consummation (e.g. liking). The rewarding effects of cannabis may be caused by its primary psychoactive constituent, delta-9-tetrahydrocannabinol (THC). Cannabidiol (CBD) is another cannabis constituent that can inhibit some effects of THC. This study had the following objectives: 1) to investigate the acute effects of cannabis on reward anticipation and consummation, 2) to establish whether these effects are blocked by CBD. Methods: Across 3 sessions, 16 healthy cannabis users inhaled vaporized cannabis preparations containing 8mg THC, 8mg THC + 10mg CBD, and placebo. Reward consummation was indexed using functional Magnetic Resonance Imaging, evidenced by greater BOLD signal whilst listening to classical music versus scrambled sound (Menon and Levitin, 2005, Neuroimage, 28, 175–184). Regions of interest were selected from a meta-analysis of music-evoked emotion, and all results were False Discovery Rate corrected. Reward anticipation was recorded using the visual analogue scale ‘want to listen to music’; post-hoc tests were Bonferroni-corrected. Results: Analysis of consummatory reward showed that cannabis containing THC only reduced activation in bilateral temporal gyrus (right: p=0.005, left: p=0.008), right hippocampus (p=0.025), right amygdala (p=0.025), right insula (p=0.026) and right medial orbitofrontal cortex (OFC, p=0.033). Cannabis containing THC and CBD did not alter signal in any regions. Across all scans, OFC activation correlated with subjective pleasure ratings (r(48)=0.463, p<0.001). Both types of cannabis increased reward anticipation to a similar extent (THC: p=0.001, THC+CBD: p=0.006). Conclusions: Reward anticipation is primed by cannabis, irrespective of whether it contains CBD or not. By contrast, cannabis reduces neural activation to music, and this effect is blocked by CBD. These dissociable effects support a role of the endocannabinoid system in anticipatory and consummatory reward processing. This study was supported by Drugscience.

B17 THE ACUTE EFFECTS OF ALCOHOL ON MEMORY AND SLEEP: DOES SLOW-WAVE SLEEP MEDIATE THE RETROGRADE FACILITATION EFFECT? Carlyle M, Psychology, Univ of Exeter, Washington Singer Labs, College of Life and Environmental Sciences, Perry Rd, Exeter EX4 4QG [email protected] Bush G(1), McAndew A(1), Stevens T(1), Morgan CJA(1) (1) As presenting author Introduction: Alcohol given before learning can impair memory (anterograde amnesia). Paradoxically, alcohol can facilitate memory if given after learning (retrograde facilitation). Studies reporting retrograde facilitation have often comprised a period of sleep after drinking. Since high doses of alcohol can increase the proportion of slow-wave sleep (SWS), and SWS is critical for declarative memory, it was hypothesised that this paradoxical effect of alcohol on memory is a consequence of increased SWS. Method: Forty- eight participants were administered 4 drinks, either alcohol or placebo, in a two-session, double blind, randomised placebo-controlled design. The alcohol condition were administered a total of 50 grams and 39 grams of vodka to men and women, respectively. To assess memory consolidation, participants were required to recall 24 previously learned novel words before drink consumption during session 1, and again the next day during session 2. Acute memory was assessed using a simple pattern separation task completed immediately following intoxication. SWS was measured using a discrete sleep monitoring device worn overnight, accompanied by a continuous transdermal measure of blood alcohol content (SCRAM-X monitoring device) and regular breathalysers. Results: The alcohol condition was significantly more impaired on the acute memory task immediately following intoxication (p<.001). In contrast, the alcohol condition did not recall more novel words than placebo the following day (p=.592), and thus did not exhibit enhanced memory consolidation. However, recall did significantly improve in both conditions (p=.005). Post-hoc correlations revealed that the extent of impaired performance experienced during intoxication in the alcohol condition predicted how many words were recalled the subsequent day (r=- .57, p=.007). The alcohol condition did not experience significantly more WSS relative to placebo (p=.416). Conclusions: These findings support previous research by replicating anterograde amnesia. In addition, post-hoc findings suggest that this alcohol-induced amnesia is related to enhanced consolidation. However, ABSTRACTS A55 they did not reproduce retrograde facilitation or increased proportion of SWS as a result of alcohol, contrasting with previous research. Funding for this project came from the Univ of Exeter.

B18 DIFFERENTIAL STIMULANT AND SEDATIVE EFFECTS OF N2O IN INDIVIDUALS WITH A FAMILIAL VULNERABILITY TO DRINKING PROBLEMS Walsh KH, Clinical, Educational, and Health Psychology, UCL, Clinical Psychopharmacology Unit, 1–19 Torrington Place, London WC1E 7HB [email protected] Das RK(1), Hannaford JR(1), Kamboj SK(1) (1) As presenting author Introduction: A family history of alcohol abuse disorder (AUD) is a significant risk factor for alcoholism. Responses to NMDA receptor antagonists, such as alcohol, are altered in healthy individuals with a family history of AUDs, indicating a possible genetic component of ethanol dependence. Compared to those without a family history, those with a familial vulnerability experience greater rewarding and reduced aversive effects in response to ketamine, a potent, non-competitive NMDAR antagonist (Yoon et al., In Press). Like ketamine, (N2O) is an NMDAR antagonist and thus a similar response to N2O inhalation might be predicted in those with a family history of alcohol problems. Methods: Participants were “heavy drinkers” ( >3/4 alcoholic drinks, three times a week for men/women respectively; AUDIT score >8) *without* a history of AUD. Responses to inhaled 50% N20 (with 50% oxygen; 30 min) were recorded using a visual analogue scale of bodily symptoms (BSS) related to expected N2O and alcohol effects. These were recorded at pre-, during, and post N2O. Individuals with a family history (at least one first degree relative) of alcohol problems (N=23) were compared to those with no family history (N=35) on the euphoria and drowsiness items of the BSS, which corresponded to the stimulant and sedative (‘biphasic’) responses to alcohol respectively. Results: A mixed 2 (time; pre, during N2O) x 2 (item; Drowsiness, Euphoria) x 2 (family history; no family history) ANOVA revealed a significant interaction between time, subscale item, and family history F(1,54)=6.98, p=.01, np2=.115, with Bonferroni-adjusted individual comparisons suggesting both a significant increase in euphoria (p<.05) and a significant decline in drowsiness (p<.01) occurred only in those with a family history of alcohol problems. Pairwise comparisons similarly suggest that while no differences were observed at baseline, a significant difference between groups was recorded on drowsiness scores during N2O administration (p<.05). Conclusions: The primary finding was an attenuated response to the sedative effects of N2O as well as an increased sensitivity to its stimulant effects in those with a family history of alcohol problems. These findings are consistent with findings with ketamine and may reflect the detection of altered NMDAR functioning in those with a genetic vulnerability to alcohol dependence. This research was funded by an MRC project grant to SKK and RKD.

B19 THIS ABSTRACT HAS BEEN WITHDRAWN

B20 THE DIFFERENCE IN SMOKING MOTIVES, HABITS, AND ATTENTIONAL BIAS IN NICOTINE DEPENDENT AND NON-DEPENDENT SMOKERS Bartlett JB, School of Psychological, Social, and Behavioural Sciences,Coventry Univ, Priory Street, Coventry, West Midlands CV1 5FB [email protected] Farias MF(1), Jenks RJ(2), Wilson NW(2) (1) Centre for Research in Psychology, Behaviour and Achievement, Priory Street, Coventry CV1 5FB; (2) Richard Crossman Bldg, School of Psychological, Social, and Behavioural Sciences, Priory Street, Coventry CV1 5FB Introduction: Attentional bias is the automatic allocation of attention to smoking-related stimuli and has been investigated as it may have a role in the maintenance of smoking and hinder cessation attempts. Despite a lack of nicotine dependence, non-dependent smokers have a cessation failure rate comparable to A56 ABSTRACTS dependent smokers. The present study aimed to explore attentional bias as a possible mechanism involved in the maintenance of drug use. Furthermore, the present study aimed to create a comprehensive profile of dependent and non-dependent smokers by measuring their smoking motives and habits. Method: A visual dot probe task with three stimulus onset asynchronies (SOA: 200ms, 500ms and 2000ms) explored different aspects of attention in dependent and non-dependent smokers relative to non-smokers. This methodology was used to provide a comparison to previous literature as there have been inconsistent results. Smoking (dependent n= 11; non-dependent n= 7) and non-smoking (n= 17) Coventry Univ students and staff also completed measures of smoking dependence (FTND and exhaled CO) and smoking motives (WISDM). Results: Dependent smokers scored significantly higher on both objective and subjective measures of nicotine dependence. Dependent smokers reported to be motivated by factors relating to physical dependence such as loss of control and tolerance. On the other hand, the factors most important for non-dependent smokers were cue exposure and the sensory benefits of smoking. For attentional bias, there was no main effect of group (p=0.46) or SOA condition (p=0.66). Moreover, there was no interaction effect between the two (p=0.83). Conclusions: The lack of attentional bias effects across smokers and non-smokers is inconsistent with previous literature. The methodology may have contributed to this lack of difference and the limitations are discussed. However, we found that dependent and non-dependent smokers report that their smoking habit is maintained by different facets of smoking motivation. This may have potential practical implications and suggest that individualised smoking cessation interventions may be more effective for different smoking groups. There was no internal or external funding associated with this study.

B21 A NOVEL BEHAVIOURAL TASK MEASURING MOTIVATION FOR REWARDS WITH INCREASING LEVELS OF COGNITIVE EFFORT Valton V, Inst of , UCL, 17 Queen Square, Alexandra House, London WC1N 3AZ [email protected] Gray A(1), Mkrtchian A(1), Samborska V(1), VanUrk S(1), Roiser JP(1) (1) As presenting author Introduction: Motivation can be defined as the willingness to exert effort in order to attain a preferred outcome (Niv et al. 2007, Psychopharmacology, 191,507–520), which is strongly dependent on the magnitude of anticipated reward and required effort (Walton et al. 2006, Neural networks,19,1302–1314). Motivation to obtain rewards is an important cognitive component of anhedonia; a symptom that has profound effects on everyday function in a number of disorders, especially depression (Treadway et al. 2009, PLoS One,4,8). A novel physical effort task (Bonnelle et al. 2015, Journal of Physiology,109,16–26) demonstrated how motivation for rewards can be accurately measured and manipulated by varying reward magnitudes across different physical effort demands (measured by a hand dynamometer). While this task can be used to assess physical motivated effort, motivation for rewards may be differentially affected by physical vs. cognitive demands. Hence, we developed a novel effort paradigm similar in design to the physical effort task in which effort was manipulated according to varying demands on attention and working memory. Methods: We tested 103 healthy subjects, who were required to perform categorization of 10 odd/even digits under time pressure in order to win points. Effort requirements were manipulated by adjusting the time constraints on a given trial such that effort levels of 20%, 40%, 60%, and 80% corresponded to 180%, 160%, 140%, and 120% of the subject’s best categorization speed during a practice calibration period. Participants performed 5 blocks of 16 trials, testing all combinations of reward magnitudes (3,6,9 and 12 points) and effort levels. On each trial, subjects could either accept the offer and perform the sequence, or refuse and move to the next trial. The primary outcome measure was the proportion of offers accepted. Results: Similarly to the physical effort task we found a significant interaction between Effort Level and Reward Magnitude for the probability of accepting an offer (F(7.087,637.735)=7.868, p<0.001). The probability of accepting an offer decreased significantly with increasing Effort Level (F(1.584,142.544)=53.45, p<0.001), and increased significantly with increasing Reward Level (F(1.626,146.37)=56.487, p<0.001 ). Effort sensitivity correlated negatively with measures of optimism ABSTRACTS A57

(LOTR – r=-.220, p=0.034), and cognitive apathy (AES – ρ=-.214, p=0.037). Conclusions: This novel cognitive effort task appears to appropriately manipulate motivation for rewards in the cognitive domain, producing results similar to these of an existing physical effort task. This enables a direct comparison between motivation for rewards in the physical and cognitive domains. This project was funded by the Wellcome Trust.

B22 PROMOTING HEALTHIER BEHAVIOUR VIA INFORMATION: THE EFFECT OF UNIT INFORMATION ON ALCOHOL CONSUMPTION IN THE LABORATORY Langfield TIR,Experimental Psychology, Univ of Bristol, The Priory Road Complex, 12a Priory Road, Clifton, Bristol BS8 1TU [email protected] Maynard OM(1), Attwood AS(1), Allen E(1), Munafò MR(1) (1) As presenting author Excessive alcohol consumption is associated with a number of health, social and economic problems. Developing effective interventions to reduce alcohol intake is thus high on the UK government agenda. Given that knowledge of the unit content of alcoholic drinks is generally low, improving unit awareness could be one such intervention. We investigated this possibility by exploring whether unit information influences drinking behaviour in a laboratory setting. We used a 2 × 2 between-subjects design, with unit and calorie information as between-subjects factors. Social alcohol consumers (n = 123, 50% female) were randomised to a unit-condition (no information; unit information) and a calorie-condition (no information; calorie information), and completed a bogus ‘taste-test’ of two alcoholic beers. The primary outcome measure was total percentage of beer consumed during the taste test. Secondary outcome measures included mood, alcohol craving, drink enjoyment, and intentions to consume the beverage in the future. A power calculation indicated that we required a minimum sample of N = 264 to detect an effect of d = 0.35, based on previous studies using the taste test method. ANOVA was used to compare amount of beer consumed between groups (unit vs. no unit information), with and without adjustment for age, gender, TFEQ scores, and hazardous drinking. Amount of beer consumed was higher in the unit-condition (M 55%, SD 28) compared with the no unit information condition (M 48%, SD 28), but there was no clear statistical evidence that this difference was meaningful in either the unadjusted (p = .21) or adjusted (p = .16) analyses. Independent samples t-tests indicated no difference between groups on positive or negative mood, drink enjoyment, or future intentions to consume the beverage (ps > .35). There was weak evidence that alcohol craving was higher in those who received unit information (p = .021). These analyses provide no evidence to support an effect of unit information on drinking behaviour. Critically, the amount of beer consumed in the unit-condition was in fact higher than in the no unit information condition, although there was no statistical evidence that this difference was due to anything other than chance effects. It is possible that compulsory unit labelling, a public-health initiative supported by the alcohol industry, would have no effect on alcohol intake. However, based on our power calculation, a larger sample is required before we can conclude that unit information has no effect on drinking behaviour. This work was supported by the Medical Research Council and the Univ of Bristol (MC_UU-12013/6).

B23 INTERACTIONS BETWEEN SHYNESS LEVEL AND ALCOHOL EXPOSURE - ACUTE ANXIOLYSIS, HANGOVER ANXIETY AND IMPLICATIONS FOR ALCOHOL USE DISORDER RISK Marsh BR, Psychopharmacology Lab, Univ of Exeter, Washington Singer Labs, Perry Road, Prince of Wales Road, Exeter EX4 4QG [email protected] Morgan CJA(1), McAndrew A(1), Stevens T(1), McGahey S(1), Carter E(1) (1) As presenting author Introduction: Social anxiety disorder (SAD) has been related to the incidence of alcohol use disorders (AUD). Shyness can be viewed as a subclinical analogue of SAD but there is little research into the A58 ABSTRACTS effect of alcohol on anxiety levels in highly shy individuals. Therefore, the current study aimed to investigated the acute and sub-acute effects of alcohol in high and low shy social drinkers tested in a naturalistic setting. Methods: 97 individuals were tested in their own homes, who were assigned either consume alcohol to normal levels (n=50) or to remain sober (n=47). Measures were of shyness, state anxiety, alcohol use disorders and hangover symptoms. Results: Our results suggested marginally decreased anxiety resulting from alcohol consumption in high shyness group on the night of alcohol use, with a Time 1 to Time 2 decrease of: t(19) =2.29, p =0.034). Alongside this, there was a significant increase in anxiety the day following drinking: t(19) =0.1.66, p =0.004). There were no differences in low shyness participants in either group or high shy individuals who had not consumed alcohol. There was a significant correlation between anxiety elevation on the second day and AUDIT scores in high shy participants. Conclusions: This study suggests that rebound anxiety the day after drinking is linked to AUD measures in highly shy individuals. This provides a potential marker for increased risk of AUD in a student population which may be helpful for targeting prevention programmes and informing treatments.

B24 DEVELOPING A HUMAN LABORATORY MODEL OF TOBACCO WITHDRAWAL: THE EFFECT OF ACUTE ABSTINENCE ON TASK PERFORMANCE IN SMOKERS Grabski MG, School of Experimental Psychology, Univ of Bristol, 12a Priory Road, Bristol BS8 1TU [email protected] Curran VC(1), Nutt DN(2), Husbands SH(3), Munafo MM(4), Ferguson SF(5) (1) Clinical Psychopharmacology Unit, Research Dept of Clinical, Educational and Health Psychology, UCL, Gower Street, London, WC1E 6BT, UK.; (2) Dept of Medicine, Imperial College, Burlington Danes Building, Du Cane Road, London, W12 0HR, UK.; (3) Dept of Pharmacy and Pharmacology, Univ of Bath, Claverton Down, Bath, BA2 7AY, UK; (4) School of Experimental Psychology, Univ of Bristol, Bristol, UK.; (5) School of Medicine, Univ of Tasmania, Hobart TAS 700, Australia Tobacco withdrawal symptoms are considered to be a main obstacle to successful smoking cessation, and are therefore a target for treatment efforts. Tobacco abstinence has been linked to cognitive performance deficits, as well as to increased attentional bias towards smoking related cues. asksT assessing these domains might serve as a model to indicate the withdrawal state of a smoker, which could be used as an early indicator for the effectiveness of novel smoking cessation treatments. Informed by a systematic review and meta-analysis of the relevant literature, we evaluated a battery of four laboratory tasks measuring response inhibition (go/no-go task), working memory (N-back task), impulsivity (delay- discounting task), and attentional bias (dot probe task with eye tracking) for their sensitivity to acute abstinence. After two morning laboratory sessions (overnight abstinent, smoking satiated), ecological momentary assessment (EMA) of craving levels via a smartphone app was conducted over the course of the afternoon. Participants (N=70, 41% female) were on average aged 22 years (SD 5), smoked 11 cigarettes per day (SD 4) and had an FTCD score of 4.4 (SD 1.6). Abstinence resulted in more commission errors on the go/no-go task (t[69]=-3.07, p=0.003). However, no clear difference was found in omission errors on the N-back task (t[67]=1.49, p=0.14) or the discounting parameter k on the delay discounting task (t[69]=0.18, p=0.86). A repeated measures ANOVA of the eye-tracking data for the dot probe task indicated a main effect of dwell time on picture type (F[1, 57]= 42.53; p<0.0001), but no main effect of abstinence (F[1, 57]=0.60; p=0.81), and no picture type × abstinence interaction (F[1,57]=0.22; p=0.64). Abstinence strongly predicted EMA of craving (ß=.56, p<.001). Go/no-go task performance, augmented by EMA of craving, might serve as a model for tobacco abstinence. Ongoing work is assessing whether the go/no-go-task is also sensitive to anxiety, a risk factor for relapse, by inducing anxiety experimentally via the inhalation of 7.5% CO2 enriched air. This research was supported by an Economics and Social Research Council PhD scholarship and Rusan Pharma Ltd. ABSTRACTS A59

B25 A REDUCTION IN VOLUME OF THE PUTAMEN IN ALCOHOLISM Sadler C, Imperial College London, Centre for Neuropsychopharmacology, Division of Brain Sciences, London W12 0HS [email protected] Myers J(1), Ribeiro A(1), Mick I(1), Nutt DJ(1), Turton S(1), Lingford-Hughes AR(1), ICCAM Consortium (1) (1) As presenting author Introduction: Structural changes in the brain have been reported in substance dependence, particularly in alcoholism with reductions in grey matter volume (Oscar-Berman et al,2003,Alcohol Research and Health,27:125–133). The toxicity of alcohol consumption and withdrawal are major contributions, with volumetric changes improving with abstinence. We compared the volumes of three regions of interest (ROIs) involved in addiction: the caudate, putamen and hypothalamus, in Alcohol Dependent Patients (ADPs) and Healthy Controls (HCs), to test the hypothesis that ROI volume would be smaller in ADPs and related to the severity of dependence and length of abstinence. Methods: Twenty four abstinent ADPs and 22 HCs were recruited for a PET imaging study and the ICCAM fMRI study (Paterson et al,2015,Journal of Psychopharmacology,29:943–60). Participants underwent a structural MRPRAGE on a 3T Siemans Trio MRI scanner. Subjects with additional substance dependence (not including nicotine) were excluded. The caudate, putamen and hypothalamus were manually delineated to ascertain their volume (Tziortzi et al,2011,NeuroImage,54:264–277). Severity of Alcohol Dependence Questionnaire (SADQ) was completed for 11 ADPs and length of abstinence was recorded for all ADPs. We performed one tailed t-tests and Pearson’s correlation for normally distributed data and Mann Whitney U tests and Spearman’s correlation for non-normally distributed. Results: ADPs had smaller caudate, putamen and hypothalamus volume (mean±SD=8.17±1.53, 9.14±1.19, 1.08±0.19) compared with HCs (8.74±1.43, 10.37±1.19, 1.12±0.21). This was statistically significant for the putamen (p=0.0005). There was a significant correlation (r=0.511,p=0.011) between the volume of the hypothalamus and duration of alcohol abstinence, however no correlation for other ROIs. Abstinence ranged from 14 to 2920 days. In an ADP subsample (n=11), SADQ score positively correlated with caudate (r=0.46,p=0.154), putamen (r=0.162,p=0.634) and hypothalamus (r=0.638,p=0.035) volume, however these were not significant after correcting for multiple comparisons. There were no significant correlations between age and any ROIs in HCs or ADPs. Discussion: We found a smaller putamen in ADPs but not caudate or hypothalamus. ROI volumes were not related to severity of dependence or age; however a longer duration of abstinence was associated with larger hypothalamic volume. These results could reflect a pre-existing vulnerability to alcohol dependence, a particular vulnerability of the putamen to the toxicity of alcohol and/or less recovery of the caudate and putamen compared with the hypothalamus. In this study, ADPs were stable in recovery with substantial range of abstinence length, raising questions about the potential and differing reversibility of structural changes with recovery in alcoholism. The study was funded by MRC G100226 & G1000018.

B26 ANALYSIS OF THE REWARD SYSTEM IN DISORDERED GAMBLING WITH MR AND 11C-CARFENTANIL PET Sohail A, Neuropsychopharmacology, Imperial College London, Burlington Danes, Hammersmith Hospital, 160 Du Cane Road, Hammersmith, London W12 ONN [email protected] Myers JFM(1), Mick I(1), Nutt DJ(1), Lingford-Hughes AR(1), Turton SP(1), Clark L(2), Bowden-Jones HM(3) (1) As presenting author; (2) Centre of Gambling Research, Dept of Psychology, Univ of British Columbia, Vancouver, Canada, V6T 1Z4; (3) National Problem Gambling Clinic, CNWL NHS Foundation Trust, London, UK, SW6 1NH Introduction: Evidence implicates the opioid system and cortico-striatal dysregulation in the pathophysiology of disordered gambling (DG; Grant et al., 2006, Am J Psychiatry. 11(12)924–930). Recently we reported blunted opioid release with no change in mu-opioid receptor availability in DG (Mick et al., 2015, Neuropsychopharmacology doi: 10.1038/npp.2015.340). This study focuses on three regions of interest (ROIs) important in mu-opioid signalling: caudate, putamen and hypothalamus. We hypothesised a decrease in volume of the caudate-putamen. We further explored the relationship between the opioid A60 ABSTRACTS system with 11C-carfentanil PET and hypothalamic volume; the hypothalamus being the location of opioid cell-bodies. Methods: Magnetic resonance imaging (MRI) scans were acquired in two studies, both using a 3-Tesla Siemens-Trio MRI scanner (Mick et al., 2015, Neuropsychopharmacology doi: 10.1038 / npp.2015.340) (Colasanti et al., 2012, Biol Psychiatry 72:371–7). Subjects included healthy volunteers (HV; n=22) and DG (n=15). ROIs were hand-drawn following existing guidelines (Tziortzi et al., 2011, Neuroimage 54:264–77). T-tests were used to assess statistical significance of any ROI volume difference. 11C-carfentanil PET-data were available for some participants (HV:15; DG:14; Mick et al. 2015 Neuropsychopharm) with 11C-carfentanil PET scans before (baseline) and 3 hours after a single oral dose of 0.5mg/kg d-amphetamine, which stimulates the release of endorphins. Baseline 11C-carfentanil (BPND) (reflecting mu opioid receptor availability) and reduction in 11C-carfentanil binding from baseline to post-amphetamine scans (delta- BPND) (reflecting level of endorphin release) was assessed in the hypothalamus. delta-BPND was calculated as: delta-BPND = (BPNDpre – BPNDpost)/BPNDpre. Pearson rank correlation-coeffiecients were calculated between both 11C-carfentanil BPND and delta-BPND values and hypothalamus volume in DGs. Results: In DG compared with HV, the putamen was smaller (mean+ S.D.: HV 8.72+1.22 vs DG: 8.07+ 0.56; p=0.04), whilst the hypothalamus was larger (mean+ S.D. HV: 1.07+0.13 vs DG: 1.18+0.12; p=0.01) compared with DGs. A smaller caudate was also found in DG but was not significant (mean+ S.D. HV: 6.1+0.98 vs DG: 5.53+0.85; p=0.07). A significant positive correlation was found between hypothalamus volume with baseline 11C-carfentanil BPND and (r = +0.55, p=0.04), but not with delta-BPND (r = -0.11, p=0.71). Conclusion: We found that DG was associated with smaller putamen and larger hypothalamus volume and the latter was positively associated with mu opioid receptor availability but not endorphin release. Smaller putamen volume may contribute to and reflect poor-decision making in DG; facilitating the cycle of addiction. The role of greater hypothalamic volume is not clear in DG and deserves further attention given its role in addictive behaviours. Financial Sponsorship: The study was financially supported by the UK Medical Research Council (MRC G100226).

B27 DECISION-MAKING IMPAIRMENTS IN CURRENT AND EX USERS OF OPIATES: A META-ANALYTIC REVIEW Biernacki KC, Cognition and Emotion Research Centre, Dept of Psychology, Australian Catholic Univ, 115 Victoria Pde, Fitzroy, Victoria Australia 3065 [email protected] McLennan SN(1), Terrett G(1), Labuschagne I(1) (1) As presenting author Introduction: Opiate use is associated with deficits in decision-making, with literature suggesting a consistent deficit across different samples and measures. However, many studies have included participants who have co-morbidities such as brain injuries or poly-substance addictions. The degree to which these comorbidities may account for the decision-making deficit is currently unclear. It has also been suggested that when people stop using opiates, they may experience improvements in decision- making. Only a few studies have examined this question, and findings have been mixed. Aim: The first aim of this meta-analysis was to assess the magnitude of the decision-making deficit in current and ex users. A second aim was to evaluate whether the presence of polydrug use and/or head injury moderates this relationship. A final aim was to determine whether ex opiate users demonstrate smaller decision-making deficits (relative to non-drug-using controls) than current opiate users do. Method: We analysed 21 studies which compared the performance of current (512 subjects) or ex (391 subjects) opiate users to healthy controls (833 subjects) on decision-making measures such as the Iowa Gambling Task and Delayed Discounting Task. Effect sizes were pooled to calculate the overall magnitude of the decision-making deficit, and subgroup analyses were conducted to assess the impact of moderator variables such as head injury status, as well as the effect of abstinence. Results: When ex and current users were considered together, they demonstrated a moderately strong impairment in decision- making relative to non-using controls (d = 0.639, 95%CI =, 0.46–0.82, p< .001). Moderating variables had no significant impact on the magnitude of the decision-making deficit. Contrary to expectation, the magnitude of the difference between ex-users’ performance in relation to controls (d = 0.50, 95%CI = 0.12–0.87) did not significantly differ from the magnitude of the difference between current users’ performance in relation to controls (d = 0.70, 95%CI = 0.51–0.89; Qbetween (1) = 0.88, p = 0.35). ABSTRACTS A61

Conclusion: Decision-making deficits in opiate users are relatively severe, and cannot be explained by co-morbid conditions. Furthermore, decision-making deficits in people who have stopped using opiates are just as severe as those seen in current users. This suggests ex-users may require additional support with decision-making even after abstinence has been achieved. However, further research that tracks decision-making ability from a period of active use through to a period of abstinence is needed to confirm this effect. Funding: This project has been funded by the Faculty Research Students Support Scheme (Australian Catholic Univ).

B28 SALIVA CORTISOL LEVELS IN ALCOHOLICS PRIOR TO RELAPSE DRINKING Little HJ, Addictions Dept, Inst of Psychiatry, Psychology and Neuroscience, King’s College London, Addiction Sciences Building, 4 Windsor Walk, Denmark Hill, London SE5 8BB [email protected] Brewer A-J(1), Drummond C(1) (1) As presenting author Introduction: This study was designed to examine protracted alcohol withdrawal symptoms and hypothalamopituitary adrenal (HPA) function in alcoholics during the abstinence phase. Participants with a diagnosis of alcohol dependence were recruited from two London non-statutory drug and alcohol detoxification units. Ethics approval was given by the East Dulwich Research Ethics Committee and all participants gave informed consent prior to the study. Method: Following screening interviews and application of exclusion criteria, participants attended meetings with the researcher at 31 (±4) and 59 (±7) days from cessation of drinking. All meetings were between 11 am and 2 pm and cortisol levels were measured in participants who had refrained from eating, drinking or smoking for one hour before the meeting and showed breath alcohol levels below detectable limits. Saliva samples were taken, using salivettes, immediately after the researcher met each participant (0 minutes) and at 30, 60 and 90 minutes later. Between these time intervals, participants were completing psychological questionnaires. Salivary cortisol levels were measured by radioimmunoassay. Relapse drinking was determined at the 59 day meeting, using the Time Line Followback questionnaire, and was defined as consumption of more than 8 units of alcohol for a man or 6 units for a woman on at least one day between the 31 day and 59 day follow up points. Results: Those participants who were abstinent from drinking at the 31 day follow up were subsequently divided into those who by the 59 day point had remained abstinent (n=22) or had relapsed back into drinking alcohol (n=18). Cortisol concentrations, nmol/L (interval minutes) at the 31 day point for those who remained abstinent were 4.99 (0), 5.18 (30), 5.39 (60) and 4.31 (90) and for those who relapsed were 5.75 (0), 7.69 (30), 6.91 (60) and 5.3 (90). Two way ANOVA showed salivary cortisol levels in those participants who relapsed between the 31 and 59 day follow up points were significantly higher than the levels in those who did not relapse during this time (P<0.05, F[1,152] = 3.96). Time comparison and interactions were not significant (P>0.05). Conclusions: The data suggests a difference in HPA function in people with alcohol dependence during the protracted withdrawal period and before the actual occurrence of relapse drinking, that could be a predictive factor for relapse. This work was supported by an MRC Studentship to A-J Brewer.

B29 EFFECTS OF CANNABIDIOL (CBD) ON BRAIN FUNCTION, STRUCTURE, NEUROCHEMISTRY, COGNITION AND SYMPTOMS: PRELIMINARY FINDINGS FROM TWO TRIALS OF ACUTE AND PROLONGED CBD ADMINISTRATION TO CANNABIS USERS Solowij N, School of Psychology and Illawarra Health & Medical Research Inst, Univ of Wollongong, Northfields Ave, Wollongong, NSW Australia 2522 [email protected] Broyd S(1), vanHell H(1), Greenwood L-M(1), Beale C(1), Suo C(2), Yucel M(2) (1) As presenting author; (2) Monash Inst of Cognitive and Clinical Neurosciences and School of Psychological Sciences, Monash Univ, Melbourne, Victoria, Australia Introduction: Cannabis use is associated with altered brain structure and function, elevated psychological symptoms and impaired cognition. These adverse effects are attributed to delta-9-tetrahydrocannabinol A62 ABSTRACTS

(THC). Cannabidiol (CBD) has been shown to ameliorate some of the adverse effects of THC and to possess antipsychotic and neuroprotective properties. We reported that exposure to CBD may protect against hippocampal volume loss in chronic cannabis users (Yücel et al, 2016, Transl Psychiatry, 6, e710). We subsequently conducted two randomised controlled trials (RCTs) of acute and prolonged CBD administration to frequent and infrequent cannabis users to investigate potential mechanisms of its action. Methods: In RCT-1, the acute effects of vaporised CBD alone and in combination with 8mg THC at high (400mg) and low (4mg) CBD doses were examined on brain function (electrophysiology, neurochemistry (spectroscopy), resting state functional magnetic resonance imaging), cognition and symptoms, in subsamples of 65 volunteers with wide ranging experience with cannabis. RCT-2 is still in progress, investigating the effects of prolonged oral administration of CBD (200mg daily for 10 weeks) to 22 cannabis users with similar outcome measures as RCT-1, as well as hippocampal volumes. Results from 10 cannabis users who have completed the 10 week trial will be presented. Results: RCT-1 found that vaporised high dose CBD did not attenuate the adverse functional, psychological and cognitive effects of THC, while low dose CBD sometimes enhanced these, particularly in infrequent cannabis users. High dose CBD alone showed psychoactivity relative to placebo and significantly elevated hippocampal glutamate levels (p=.006). Prolonged administration of CBD for 10 weeks (RCT-2) also resulted in significantly elevated hippocampal glutamate levels (by 16%; p<.005) which were associated with better memory performance (p<.05). Less change in glutamate was observed with an earlier age of onset of cannabis use (p<.005). Other outcome measures are still undergoing analysis. Conclusions: The results of these studies suggest complex interactions between THC and CBD on psychological and brain function that vary with proportional dose, degree of prior exposure to and age of onset of regular cannabis use. There is significant interest in developing CBD as a medication for multiple indications but little is known regarding its mechanism of action. Our data suggest that elevation of glutamate levels in the hippocampus is one mechanism that may underlie some of its actions. Our findings have implications not only for understanding the therapeutic potential of cannabinoids, but also for recreational use in the context of increasing trends toward legalisation. Funded by the National Health and Medical Research Council of Australia (Project Grant 1007593) and the Australian Research Council (ARC Future Fellowship FT110100752).

B30 VALIDITY OF SELF-REPORTED CANNABIS USE HISTORY: THE ARGUMENT FOR HAIR ANALYSIS Mokrysz C, Clinical Psychopharmacology Unit, UCL, 1–19 Torrington Place, London WC1E 7HB [email protected] Freeman TF(1), Korkki SM(1), Griffiths K(1), Curran HV(1) (1) As presenting author Retrospective drug histories can be unreliable, often depending on memory of drug use decades earlier e.g. age first used, total lifetime uses. Cannabis use history is particularly complex since cannabis spliffs are often shared with friends (and the drug impairs memory) so users may be unaware how much drug they personally consumed. Further complication comes from the wide variation in available cannabis strains and potency, poor estimation of cannabis weight, and changing meaning of common terminology (e.g. an “eighth”). Accurate quantification of actual dose consumed is one of the major challenges for cannabis research, and current limitations must be addressed. As part of a wider study we administered detailed cannabis use history to non-dependent weekly cannabis users, forming two groups: adolescents (aged 16–17; n=17) and adults (aged 24–28; n=18). Participants completed standard questions (e.g. years used; days per month; time to individually smoke an “eighth”) regarding three different cannabis types (“skunk”/“hash”/“Thai”). Next participants provided detailed timeline follow-backs, starting from age of first use, with physical cannabis samples of 0.05g/0.1g/0.2g/0.3g provided to aid weight estimation. For each timeline, participants recorded type of cannabis, frequency of use, grams per session, and tobacco:cannabis ratios. New timelines were started to reflect a change in usage pattern. Hair and urine samples were collected for cannabinoid analyses. Timeline-derived variables and standard question responses were then compared to cannabinoid analyses, to assess measure validity. According to standard questions adolescents reported currently using “skunk” 1.9 days (SD=0.8) per week, and adults 0.8 days ABSTRACTS A63

(SD=0.9) per week. However timeline-derived variables suggested adolescents currently used “skunk” 2.9 days (SD=1.0) and adults 1.1 days (SD=1.1) per week. Additionally there were discrepancies between the single-item “time to smoke an eighth” (adolescents= 2.5 days (SD=4.5), adults= 26.5 days (SD=25.7); p=.003) and timeline-derived variables (adolescents= 6.9 days (SD=6.0), adults= 48.5 days (SD=40.2); p=.004). We also found large variation in response to “How much does an eighth weigh?” (adolescents= 1.8g (SD=0.3); adults= 4.7g (SD=4.8); p=.045), questioning the validity of this commonly used variable. Estimates were then compared to urine and hair analysis results. There is substantial variation in reporting of frequency and quantity of cannabis use, according to how questions are asked and interpretation of terms used. As such there is a pressing need to improve measures of cannabis use history, and our findings suggest biological quantification may be necessary to improve reliability and consistency across studies. This study was funded by an MRC Studentship to CM.

B31 NO SMOKE WITHOUT TOBACCO - A GLOBAL OVERVIEW OF CANNABIS AND TOBACCO ROUTES OF ADMINISTRATION AND THEIR ASSOCIATION WITH INTENTION TO QUIT Hindocha C, CPU, UCL, 1–19 Torrington Place, London WC1E 7HB [email protected] Winstock AR(1), Freeman TP(2), Lynskey MT(3), Ferris JA(4) (1) Dept of Addiction, Inst of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK; (2) As presenting author; (3) Dept of Addiction, Inst of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK; (4) Inst for Social Science Research, Univ of Queensland, Australia Introduction: Cannabis and tobacco are common drugs of abuse worldwide and are often used in combination through a variety of routes of administration (ROAs). Recent prevalence statistics suggesting one billion people worldwide smoke tobacco (22.6% of adults) and 174 million using cannabis (3.5% of adults) with highest rates of use in Oceania (10.3%) (Gowing et al., 2015, Addiction, 110, 904– 919). ROAs are important to investigate as they may confer to higher rates of DSM-IV cannabis abuse, produce aero-respiratory changes to favor cannabis inhalation, and alter the subjective experience of the drug. Here we aimed to provide an overview of how cannabis and tobacco ROAs varied across countries and assess the impact of tobacco ROAs on motivation to use less cannabis and tobacco. Method: A cross-sectional online survey (Global Drugs Survey 2014) was completed by 33687 respondents (mean age = 27.9; %female = 24.8) who smoked cannabis at least once in the last 12 months. Most common ROA, frequency of cannabis/tobacco use, and questions about motivation to use less cannabis/tobacco were recorded. Results: Tobacco-based ROAs were used by 65.6% of respondents. These were most common in Europe (77.2–90.9%) and Australasia (20.7–51.6%) and uncommon in the Americas (4.4–16.0%). Vaporizer use was most common in Canada (13.2%) and the United States (11.2%). Non-tobacco based routes of administration were associated with a 10.7% increase in odds for ‘desire to use less’ tobacco (OR: 1.107, 95% CI: 1.003, 1.221), 80.6% increase in odds for ‘like help to use less tobacco’ (OR: 1.806, 95% CI: 1.556, 2.095) and a 103.9% increase in the odds for ‘planning to seek help to use less tobacco’ (OR: 2.039, 95% CI: 1.638, 2.539) in comparison to tobacco based routes of administration. Associations between ROA and intentions to use less cannabis were inconsistent. More males (63.8%) than females (36.2%) used tobacco ROAs (χ2(1)=48.51, p<0.001). These users were younger (M=26.23, SD=8.48) than those using non-tobacco ROA users (M=30.79, SD=12.76) (F(1,14622)=1058.94, p<0.001). Conclusions: Results support considerable global variation in cannabis and tobacco ROAs. Tobacco ROAs are common, especially ‘joints with tobacco’, especially in Europe, but not in the Americas. Tobacco routes are associated with lower motivation to change tobacco use. Given the scale of cannabis and tobacco use, interventions addressing the comorbidity need to accommodate this finding and encourage non-tobacco routes. Alternative ROAs such as vaporizers should be investigated for their harm reduction properties. This work was supported by the Global Drug Survey (GDS) and Medical Research Council (PhD to CH). AW is the founder of GDS. A64 ABSTRACTS

C01 AN INVESTIGATION OF HUMAN PLASMA PROTEIN BIOMARKERS IN BURNING MOUTH SYNDROME O’Driscoll D, Transpharmation Ireland Ltd, 3.47 Lloyd Inst, Trinity College , Dublin, 2 [email protected] Prenderville JA(1), Bianchi M.(1), McCreary C(2), McKenna JP(2), Downer EJ(3), Barry A(4), O’Halloran KD(4) (1) As presenting author; (2) Cork Univ Dental School and Hospital,Univ College Cork, Cork, Ireland; (3) Dept of Physiology, School of Medicine, Trinity Biomedical Sciences Inst, Trinity College Dublin, Univ of Dublin, Dublin 2, Ireland; (4) Dept of Physiology, School of Medicine, Western Gateway Building, Univ College Cork, Cork, Ireland Introduction: Burning mouth syndrome (BMS) is a heterogeneous neuropathic pain condition of unknown aetiology, however inflammatory and synaptic structural processes may underlie the pathology. BMS patients exhibit high psychiatric co-morbidity. Neuronal microtubules are fundamental to structural neuronal plasticity. Post-translational modifications (PTMs) of α-tubulin can be measured to assess microtubule dynamics. Alterations in brain α-tubulin PTMs, including acetylated tubulin (Acet-Tub), tyrosinated tubulin (Tyr-Tub), detyrosinated tubulin (Glu-Tub) and Δ2-tubulin (Δ2-Tub), have been implicated in depressive disorders in animal models. The aim of this study was to assess microtubule profile in plasma isolated from BMS patients. This study also assessed the plasma cytokine signature in BMS plasma, and correlated this signature with plasma microtubular proteins. Methods: Ethical approval was obtained from the Clinical Research of Cork Teaching Hospitals, the School of Medicine Research at Trinity College Dublin, and the Tallaght Hospital/St. James’s Hospital. Healthy controls (n=10) and BMS patients (n=10) completed the self-rated 16-item Quick Inventory of Depressive Symptomatology (QIDS- SR16) questionnaire and a Visual Analogue Scale (VAS) to quantify oral pain. Peripheral venous blood samples were obtained by venipuncture, plasma aspirated and stored at -80°C until analysis. Plasma from all subjects were analysed by infrared Western Blot to quantify plasma α-tubulin PTMs. Plasma cytokines were also measured using V-PLEX human pro-inflammatory panel assays. Data were analysed using Student’s t-test, one-way ANOVA and linear regression. Results: BMS was associated with enhanced oral cavity pain (51.1±4.9mm BMS vs. 0.5±0.5mm control; p<0.001) and increased depressive symptomatology (7.9±1.3 BMS vs. 2.5±0.8 control; p<0.01). Plasma Acet-Tub expression demonstrated a non-significant increase in BMS patients (111.2±12.87%) that were not on antidepressant medication (n=6), compared to control subjects (100.4±9.9%). Post-hoc analysis of clinical subgroups showed a non-significant reduction in plasma Acet-Tub in BMS patients treated with antidepressants (83.8±12.0%; n=4), compared to BMS patients not treated with antidepressants (129.4±16.5%; n=6) (p=0.07). Increased depression scores correlated with increased plasma Glu-Tub, Acet-Tub, and Δ2-Tub in control individuals (p<0.05), while decreased Tyr-Tub correlated to depression severity in BMS patients (p<0.05), suggestive of reduced microtubule dynamics. Increased plasma IL-8 was also observed in BMS patients compared to controls (8.89±1.13pg/ml vs. 5.59±0.78pg/ml, respectively; p<0.01). Analysis of clinical subgroups revealed that increased plasma IL-8 was associated (P<0.01) with antidepressant treatment in BMS (11.70±1.19pg/ml; n=4) compared to controls (5.59±0.78pg/ml; n=10). Conclusion: These findings support the translational validity of plasma α-tubulin PTMs and IL-8 as indicators of treatment intervention or pharmacological efficacy in BMS patients. Funding: Transpharmation Ireland Ltd.

C02 INHIBITION OF RACEMASE AS A NOVEL APPROACH TO MODULATING NMDA RECEPTOR FUNCTION Koulouris CR, Sussex Drug Discovery Centre, Uni. Sussex, Chichester II, North-South Road, Falmer Brighton BN1 9QJ [email protected] Walker S(1), Atack J(1), Roe M(1) (1) As presenting author The N-methyl-D-aspartate (NMDA) receptors (NMDAR) are a subtype of ionotropic glutamate receptors that are highly expressed in the central nervous system (CNS) and are involved in the excitatory ABSTRACTS A65 synaptic transmission and synaptic plasticity that form the basis of many critical CNS functions. NMDAR dysfunction has been implicated in Alzheimer’s disease, amyotrophic lateral sclerosis (ALS), neuropathic pain, schizophrenia, and depression, among others. Most non-selective NMDAR antagonists (e.g. ketamine) have undesirable side-effects that restrict their clinical utility for relieving symptoms of neuropathic pain and treatment-resistant depression. Modulators of NMDAR function that either act indirectly or target particular NMDAR subtypes offer the potential to have reduced side-effects relative to the non-selective antagonists. One approach is to inhibit the production of the NMDAR co-agonist, D-serine, which is produced endogenously by conversion of L-serine to D-serine by the enzyme serine racemase (SR). Inhibitors of SR that reduce the production of D-serine may therefore have therapeutic benefits in disorders associated with NMDAR hyperfunction. Wild-type hSR was expressed with a polyhistidine-tag in BL21 CodonPlus (DE3)-RIL cells and after a two-step purification (immobilized metal affinity chromatography followed by size-exclusion chromatography) was isolated to a high degree of purity (>90%) and yield of ~3mg/L. The presence and purity of SR were verified yb western blot, mass spectrometry and crystallography, with the crystal structure of SR in complex with malonate being solved in-house to a resolution of 2.2Å. The conversion of L-serine to D-serine by SR was quantified by a coupled biochemical assay in which D-serine is broken down by D-amino acid oxidase (DAO) and the resulting H2O2 takes part in a chemiluminescence reaction with HRP and luminol to produce light. The Km of L-serine was determined to be 6.4 ± 0.5mM and at a substrate concentration of 2mM, the IC50 for the competitive inhibitor malonate was 82μM, in agreement with literature values. A single-point screen of 3000 fragments (molecular weight <300) has been performed (n=2). Hits have been reordered and IC50s will be determined in the SR assay and counter-screened in the DAO assay (which comprises the second part of the coupled assay). Specific hits will then be confirmed using biophysical methods such as thermal shift, MicroScale Thermophoresis with the ultimate mode of action being established using X-ray crystallography. These structural data will then be used to direct medicinal chemistry efforts aimed at increasing the potency and physicochemical properties of the hits. Fully funded by the Univ of Sussex.

C03 L-αAMINOADIPIC ACID INDUCED GLIAL ABLATION IN THE PRE-LIMBIC CORTEX INDUCES DEPRESSIVE AND ANXIETY -LIKE BEHAVIOURS IN MICE David J, Neuroscience, Trinity College, College Green, Dublin 2, Ireland [email protected] Mcintosh AL(1), Harkin A(1) (1) As presenting author Astrocytic glial cells play important roles in many CNS functions. Developing an animal model of astroglial dysfunction is necessary to further understand the role of glia in the pathophysiology of depression (Rajkowska & Stockmeir, 2013, Curr. Drug Targets 14, 1225–1236) Male C57BL6/J mice (12 weeks) received a stereotactically bilateral injection of the astrocytic glia toxin L-alpha-aminoadipic acid (L-AAA) [50μg/μl; 1μl] or saline into the prelimbic Cortex (PLC) [1.7mm anterior and ± 0.4mm dorsolateral to bregma and depth -2.5mm] adapted from (Lee, Y. et al., 2013, J. Psychiatry Neurosci. JPN 38, 183.). A range of behavioural tests were carried out at different time points adapted from (Banasr, M. and Duman, R.S., 2008, Neurodegener. Dement. Depress. 64, 863–870) including the saccharin preference (hedonic responsivity) during the all experiment, open field test (locomotor activity and anxiety and tail suspension test (despair). In a first experiment after a single injection, three groups of animals performed these tests at different time points (Open-field at 24h, TST at 48h), (Open-field at 48h, TST at 72h), (Open-field at 6 days, TST at 7 days) following by brain collection. A preliminary study evaluates the effect of a two daily injection on open-field at 24h and TST at 48h. Immunohistochemistry for the astrocyte marker Glial Fibrillary Acidic Protein (GFAP) was performed to confirm changes in glial cell number in the Pre-Limbic-Cortex (PLC). After a single injection of L-AAA into the PLC, 2 Way ANOVA on locomotor activity and open field behaviour didn’t reveal any effect at all the time-point investigated (24h, CTL: 98.34 ± 0.2865 N=10, L-AAA: 98.34 ± 0.3554 N=10, 48h, CTL: 98.09 ± 0.6471 N=10, L-AAA: 96.41 ± 1.338 N=9 or at 6 days CTL: 97.26 ± 0.8054 N=8, L-AAA: 97.97 ± 0.2744 N=10), 2 Way ANOVA on saccharin preference revealed an effect treatment (decrease of saccharin preference) over the 24h period A66 ABSTRACTS following L-AAA administration (***, P<0.001) with no effect at 48h or 72h. (baseline before surgery (CTL: 85.25 ± 1.337 N=27, L-AAA: 84.42 ± 1.482 N=28), 24h (CTL: 78.77 ± 3.080 N=23, L-AAA: 64.97 ± 2.916 N=26), 48h (CTL: 77.28 ± 3.631 N=21, L-AAA, 81.05 ± 2.480 N=22) and 72h (CTL: 88.42 ± 2.287 N=13, L-AAA: 84.99 ± 2.581 N=14). Student t-test on the TST reveals a trend towards increase in immobility at 48h (CTL: 91.99 ± 14.09 N=10, L-AAA: 122.3 ± 8.949 N=10), a significant increase in immobility at 72h (*, p<0.05) (CTL, 102.1 ± 9.291 N=10, L-AAA: 132.1 ± 8.380 N=9) and no difference after 7 days (CTL: 91.16 ± 16.73 N=7, L-AAA: 109.1 ± 8.808 N=10) when compared to saline treated controls. Students t-test on GFAP positive astrocytes reveal that L-AAA significantly reduced the number of GFAP-positive astrocytes 48h post- injection (*,p<0.01) (CTL: 297.7 ± 23.91 N=8; L-AAA: 166.9 ± 26.14 N=10) with a gradual recovery at 72h (p<0.05) (CTL: 308.8 ± 26.84 N=10; L-AAA: 204.3 ± 30.33 N=9) and no difference at 7 days when compared to saline controls (CTL: 326.7 ± 32.39 N=8; L-AAA: 334.9 ± 36.15 N=10). By injecting L-AAA over two consecutive days anxiety-like behaviours were revealed in the open field (latency to enter the centre) by a student t-test 24h following the second injection of L-AAA (*,P< 0.05) (CTL: 24.06 ± 5.500 N=10, L-AAA: 77.14 ± 21.19 N=10). Moreover, student t-test reveals that L-AAA increased the immobility time in the TST 48h following the second L-AAA injection (*,P<0.05) (CTL: 52.19 ± 13.90 N=10, L-AAA: 96.48 ± 13.53 N=9) and repeated measure 2 way ANOVA reveals a reduction in saccharin preference (main effect of treatment *,p<0.05 over 3 days) (baseline, CTL: 84.81 ± 1.826 N=9, L-AAA: 83.49 ± 1.361 N=10, 24h, CTL: 78.31 ± 3.060 N=9, L-AAA: 69.65 ± 3.473 N=10, 48h, CTL: 77.04 ± 2.572 N=9, L-AAA: 72.25 ± 2.680 N=10, 72h, CTL: 86.10 ± 2.751 N=9, L-AAA: 81.72 ± 3.131 N=10). Student t-test on GFAP positive cell number reveals a decrease (*, P<0.05) 48h following the second L-AAA injection (CTL: 269.6 ± 30.55 N=9, L-AAA: 186.4 ± 22.43 N=9). In conclusion, L-AAA administration in the pre-limbic cortex induced a depressive- like phenotype with anhedonia related behaviour and increased immobility related to behavioural despair (1 or 2 doses) and L-AAA anxiety related behaviour (2nd dose). These deficits in behaviour were associated with a transient decrease in GFAP-positive cell number that recovers 7 days following L-AAA administration. These results indicate that L-AAA administration into the PLC may be valuable model to investigate the role of astrocytes in depression or anxiety related behaviours. This model could be used to investigate novel glutamatergic compounds for depressive or anxiety disorders. This work was funded by the Marie Curie Brain Imaging Return to Health (r’Birth) consortium.

C04 MODELLING DECISION-MAKING IN A RODENT JUDGEMENT BIAS TASK USING A BIOPHYSICALLY REALISTIC NEURAL NETWORK MODEL Hales CA, Dept of Physiology, Pharmacology & Neuroscience, Univ of Bristol, Biomedical Sciences Building, Univ Walk Bristol BS8 1TD [email protected] Robinson ESJ(1), Houghton CJ(2) (1) As presenting author; (2) Merchant Venturers Building, Dept of Computer Science, Woodland Road, Bristol, BS8 1UB Biophysically realistic neural network models have been used to accurately simulate decision-making performance in simple, two-choice decision-making tasks in humans and non-human primates. These networks comprise competing populations of neurons, each selective for a single choice, and exhibit attractor dynamics which signify evidence accumulation for one or other alternative. A winner-takes-all cortical network model [Wang, 2002, Neuron, 36:955–968] was modified and used to simulate baseline decision-making performance in a rodent judgement bias task (JBT). The effect of altering network activity through simulated neuromodulatory changes was then investigated. A cortical neural network model (based on Wang [2002]) was implemented in the Brain simulator. The parameters and network architecture were identical with the following exceptions: i) increased external conductance to inhibitory neurons was used to increase decision latency without destabilising attractor and decision-making dynamics; ii) the decision threshold was increased to 40Hz. Model simulations were compared to baseline decision-making performance from the reward rodent JBT [Hales et al., 2016, Plos One]. The effect of neuromodulation on the network was simulated by altering cellular or postsynaptic conductances in the model (as in Eckhoff et al. [2009, J Neurosci, 29:4301–4311]). Model simulations produced output comparable to behavioural performance on the JBT. For each of the three cues in the task (high reward, ABSTRACTS A67 midpoint and low reward) behavioural response latencies were 1.99±0.13, 3.20 ±0.24 and 3.76±0.29 seconds respectively, whilst model decision times were 1.86±0.03, 3.18±0.02 and 3.49±0.15 seconds. In the JBT, percent of high reward responses for the three cues were 95.67±0.77%, 53.92±4.75% and 33.27±2.80%, whilst equivalent measures from model simulations were 96.17±0.60%, 52.65±0.68% and 30.73± 5.04%. Simulation of neuromodulatory effects by altering conductances of either excitatory, or inhibitory populations only tended to rapidly destabilise the network, however altering AMPA, NMDA or GABA conductances to both populations caused directional modulation-dependent changes in percentages of model decisions reaching the high reward threshold. A biophysically realistic neural network model can be used to simulate decision-making performance on a rodent JBT. Although this single-layer model is a simplification of the decision-making processes that may be occurring, and must be embedded within much larger networks in the brain, initial results from simulation of the effect of neuromodulation on decision-making have provided testable predictions. These predictions could be investigated using the rodent JBT by carrying out pharmacological manipulations that replicate the neuromodulatory simulations. This study was funded by Wellcome Trust Doctoral Training Programme in Neural Dynamics.

C05 EFFECTS OF ACUTE HORMONE TREATMENTS ON AFFECTIVE BIASES IN SPRAGUE DAWLEY RATS Papciak JK, School of Physiology, Pharmacology and Neuroscience, Univ of Bristol, Biomedical Sciences Bldg, Univ Walk, Bristol BS8 1TD [email protected] Robinson ESJ(1), Mendl M(2) (1) As presenting author; (2) School of Veterinary Sciences, Univ of Bristol, Langford House, Langford, Bristol BS40 5DU Studies have shown that cognitive processes such as memory, attention and decision–making are biased by underlying emotional state in humans (Roiser et al., 2012, Neuropsychopharmacol., 37:117–136). Similar affective biases have been observed after pharmacological and social manipulations of the affective state in non-human animals, such as rats (Stuart et al., 2013, Neuropsychopharmacol., 38:1625–35). Also, there is some evidence that changes in circulating hormone levels influence emotional processing, mood and memory (Swaab et al., 2005, Ageing Res Rev., 4(2):141–94). The aim of our study was to investigate how hormonal manipulations influence learning and recall of two independent experiences using the affective bias test in rats (Stuart et al., 2013, Neuropsychopharmacol., 38:1625–35). Male Sprague Dawley rats (Charles River, N=12) were trained in this bowl digging task, that involves repeated pairing of a rewarding outcome (obtaining a food reward) with a neutral stimulus (specific digging substrate) under control conditions or following a hormonal manipulation. The value of each rewarding experience was equal and all factors (i.e. bowl location, substrates) were fully counterbalanced. The affective bias is measured using a choice test where both previously rewarded substrates are presented together over 30 randomly reinforced trials. Animals were tested using acute treatment with gonadal hormones (oestradiol, 0.0, 1.0, 10.0µg/kg, SC vs vehicle treatment; progesterone, 0.0, 1.0, 10.0mg/kg, SC vs vehicle treatment; testosterone, 0.0, 1.0, 10.0mg/kg, SC vs vehicle treatment), androgen receptors antagonist (flutamide, 0.0, 1.0, 10.0mg/ kg, SC vs vehicle treatment), aromatase inhibitor (formestane, 0.0, 1.0, 10.0mg/kg, SC vs vehicle treatment), xenoestrogen and endocrine disruptor (bisphenol A, 0.0, 0.05, 0.5mg/kg, oral vs vehicle treatment), and oxytocin and vasopressin analogues (carbetocin 0.3mg/kg and desmopressin 0.1mg/kg, respectively, SC vs vehicle treatment). Results were analysed using a Repeated Measures ANOVA and post-hoc, one-sample t-test against the null hypothesised mean of 0% choice bias. Acute treatment with oestradiol (F2,11=4.033, p=0.0322), flutamide (F2,10=4.593, p=0.0228), bisphenol A (F2,11=6.339, p=0.0082), and carbetocin (F2,11=6.875, p=0.0048) resulted in a positive affective bias, whilst progesterone (F2,11=4.592, p=0.0215), testosterone (F2,10=19.80, p=0.0001), and formestane (F2,11=6.897, p=0.0047) induced negative affective bias in Sprague Dawley rats. Only desmopressin treatment showed no significant effect. Our results support recent findings that hormones play a role in modulation of affective processes and demonstrated that acute hormonal manipulations alter the formation of affective biases in rats. JKP was supported by UoB Studentship, and also MRC and BBSRC project grants awarded to ESJR. A68 ABSTRACTS

C06 INVESTIGATION OF REPRODUCTIVE AND LITTERING PARAMETERS IN THE OLFACTORY BULBECTOMY RODENT MODEL OF DEPRESSION Doherty H, Dept. of Pharmacology and Therapeutics, NUI Galway, Galway, Ireland, 0000 [email protected] Kelly JP(1) (1) Dept. of Pharmacology and Therapeutics, NUI Galway Depression is a common neuropsychiatric disorder that affects roughly twice as many more women than men and as a consequence is commonly encountered during or shortly after pregnancy. Maternal depression has been shown to have significant effects on the physical health of offspring, causing problems such as low birth weight, growth retardation and chronic illness. The aims of this study were to determine if there are any differences in reproductive or birth parameters in a commonly used rat model of depression, namely the olfactory bulbectomized (OB) model of depression. Male and female Sprague-Dawley rats received olfactory bulbectomy or sham surgery and 5 weeks thereafter were divided into 4 mating groups (namely male sham x female sham (n=15); male sham x female OB (n=15); male OB x female sham (n=15); male OB x female OB (n=16)). Following mating, females were singly housed and the reproductive performance , maternal weight gain and litter characteristics were recorded. Data was analysed using a 2 way ANOVA followed by post hoc test or by Chi Square test where appropriate. The results showed that of the 55 females that tested positive for sperm after mating, OB females (n=28) were less likely to become pregnant than their sham operated counterparts (n=27) (p<0.05). The OB females mated with OB males (n=11) put on the least weight over the course of the pregnancy compared to the female sham groups (p<0.05). At littering, the number of pups born to OB mothers was significantly lower than the number born to sham operated rats (p<0.001); these pups weighed more at birth compared to those from litters with sham mothers (p<0.01). There were no differences in the number of male pups born in the groups but there were significantly less female pups born in the groups with OB operated mothers (p<0.001). There were no differences in the number of stillborn pups or the number that were killed/died in the immediate neonatal period. In conclusion, we found that although there were no differences in the reproductive performance of sham and OB rats, the OB operated females were less likely to see pregnancy to term, and of those that did lower litter numbers resulted. These findings have a clinical relevance in that maternal depression affects birth parameters suggesting that the OB model may be of value in modelling depression in pregnancy. This study was financially supported by a postgraduate fellowship from the College of Medicine, Nursing and Health Sciences, National Univ of Ireland, Galway.

C07 THE WISTAR-KYOTO MODEL OF DEPRESSION SHOWS IMPAIRED ULTRASONIC VOCALIZATION RESPONSE IN THE FEMALE URINE SNIFFING TEST WHEN COMPARED WITH SPRAGUE-DAWLEY Di Capua G, Transpharmation Ireland Ltd, 3.47 Lloyd Building, Trinity College Dublin, Dublin 2, Ireland [email protected] Prenderville JA(1), McDonnell C(1), Bianchi M(1), Rouine J(2) (1) As presenting author; (2) Trinity College Inst of Neuroscience, Room 3.14 Lloyd Building, Trinity College Dublin, Dublin 2, Ireland Rats use ultrasonic vocalizations (USVs) to communicate. The 50kHz-USVs have positive value and are emitted in reward states including social interactions. The female urine sniffing test is used to assess hedonic behaviour in male rodents and consists in measuring the time spent sniffing oestrus female urine (Malkesman et al, 2010; Biol. Psychiatry; 67:864–871). Current preclinical assays have limitations in measuring anhedonia, a key feature of depression. Our aim was to develop a new assay by simultaneously measuring sniffing and 50kHz-USVs in the endogenous “depressed” Wistar-Kyoto rat (WKY) compared to Sprague-Dawley (SD) as “healthy” control strain when exposed to female urine. Six-month-old WKY (n=10,350–300g) and SD (n=8,350–400g) were singly-housed 7 days before the test to increase social motivation. Urine collected from animals of the same strain was kept frozen at -80°C until use. One hour before the test a cotton tip was inserted in the animals’ home cage for habituation. They were then ABSTRACTS A69 transferred to a dimly-lit (3lux) room to perform a three phases test consisting of being exposed for three minutes to a cotton tip infused with 60μl of i) sterile water (W); ii) oestrus female urine (FU); iii) male urine (MU; used as control odour). Each phase was preceded by 5min of habituation and separated by an interval of 45min. Sniffing time was measured (s) while USVs were recorded and counted using the Sonotrack equipment (Metris). Data were analysed with Two-Way ANOVA for repeated measures followed by Fisher’s LSD test. SD emitted significantly (p<0.001) more 50kHz-USVs during the exposure to FU (FU: 39±8.33 vs. W: 0.13±0.22 vs. MU: 3.38±1.26), although they sniff MU (55.30±13.88) more than FU (41.13±12.21). SD emitted a significantly (p<0.001) higher number of 50kHz-USVs during exposure to FU (SD: 39±8.33 vs. WKY: 0, p<0.001), although no difference was observed in sniffing time between SD (41.13±12.21) and WKY (27.34±8.99). SD spent significantly (p<0.01) more time sniffing MU compared to WKYats r (SD: 55.3±13.88 vs. WKY: 16.41±7.14), but the emission of USVs wasn’t different (SD: 3.37±1.26 vs. WKY: 0). In conclusion, WKY rats demonstrate a weaker positive response to a purported rewarding stimulus such as oestrus urine which might be analogous to the anhedonia observed in depression. Male urine could represent an attractive social stimulus for SD but not rewarding as the oestrus urine. The validity of our assay will be tested by measuring 50kHz-USVs after antidepressant treatment. Supported by internal Transpharmation Ireland funding.

C08 CHARACTERISING THE PREDICTIVE VALIDITY OF REARING RATS IN ISOLATION FROM WEANING AS A MODEL FOR PSYCHIATRIC DISORDERS Dunphy-Doherty F, School of Life Sciences, Univ of Nottingham, Medical School, Queen’s Medical Centre, Nottingham NG7 2UH [email protected] King MV(1), Wigmore P(1), Fone KCF(1) (1) As presenting author Rearing rats in social isolation (SI) from weaning causes persistent behavioural and neurochemical alterations resembling changes seen in both depression and schizophrenia (Fone & Porkess 2008, Neurosci Biobehav Rev, 32, 1087–1102). Antipsychotics reverse several components of the isolation syndrome in rats but few studies have investigated sensitivity to antidepressants. This study determined whether chronic administration of the selective serotonin , fluoxetine (FLX), prevents development of the behavioural isolation syndrome (enhanced responsivity to stressors and memory deficits) to further characterise the predictive validity of this neurodevelopmental model. Methods: Sixty male Lister-hooded rats were either housed in littermate groups (GH n=3) or alone (SI) from weaning (post-natal day 21). On PND 42 half the GH and SI rats (n=15 per group) received FLX (10mg/kg/day) continuously in drinking water and half water (CON) and all received BrdU (PND 47, 150 mg/kg i.p.) and a subcutaneous temperature transponder. From PND 54 behaviour was measured in an open field, infra-red activity chamber and in novel object discrimination (NOD), elevated plus maze (EPM), conditioned freezing response (CFR) and restraint stress paradigms at one week intervals. Immediately after the last test plasma was collected for corticosterone analysis and the brain hemisected, fixed and processed for doublecortin (DCX, neuronal marker) immunohistochemistry in the hippocampal dentate gyrus (DG). Results: In the open field there was a main effect of housing and FLX on distance, velocity and central zone time (P<0.05); SI-FLX rats having highest velocity and travelling furthest. SI-CON were unable to discriminate novel from familiar objects in NOD (P=0.07); a deficit reversed with FLX (P<0.01). In the EPM SI-CON spent less time in the more aversive open arms (P<0.05) but this was unaltered by FLX. At 24 and 48h post-footshock conditioning in the CFR task SI-CON froze significantly less than GH-CON which was reversed by FLX. Restraint significantly elevated body temperature (stressed-induced hyperthermia, P<0.001) irrespective of housing or treatment and plasma corticosterone levels in the SI-FLX group (P<0.05). Plasma corticosterone was unaltered by housing or treatment. Analysis of DCX showed a main effect of housing (P<0.001) and treatment (P<0.01); FLX significantly increased the number of DG DCX cells in both GH and SI. Conclusions: Chronic FLX appeared to attenuate SI-induced anxiety in the open field but not the EPM, partially reversed visual learning and memory (NOD) and associative learning (CFR) deficits and increased DG neurogenesis. FD is funded by a Marie Curie Initial Training Network grant (rBIRTH). A70 ABSTRACTS

C09 BLOCKADE OF HIPPOCAMPAL 5-HT7 RECEPTORS ATTENUATES STRESS-INDUCED DEFICITS ON EXPLORATORY ACTIVITY Padovan CM, Psychology, FFCLRP-Univ of São Paulo, Avenida Bandeirantes, 3900 Ribeirão Preto SP, Brazil 14040901 [email protected] Prado TAR(1) (1) As presenting author Serotonin has been implicated in the aetiology of mood disorders. Serotoninergic projections from the Median Raphe Nucleus (MnRN) to the dorsal hippocampus (dH) have been considered an important mechanism on the disconnection of aversive memories. In animal models of depression, activation of 5-HT1a receptor in the hippocampus and median raphe nucleus has antidepressant effects. Recently, 5HT7 receptors (5-HT7) were implicated on the neurobiology of depression, since blockade of these receptors also induces antidepressant effects. However, the brain sites involved have not yet been elucidated. Therefore, the aim of this work was to investigate the role of hippocampal 5-HT7 on the restraint stress-induced deficits on exploratory activity of an Elevated Plus Maze (EPM). All procedures were approved by the Animal Research Ethics Committee (#2014.1.231.53.9). Male wistar rats (7 weeks old) with bilateral cannula aimed to the dorsal hippocampus were used for this study. Doses used in this study were chosen based on a first experiment where rats received simultaneous bilateral injections of Saline (SAL; 0.5µL) or SB 258741 (SB; 1, 3 or 10 nmoles/0.5µL) and were tested on the EPM five minutes later. The number of enclosed arm entries (EAE) and percentage of entries (OAE) and time spent (TOA) in the open arms were analyzed by ONEWAY ANOVA followed by Duncan test, considering significant p<0.05. In other experimental conditions, immediately before or after being restrained in a wire chamber with an adjustable roof for 2h, rats received bilateral injections of SAL or SB 3nmoles/0.5µL and were tested on the EPM 24h later. In control condition rats were treated as described but were not exposed to restraint. For these conditions, data was analyzed by Student’s t test. After test rats were sacrificed and had their brain removed for histological analysis of the injection site. The dose-response protocol showed that SB 1nmol significantly (p<0.05) decreased OAE (F3,23=2.51) and TOA (F3.23=2.64) when compared to SAL-treated rats. Restraint-induced deficits were prevented by treatment with SB 3 nmoles (OAE: t16=-2.49; p<0.05; TOA: t16=-1.98; p<0.05), without changing EAE (t16=-0.81; p>0.05). SB also attenuated the effects of stress on open arms exploration (OAE: t15=-4.93; p<0.05; TOA: t15=-5.15; p<0.05), but did not change EAE (t15—0.23; p>0.05). No differences were detected between treatment on the control condition (EAE: t10=-1.14; p>0.05; OAE: t10=0.05; p>0.05; t10=0.34; p>0.05). Our data suggest that the neurobiological processes underlying stressful memories involve activation of 5-HT7 and that blockade of these receptors has antidepressant effects. Financial Support: FAPESP (2014/20837-5).

C10 EFFECTS OF GABA SUPPLEMENT ON GABA, GLUTAMATE AND GLUTAMINE CONCENTRATIONS IN THE FRONTAL CORTEX AND HIPPOCAMPUS IN RATS UNDERGOING FORCED SWIM STRESS Thaweethee B, Dept. of Anat. and Centre of Excellence in Medical Biotechnology, Faculty of Medical Science, Naresuan Univ, Muang Phitsanulok, Phitsanulok, Thailand, 65000 [email protected] Thanoi S(1), Nudmamud-Thanoi S(1) (1) As presenting author Introduction: Abnormalities of GABAergic neurotransmission have been reported in psychiatric disorders such as anxiety and depression [Cryan and Kaupmann, 2005, Trends in Pharmacological Sciences, 26(1), 36–43]. Recently, GABA supplement has been widely used for improvement of stress, anxiety and mood. However, reliable scientific evidence to support this use is lacking [Boonstra and et al., 2015, Frontiers in Psychology, 6, 1–6]. Therefore, this study aimed to determine the effects of GABA supplement on GABA, glutamate and glutamine concentrations in the frontal cortex and hippocampus in the forced swimming test (FST)-induced depression-like behaviours in rats. Methods: Male Sprague-Dawley rats were divided into three groups: control, stressed and GABA-treated groups. GABA supplement was dissolved in distilled water (DW) and given at dosage of 0.8mg/kg/day for 28 days before FST exposure, while the control and stressed groups were administrated orally with DW. Rats in stressed and GABA-treated groups were ABSTRACTS A71 subjected to standard FST to induce a depressive-like behaviour such as immobility. Immobility in the stressed group was classified into low-immobility (LI) and high-immobility (HI) subgroups. Glutamine, glutamate and GABA concentrations in frontal cortex and hippocampus were investigated using HPLC with electrochemical detection. The data were analysed by one-way ANOVA followed by Fisher’s LSD test. Results: This study found significant decrease in glutamate (P<0.01) and increase in GABA (P<0.01) in the frontal cortex of LI group compared with the control group. Both HI and LI groups showed a significant decrease of glutamine in the hippocampus compared with the control group (P<0.05 in both groups). However, there was no significant difference of glutamine, glutamate and GABA between the GABA-treated group and the control group in either frontal cortex or hippocampus. Interestingly, significant increase in glutamate (P<0.05) and decrease in GABA (P<0.05) in the frontal cortex were observed in the GABA- treated group compared with LI group. Conclusion: This study found changes of glutamate and GABA in the frontal cortex in response to FST. These findings indicate that the frontal cortex may be a major region, which regulates stress response. However, GABA supplement can prevent alterations of glutamate and GABA in the frontal cortex, and glutamine in the hippocampus. Nevertheless, the mechanism whereby GABA supplement can induce effects on GABA, glutamate and glutamine in the brain needs to be study further. This study was financially supported by the Agricultural Research Development Agency, the National Research Council of Thailand and the Royal Golden Jubilee Ph.D Programme.

C11 PLASMA ACETYLATED A-TUBULIN EXPRESSION AS AN INDICATOR OF ANTIDEPRESSANT EFFECTIVENESS McDonnell CW, Transpharmation Ireland Ltd , 3.47 Lloyd Building, Trinity College Dublin, Dublin, Ireland 2 [email protected] Prenderville JA(1), Rouine J(1), DiCapua G(1), Bianchi M(1), McKenna JP(2), McCreary C(2), O’Driscoll D(3), Downer EJ(3) (1) As presenting author; (2) Cork Univ Dental School and Hospital, Univ College Cork, Cork; (3) Dept of Physiology, Trinity College Dublin Microtubules (MTs) are the primary component of the cytoskeleton. Neuronal MTs are fundamental to dendritic remodelling and neuronal plasticity. Post-translational modifications to α-tubulin, a main constituent of MTs, can be measured to assess MT dynamic states. Alterations in brain acetylated α-tubulin (Acet-Tub; marker of decreased MT dynamics), have been implicated in major depressive disorder (MDD). Hippocampal Acet-Tub expression was increased in rat models of depression and rescued by antidepressant treatment (Bianchi et al., 2009, PNAS 109:1713–1718). Here we show that Acet-Tub is detectable and measurable in both rat and human plasma, offering a way to translate findings in animal models to clinical samples. Plasma samples from both a rat model and clinical volunteers were used to validate Acet-Tub as a potential indicator of pharmacological efficacy. A limited number of clinical samples were obtained from patients affected by Burning Mouth Syndrome (BMS), a neuropathic disorder having high comorbidity with MDD. Male Wistar rats (3–4months;300–350g;n=10/group) received fluoxetine (10mg/kg,i.p.) or saline (1ml/kg,i.p.) 24h, 5h, and 1h before forced swimming test (FST) to verify antidepressant efficacy of fluoxetine as decreased immobility. Animals were sacrificed immediately following FST and plasma samples were obtained from trunk blood. Plasma was collected from venous blood of healthy control participants (n=10) and BMS patients (n=10). BMS samples were divided into two groups based on whether patients received antidepressant treatment (n=4) or not (n=6). Infrared Western Blotting (IFWB) was used to measure plasma expression (optical density) of Acet-Tub normalised to transferrin and expressed as % of control. Data were analysed using a student’s t-test or one-way ANOVA. Fluoxetine reduced immobility (s) in FST (Wistar+saline:133.00±8.36 vs. Wistar+fluoxetine:100 .90±8.59,p<0.05), showing antidepressant efficacy. Acet-Tub expression (%control) was decreased in rat plasma following fluoxetine treatment (Wistar+saline:100±9.47 vs. Wistar+fluoxetine:62.46±6.62, p<0.01). IFWB protocol for analysing Acet-Tub in human plasma was optimised to a signal linearity of R2=0.9685. Plasma Acet-Tub expression (%control) showed a tendency to be increased in BMS patients not taking antidepressants compared to healthy controls (Control:100±9.97 vs. BMS:129.4±16.51,p=0.1210). Plasma Acet-Tub expression in BMS patients taking antidepressants showed a clear tendency to be decreased A72 ABSTRACTS compared to BMS patients not taking them (BMS:129.4±16.51 vs. BMS+Antidepressant:83.83±12.01,p=0.07 97). The results show that plasma Acet-Tub may represent an indicator of antidepressant efficacy in rats. Acet-Tub can be measured in human plasma and the first clinical data on a limited number of samples suggests the translational validity of Acet-Tub as an indicator of antidepressant efficacy. Supported by internal Transpharmation Ireland Ltd. funding.

C12 ANTICIPATORY VERSUS CONSUMMATORY AROUSAL: THE ROLE OF THE SUBGENUAL CINGULATE IN ANHEDONIA Alexander L, Physiology, Development and Neuroscience, Univ of Cambridge, Downing Street Cambridge CB2 3EG [email protected] Clarke HF(1), Roberts AC(1) (1) Dept of Physiology, Development and Neuroscience, Cambridge CB2 3EG INTRODUCTION: Anhedonia is defined as a lack of ability to experience pleasure and is a core symptom of Major Depressive Disorder (MDD). Whilst clinical descriptions focus on the consummatory aspect of anhedonia, it is now recognised that it also has anticipatory and motivational elements. Although caudal sgACC has been shown to be over-active in depression, whether this over-activity causes symptoms of anhedonia and if so, what aspects, is poorly understood. We aimed to clarify the role of marmoset caudal sgACC during reward anticipation (conditioned, CS) and consumption (unconditioned, US) in an appetitive Pavlovian conditioning paradigm. In doing so we hoped to gain insight into how dysfunction of this region in humans could contribute to anticipatory and consummatory anhedonia. METHODS: Five common marmosets (three female, two male; average age 4 years) were implanted with telemetry probes for cardiovascular monitoring and intra-cranial cannulae targeting caudal sgACC. Animals were trained to discriminate between two auditory cues - a CS+ predicting food reward, and a CS- predicting no reward - on an appetitive Pavlovian conditioning paradigm. Once animals learnt the discrimination, caudal sgACC was inactivated (muscimol-baclofen) or over-activated (using either CGP-52432/LY-341495 or Dihydrokainic acid [DHK]) once a week prior to a session and the effect on cardiovascular and behavioural anticipatory and consummatory arousal was measured. RESULTS: Temporary over-activation of marmoset caudal sgACC resulted in a significant reduction in autonomic (systolic blood-pressure, p<0.05) and behavioural (orienting, p<0.05) anticipatory arousal. There was no effect on autonomic consummatory arousal or behavioural measures of consumption (latency to retrieve food reward or amount consumed). Inactivation of caudal sgACC had no effect on either anticipatory or consummatory arousal. CONCLUSIONS: These results demonstrate a potential link between over- activity in caudal sgACC - as is commonly seen in depressed patients - and anticipatory anhedonia. The anhedonic state induced by over-activation of caudal sgACC resembles the anhedonic state in depressed subjects, where deficits tend to be in the anticipation of reward rather than in consumption. Future work will identify the downstream changes associated with the anticipatory anhedonia induced by over- activation of this brain region. FUNDING STATEMENT: The research was funded by a Medical Research Council (MRC) Programme Grant to A.C.R. (MR/M023990/1). L.A. was funded by an MRC Studentship. H.F.C. was funded by an MRC Career Development Award.

C13 RELIABILITY OF AMYGDALA AND SUBGENUAL ANTERIOR CINGULATE CORTEX RESPONSES DURING EMOTIONAL FACE PROCESSING: IMPLICATIONS FOR FMRI BIOMARKERS Nord CL, Inst of Cognitive Neuroscience, Univ College London, 17 Queen Square, London WC1N 3AR [email protected] Alan G(1), Robinson OJ(1), Roiser JP(1) (1) As presenting author Introduction: There is considerable need to develop tailored approaches to psychiatric treatment. Numerous researchers have proposed using functional magnetic resonance imaging (fMRI) biomarkers ABSTRACTS A73 to predict therapeutic response. However, using fMRI for clinical research depends critically on the reliability and stability of the specific measure used. Methods: We scanned a group of 30 volunteers twice on a 1.5T MRI scanner (mean test-retest interval=14.3 days). We employed three tasks, all of which present emotional faces, but require different types of processing: gender classification (O’Nions et al., 2011, Journal of Cognitive Neuroscience 23:11,3681–3693), emotion identification (Robinson et al., 2012, NeuroImage,60:1, 523–529), and emotion matching (Hariri et al., 2002, NeuroImage, 17:1,317–23). Each task was performed twice on each day. We extracted the mean BOLD response during emotional face processing (versus baseline) in spheres around the group-level peak voxels, for each task separately, in the left amygdala, right amygdala, and the subgenual anterior cingulate cortex (sgACC). We then calculated the reliability of responses using the intra-class correlation coefficient (ICC), both within- session and between the two fMRI sessions. Results: As expected, all three tasks evoked reliable activation bilaterally in the amygdala, and deactivation in the subgenual cingulate (all peak Zs>4.5). In most cases, the amplitude of the BOLD response was not reliable between the first and second days. However, the left amygdala response during gender classification had a statistically reliable between- day ICC of 0.408 (p=0.027, 95%CI=-0.009—0.0653). A similar pattern of poor reliability was apparent for within-day reliability, with the gender classification task again the exception: both right and left amygdala showed reliable within-session activations on the second day (left amygdala ICC 0.653, p=0.004, 95%CI=0.25—0.839; right amygdala ICC 0.558, p=0.019, 95%CI=0.044—0.795). Interestingly, the between-day reliability of left amygdala activation during gender classification was driven by responses in the second runs (ICC 0.515, p=0.032, 95%CI=-0.04—0.776); for the first runs, measurement was not reliable across days (ICC 0.281, p=0.199, 95%CI=-0.554—0.667). Conclusions: We found surprisingly low statistical reliability for two putative fMRI biomarkers in psychiatric research, amygdala and sgACC response to emotional faces. Yet amygdala activation during gender classification showed moderate reliability, especially following initial task exposure, suggesting that fMRI reliability may be related to the specific cognitive activation paradigm employed. These data present a cautionary tale for the use of task-evoked amygdala and sgACC fMRI biomarkers as predictors of therapeutic response. This research was funded by the Wellcome Trust and the Brain Research Trust.

C14 KISSPEPTIN MODULATION OF LIMBIC BRAIN ACTIVITY AND MOOD IN HEALTHY MEN Demetriou L, Imanova Ltd, Burlington Danes Building, Imperial College London, Hammersmith Hospital, Du Cane Road, London W12 0NN [email protected] Mehta AM(1), Demetriou LD(2), Wall MW(3), Tanner MT(3), Rabiner EAR(3), Comninos AC(4), Shah SJA(4), Clarke SC(4), Narayanaswamy SN(4), Nesbitt AN(4), Izzi-Engbeaya CIE(4), Abbara AA(4), Ratnasabapathy RR(4), Salem VS(4), Nijher GMN(4), Jayasena CNJ(4), Bloom SRB(4), Dhillo WSD(4) (1) (3) Dept of Radiology, Imperial College Healthcare NHS Trust; (2) As presenting author; (3) Imanova Centre for Imaging Sciences, Imperial College London, W12 0NN; (4) Investigative Medicine, Imperial College London, W12 0NN Introduction: Kisspeptin is a crucial activator of reproductive function. It plays a critical role in the hypothalamus to activate GnRH neurons and downstream reproductive hormones. However, kisspeptin and its receptor are also expressed in limbic brain areas although little is known about their function here. Kisspeptin signalling in the amygdala modulates neuronal activity and reproductive hormone secretion in rodents, but this has not been investigated in humans. We tested whether kisspeptin modulates limbic brain activity and mood in healthy men using fMRI. Methods: 31 healthy heterosexual men (mean age 25.4±1.0y) completed a randomized blinded two-way placebo-controlled protocol. To stimulate limbic brain activity and determine if kisspeptin administration altered this response we used an Emotional Picture Task. Images fell into four themes: sexual, couple-bonding, negative and neutral. The participants viewed the images for 3 seconds each in a random order and rated their pleasantness on a 5-point scale. To investigate any global effects of kisspeptin on the brain a fast (5 minutes, 100 random trials) fMRI battery control task was used, consisting of visual, audio, motor, calculation and language tasks. Blood samples were collected throughout for analysis of reproductive hormones. Finally the participants also completed psychometric questionnaires designed to assess A74 ABSTRACTS sexual and emotional factors. Results: Kisspeptin administration resulted in an increase in circulating kisspeptin (p<0.0001), LH (p<0.0001) and FSH (p=0.0005) but not testosterone (p=0.180) for the duration of the scans, as expected. A Region of Interest (ROI) analysis was employed to investigate the BOLD activity in limbic regions containing kisspeptin receptors. Results revealed that kisspeptin (vs. placebo) significantly enhanced (p<0.05) activation in key limbic and para-limbic structures (including the amygdala and left putamen) on viewing sexual images. Additionally, kisspeptin-enhanced activity in putamen was significantly correlated with reduced sex-avoidance (r=-0.63, p=0.002). Viewing non- sexual couple-bonding images also resulted in increased activity in limbic areas (p<0.05), including left amygdala, left putamen, globus pallidus and thalamus. Increased amygdala activation by kisspeptin was also related to improved positive mood (r=0.67, p=0.001). Results of the fMRI battery control task revealed no significant differences in BOLD activation between kisspeptin and placebo. Conclusions: Collectively, these data provide the first evidence that kisspeptin administration modulates limbic brain activity in response to sexual and emotional stimuli and can influence mood in healthy men. These data have important implications for our understanding of reproductive biology and behaviour, and the development of kisspeptin as a potential therapeutic. The research was supported by the NIHR/Wellcome Trust CRF at Imperial College Health-care NHS Trust.

C15 POSITIVE IMAGERY COGNITIVE TRAINING INCREASES ROSTRAL ANTERIOR CINGULATE ACTIVITY IN HEALTHY OLDER ADULTS Murphy SE, Dept of Psychiatry, Univ of Oxford, Warneford Hospital, Oxford OX3 7JX [email protected] O’Donoghue MC(1), Nobre AC(1), Browning M(1), Blackwell SE(2), Holmes EA(3) (1) As presenting author; (2) Dept of Psychology, Ruhr Univ Bochum, Massenbergstraße 9–13, 44787, Bochum, Germany ; (3) MRC Cognition and Brain Sciences Unit, 15 Chaucer Road, Cambridge, CB2 7EF Introduction: The ability to form positive mental images about the future may be a key aspect of mental health and wellbeing. Cognitive interventions targeted at shifting biases in the processing of emotional information away from negative and towards positive are attracting increasing focus as a potential approach to the prevention or treatment of mood disorders. One such approach is to train individuals to automatically imagine positive resolutions of ambiguous information. We have recently demonstrated that it is possible to increase the vividness of positive prospective imagery using computer-based positive imagery cognitive training in healthy older adults (Murphy et al, 2015, Psychiatry Research 230(1):36–43). The rostral anterior cingulate cortex (rACC) plays a key role in the simulation of future affective episodes. The aim of this study was to investigate whether the increased vividness of positive imagery reported by older adults following positive imagery training was also reflected by altered functional brain activity in the rACC. Methods: 75 participants received 4-weeks of positive imagery training or control training. At the end of training, all participants completed an Ambiguous Sentences Task during an fMRI scan. In this task, participants were presented with ambiguous scenarios and instructed to generate mental images in response. Participants were then presented with an outcome disambiguating the initial scenario, which could be either positive or negative. fMRI data were analysed using FSL. Z (Gaussianised T/F) statistic images were thresholded using clusters determined by Z>2.3 and a (corrected) cluster significance threshold of P=0.05. Results: rACC activity was positively correlated with the pleasantness ratings of participants’ images formed in response to a cue describing an ambiguous future event. 4 weeks of positive imagery training increased this rACC activity compared with control training. Bilateral hippocampal activity was also increased in the positive imagery group compared with the control group. Conclusions: The current study demonstrates that rACC activity during positive future imagery is malleable to change by cognitive training. This is consistent with previous evidence that this training enhances the vividness of positive imagery about the future and lends weight to the idea that it is acting to increase the intensity and affective quality of imagery about the future. The rACC is involved in the cognitive control of emotion and previous studies have linked rACC activity with an attentional positivity bias in successful ageing. Conversely, late life depression has been associated with rACC dysfunction. Therefore, any intervention that is able to effectively increase rACC functioning may be both a useful therapeutic target in late-life ABSTRACTS A75 depression, and a useful approach for maintaining and promoting emotional wellbeing and successful ageing. This research was funded by the NIHR Oxford Biomedical Research Centre.

C16 EFFECTS OF THE POTENTIAL LITHIUM-MIMETIC, EBSELEN, ON BRAIN NEUROCHEMISTRY: A MAGNETIC RESONANCE SPECTROSCOPY STUDY AT 7 TESLA Masaki CO, Psychiatry, Oxford, Warneford Hospital, Oxford OX3 7JX, OX3 7JX [email protected] Singh N(1), Vasudevan SR(1), Churchill GC(1), Sharpley AL(2), Godlewska BR(2), Hashimoto T(2), Cowen PJ(2), Berrington A(3), Emir UE(3) (1) Dept of Pharmacology, Univ of Oxford, Mansfield Road, Oxford OX1 3QT; (2) Neurosciences Building, Warneford Hospital Headington, OXFORD OX3 7JX; (3) The Oxford Centre for Functional MRI of the Brain, Nuffield Dept of Clinical Neurosciences, Univ of Oxford, John Radcliffe Hospital, Oxford,X3 O 9DU Introduction: Lithium remains the most effective treatment for bipolar disorder but tolerance and safety issues complicate its clinical use (McKnight et al. 2012, Lancet. 379:721–728). The antioxidant drug, ebselen, has been proposed as a possible lithium-mimetic based on its ability in animals to inhibit inositol monophosphatase (IMPase) and lower brain inositol, actions which it shares with lithium (Singh et al. 2013, Nature communications. 4:1332). We set out to determine whether treatment with ebselen lowered levels of inositol in the human brain. We also assessed the effect of ebselen treatment on other brain neurometabolites, including glutathione, glutamate, glutamine, and glutamate+glutamine (Glx). Methods: We studied 20 healthy volunteers who were tested on two occasions receiving either ebselen (3600mg over 24 hours) or identical placebo in a double-blind, random-order, cross-over design. Two hours after the final dose of ebselen/placebo, participants underwent proton magnetic resonance spectroscopy (1H MRS) at 7 tesla (7T) with voxels placed in anterior cingulate and occipital cortex. Neurometabolite levels were calculated using an unsuppressed water signal as a reference and corrected for individual cerebrospinal fluid content in the voxel. Results: Ebselen treatment was well tolerated and no participants dropped out of study. Ebselen produced no effect on neurometabolite levels in the occipital cortex. The MANOVA for the anterior cingulate cortex showed a main effect of ebselen treatment (F= 12.48; p= 0.003) and a significant interaction between treatment and neurometabolite (F= 3.38; p= 0.044). Follow-up pairwise comparisons revealed that ebselen decreased inositol concentrations within this region (p = 0.028). There were also significant reductions in glutathione (p = 0.033), glutamine (0.024), glutamate (0.010) and Glx - a composite of glutamate and glutamine (0.001). Conclusions: The study suggests that at the dosage used, ebselen produces a functional inhibition of IMPase in the human brain. The ability of ebselen to lower indices of glutamate activity are consistent with its action, reported in animal experimental work, to inhibit the enzyme, glutaminase. Ebselen appears to have potential as a repurposed treatment for bipolar disorder and it would be of interest to see if similar biochemical alterations are produced by ebselen treatment in this patient group. The study was supported by the MRC (Grant MR/K022202/1).

C17 THE EFFECTS OF 7-DAYS TREATMENT WITH AGOMELATINE ON THE NEURAL RESPONSE TO REWARD AND AVERSION IN HEALTHY VOLUNTEERS Dean ZPC, Psychology, Univ of Reading, Whiteknights Campus, Reading RG6 6BE [email protected] Antonesei A(1), Jayabalan P(1), Butt M(1), McCabe C(1) (1) Univ of Reading, Whiteknights Campus, Reading, RG6 6BE It has been suggested that dopaminergic (DA) and noradrenergic (NE) antidepressants might be more effective at treating reward-related deficits in depression than serotonergic (5-HT) antidepressants. eW have previously found that the selective 5-HT re-uptake inhibitor, citalopram, reduces the neural response to reward and aversion in healthy volunteers. Agomelatine increases DA and NE via 5-HT2C antagonism but is yet to have its effects on reward and aversion processing in the human brain fully elucidated. 30 volunteers will be recruited and will receive 7 days of agomelatine (25 mg/day) and 7 days of placebo treatment, in a double-blind crossover design. On the 8thmorning of each medication phase, volunteers A76 ABSTRACTS undergo functional Magnetic Resonance Imaging (fMRI). Volunteers are presented with a pleasant or unpleasant cue (anticipatory phase) before investing effort via button pressing (effort phase), in order to obtain a pleasant taste or to avoid an unpleasant taste (consummatory phase). Contrast analyses examined drug vs. placebo activity during each phase. Preliminary results with 8 volunteers reveal that relative to placebo, agomelatine increased activity in the thalamus during reward consummation. During aversion processing, agomelatine increased activity in the amygdala and dorsal caudate during anticipation and reduced activity in the precentral gyrus* and bilateral insula during consummation. All p< 0.028 FWE svc for multiple comparisons. * Whole brain cluster corrected p<0.05 FWE for multiple comparisons. Our results suggest that agomelatine may act as an antidepressant by increasing neural activity related to positive information and reducing neural activity to negative information. It will be of interest to next examine the neural response to reward and aversion in a depressed sample before and after treatment with agomelatine, to elucidate if improvements in mood are associated with these neural changes. This study was sponsored by the Univ of Reading.

C18 THE EFFECTS OF THE NMDA RECEPTOR ANTAGONIST MEMANTINE ON EMOTION RECOGNITION IN HEALTHY VOLUNTEERS Upton N, Dept of Psychology & Dept of Psychiatry, Univ of Southampton, Highfield Campus, Southampton SO17 1BJ [email protected] Penton-Voak I(1), Munafo MR(1), Hughes M(2), Bamford S(2), Pinkney V(2), Woolley J(2), Baldwin DS(2), Garner M(2) (1) Univ of Bristol; (2) Univ of Southampton Introduction: The moderate-affinity NMDA receptor antagonist memantine is clinically used in Alzheimer’s disease (AD) and has positive effects across cognitive and behavioural symptoms. Pre-clinical studies suggest memantine may also have therapeutic potential in other conditions. Neuropsychological models of psychotropic drug action suggest therapeutic effects might be mediated by the early correction of maladaptive biases in cognitive and emotion processing, for example hyper-sensitivity to negative emotional information. Method: 36 healthy volunteers were randomised to receive either a 2-week course of memantine (5mg titrated to 10mg on day 7) or placebo (balanced for gender, administration double-blind). On day 0 and day 14 participants completed a computerized emotional expression recognition task that measured sensitivity to the six primary emotions; sadness, happiness, anger, fear, disgust and surprise. Results: Memantine, compared to placebo, selectively reduced the recognition of anger F(5,160) = 2.79, p = .019, np2 = .08. Follow- up analysis of Hits (correct detection of anger) and False Alarms (incorrect of anger) show that detection of anger reduced in the memantine group (t(16) = 2.62, p = .018) and was lower than the placebo group on day 14 (t(32) = 2.82, p = .008). There was no clear evidence that memantine altered the recognition or incorrect classification of other emotions. The effect of memantine on emotion recognition occurred in the absence of changes in reported mood. Conclusions: The effect of sub-chronic memantine administration on the recognition of anger indicates a mechanism through which memantine may achieve therapeutic effects in other conditions. Our findings invite future studies to examine whether memantine might correct maladaptive biases in emotional face processing that characterise mood and anxiety disorders, or affective symptoms/emotion processing in AD. Funded by MRC grant MR/J011754 awarded to MG, DSB and MRM.

C19 THE EFFECT OF FOUR WEEKS SSRI ADMINISTRATION ON FACIAL EMOTION RECOGNITION IN HIGH AND LOW NEUROTIC HEALTHY VOLUNTEERS Waltmann M, Univ Dept of Psychiatry, Univ of Oxford, Neurosciences Bldg, Warneford Hosp, Oxford OX3 7JX [email protected] Warren MB(1), Murphy SE(1), Harmer CJ(1), DiSimplico M(2) (1) As presenting author; (2) MRC Cognition and Brain Sciences Unit, 15 Chaucer Rd, Cambridge CB2 7EF Introduction: High neuroticism is associated with exaggerated emotional responses, excessive negative ABSTRACTS A77 thinking and higher risk for depression. High neurotic healthy individuals tend to avoid looking at the eye regions of emotional faces and 7-day SSRI treatment has been shown to increase gaze maintenance on faces in this population (DiSimplicio et al 2014, Neuropsychopharmacology, 39(13):3059–66). Furthermore, high neurotics display reduced amygdala response to faces (DiSimplicio et al 2013, Psychological Medicine, 44:1–12) and short term SSRI administration in this group increases amygdala responses to fearful faces (DiSimplicio et al 2014, Psychological Medicine 44(2):241–52). These results are in contrast with the reduced amygdala responses to fearful faces in healthy and depressed volunteers typically seen following antidepressants (Murphy et al 2009, British Journal of Psychiatry 194:535–40; Godlewska et al 2012 Psychological Medicine 42(12):2609–17). The current study investigated the effect of 4-week SSRI treatment in high and low neurotic healthy individuals. Methods: 68 healthy volunteers with high (≥16) and low (≤5) neuroticism (N) scores on the Eysenck Personality Questionnaire were randomised to receive 4 weeks of 20mg citalopram (High N=15, Low N=18) or placebo (High N=18, Low N=17). On day 28 of drug administration, participants completed the Facial Expression Recognition Task, in which they are asked to identify 6 basic emotions from ambiguous facial expressions. Results: There was a three-way interaction between face valence, neuroticism and drug group [F(1,64)=5.95, p=0.02]. In the high neurotic group there was a marginal main effect of citalopram (p=0.1) which reflected increased accuracy for both positive and negative faces following SSRI treatment. In the low neurotic group, there was a significant interaction between face valence and group [F(1,33)= 4.478, p=0.04], reflecting relatively increased accuracy to positive faces and relatively decreased accuracy to negative faces following SSRI treatment. Reaction times were not affected by antidepressant treatment. Conclusions: These results show that citalopram differentially affects high and low neurotic healthy volunteers. Whereas low neurotics exhibited the typical pattern of increased accuracy in recognising positive vs. negative faces following SSRI treatment, high neurotics showed overall increased accuracy in recognising facial expressions independent of their valence. This is consistent with previous work showing decreased avoidance and increased ocular exploration of faces following 7-day SSRI treatment in high neurotics. These results further highlight the effect that individual differences and personality can have on response to antidepressant treatment. This research was supported by a grant from the Univ of Oxford John Fell Fund.

C20 A SINGLE DOSE OF FLUOXETINE MODULATES EMOTIONAL PROCESSING IN HIGH TRAIT ANGER MALES Stavropoulou Deli A, Psychopharmacology and Emotion Research Lab, Dept of Psychiatry, Univ of Oxford, Warneford Hospital, Oxford OX3 7JX [email protected] Capitão LP(1), Murphy SE(1), Folloni D(1), Pike A(1), Hanson B(1), Taggart P(1), Cowen PJ(1), Harmer C(1) (1) As presenting author Introduction: Aggression and anger can be important correlates of irritable depression. Selective serotonin reuptake inhibitors (SSRIs) have been noted to act on emotional processing biases before clinical changes are seen (Harmer et al 2013 J Psychopharm; 5:435–43). From a forensic viewpoint, SSRIs (particularly fluoxetine) have been found to exert a moderate ameliorating effect in intermittent explosive disorder (Coccaro et al 2009 J Clin Psychiatry 70:653–62). Further investigation of the acute effects of fluoxetine could be relevant to the treatment of disorders where anger is a key symptom. Methods: Twenty-eight males with high trait anger were randomized to either a single 20mg dose of fluoxetine or placebo. Six hours post-administration, a facial expression recognition task and a dot-probe paradigm were administered. Subjective and autonomic psychophysiological correlates of anger were also investigated in an anger recall paradigm, which included constant elecrocardiographic monitoring. Results: In the facial expression recognition task, there was a trend for an interaction between group and emotion intensity regarding anger (p=0.094), with participants on fluoxetine being more accurate at detecting low and high intensity levels of this emotion [p=0.030, p=0.033 respectively]. In the dot-probe paradigm, participants on placebo showed a statistically significant absolute attentional bias towards angry faces (p=0.023), which was absent in the fluoxetine group (p=0.601). In the anger recall paradigm, there was a statistically significant interaction between treatment and visual analogue scales measuring anger-related symptoms (p=0.010). The fluoxetine group (N=8) reported overall lower scores in nervousness (p=0.014), irritability (p=0.044), fewer palpitations (p=0.019) and reduced feeling of being ‘out of control’ (p=0.058), compared to placebo (N=10). No group differences were seen when A78 ABSTRACTS considering the autonomic psychophysiological correlates of anger (all p’s>0.23). Conclusions: To the best of our knowledge, this is the first study that investigated the acute mechanisms of fluoxetine in high trait anger males. Overall, the findings regarding anger support the existence of an acute effect of fluoxetine on the processing of this emotion, consistent with our previous work (Capitão et al 2015 Psychol Med, 45: 2295 - 2308). There was some preliminary suggestion that fluoxetine reduced the attentional bias towards anger seen in the placebo group, but further investigation is needed in order to fully understand the direction of these effects and the dissociation between facial expression recognition vs. attention. The effects seen in the subjective report of anger may suggest that fluoxetine decreases subjective feelings associated with this emotion, but future replication is needed. Funding: We would like to thank the Medical Research Council for their generous support.

C21 A DOUBLE-BLIND RANDOMISED TRIAL TO INVESTIGATE THE ACUTE EFFECTS OF THE DOPAMINERGIC ANTIDEPRESSANT BUPROPION ON EMOTIONAL AND REWARD PROCESSING IN HEALTHY VOLUNTEERS Walsh AEL, Dept of Psychiatry, Univ of Oxford, Neurosciences Building, Warneford Hospital, Oxford OX3 7JX [email protected] Brown R(1), Browning M(1), Cowen P(1), Harmer C(1), Huneke N(2) (1) As presenting author; (2) Univ Dept of Psychiatry Academic Centre, Southampton, SO14 3DT Introduction: Previous research has shown that early in treatment, prior to an improvement in mood, antidepressants can remediate the negative biases in affective processing observed in depression. Whilst the majority of research has been conducted using antidepressants primarily influencing serotonergic and/ or noradrenergic activity, it remains unclear whether dopaminergic antidepressants, such as bupropion, exert similar actions on affective processing and whether these actions might have therapeutic implications. Therefore, here we investigate the acute effects of bupropion on emotional and reward processing in healthy volunteers. Method: 60 medication-free healthy volunteers (matched for age, gender, NART and baseline score on a variety of self-report questionnaires) participated in a double-blind randomised trial with three parallel groups of 20. Participants received either an acute dose (150 mg) of bupropion, placebo or no treatment prior to performing a probabilistic instrumental learning task and the emotional test battery (ETB). Results: Bupropion was found to increase positive emotional processing on the ETB. Specifically, compared to placebo, bupropion significantly increased the recognition of happy faces at lower happiness intensities (F = 2.21, p < 0.05; 30% t = -2.45, p < 0.05; 40% t = -2.73, p < 0.05) and significantly decreased the response bias for sad faces (F = 2.15, p < 0.05; t = -2.32, p < 0.05) indicating reduced misclassification of faces as sad. Bupropion also significantly increased the familiarity of positive self-referent words whilst decreasing the familiarity of negative self-referent words (F = 8.66, p = 0.001). There were no significant group effects on the probabilistic instrumental learning task and no significant effects of placebo vs no treatment on any of the tasks. Conclusions: Bupropion appears to increase positive emotional processing, which may be instrumental in its therapeutic effects as an antidepressant. Evidence is also provided that the ETB can be used for the specific detection of antidepressant effects on emotional processing across neurotransmitter systems without any influence of the placebo effect. Sources of financial sponsorship: Univ of Oxford & Medical Research Council.

C22 A SIMPLE PROBABILISTIC COGNITIVE SYNTAX TO FACILITATE REWARD-BASED RESPONSE BIAS IN PARTICIPANTS WITH INCREASED SYMPTOMS OF DEPRESSION Smee JA, Psychology, Reading Univ, Harry Pitt Bldg, Earley Gate Reading UK, RG6 7BE [email protected] Smee J(1), Schvartzman Benitez CM(1), Bryan S(1), Antonesei A(1), Murayama K(1), McCabe C(1), Mclaren B(1) (1) As presenting author Major depression disorder is characterized by deficits in the reward system (Pizzagalli et al., 2005, Biol Psychiatry, 57(4):319–27). One possible dysfunctional mechanism might be reduced reward sensitivity ABSTRACTS A79 instead of actual learning mechanism (Huys et al., 2013, Biol of Mood & Anxiety Disorders, 3(12):1–16). The aim of this study is to facilitate an increase in positive response bias in participants with high depression symptoms compared to healthy participants by using a simple probabilistic syntax in a reinforcement learning paradigm. So far, each of the 27 participants (N = 19 healthy controls, N=8 highly depressed participants according to a mood questionnaire) took part in a two signal-detection based probabilistic learning tasks conducted in two separate experimental sessions. For one task, unknown to the participants, the chocolate reward was delivered four times more for one stimulus (rich) compared to the other one (lean) across the three blocks. The second task was identical, apart from the first block, for which we have used an equal reinforcement ratio for both stimuli in order to create a neutral baseline block. Reward learning refers to an increase responsiveness to both, rich stimuli (hits) and false alarms (participants’ tendency to define an ambiguous stimulus as a target). We expect a steeper learning rate based on the difference between, on one hand, equal and neutral and, on the other hand, asymmetrical reward reinforcement probabilities in one of the task, compared to an equally increasing reward reinforcing rate from the other task. Results: Preliminary results showed a significant overall difference between healthy controls and participants with high depressive symptoms in reward learning in favour of the task with the different probabilistic syntax, p=.45. Further on, the results might be due to a significant difference in healthy controls in the last block between the two tasks, p=.03 (Mblock3task1 =-.13, SE = .68, Mblock3task2 =.26, SE = .52). Though not statistically significant, both healthy and high depressive symptoms participants seemed to benefit in terms of reward responsiveness from a more simple probabilistic cognitive syntax, with an average trend as expected. Conclusions: The lack of significant results within the highly depressed sample might be explained by the lack of enough statistical power. Increased sample size can help us to test the assumption that an easier-to-decode probabilistic reinforcement history facilitates an increase in reward-based response bias even in highly depressed participants. This project is funded by an MRC Ph.D. studentship.

C23 REWARD RESPONSIVENESS IN A PROBABILISTIC LEARNING TASK BY MEANS OF CORRECTIVE PUNISHMENT Antonesei A, Dept of Psychology, Univ of Reading, Earley Gate, Whiteknights,, Reading RG6 6AL [email protected] Murayama KM(1), McCabe CM(1) (1) As presenting author Major depression disorder is characterized by reduced responsiveness to hedonic stimulation. One possible dysfunctional mechanism might be blunted response to reward learning. Previous studies pointed towards a lack of response bias towards monetary gains in participants with elevated depression symptoms (Pizzagalli et al., 2005, Biol Psychiatry, 57(4):319–27) compared to healthy participants. On the other hand, people with depression are biased towards negative information (Ernst et al., 2005, Emotion, 5(4):446–55). The main aim of this study is to reverse the lack of response bias to rewards in participants with depressive symptoms compared to healthy controls by means of corrective punishment. 66 participants (N = 33 healthy controls, N = 24 mildly depressed and N=9 highly depressed according to a depression questionnaire) have taken part in a signal detection-based probabilistic learning task thus far. Unknown to the participants, the chocolate reward was delivered four times more for one stimulus (rich) compared to the other one (lean) across the three blocks; during the last two blocks, missing the rich stimulus was always punished via unpleasant taste. Reward learning refers to an increased responsiveness to both, rich stimuli (hits) and false alarms (participants’ tendency to define an ambiguous stimulus as the target), as opposed to avoidance of punishing cues (missing the hits). Results: Based on preliminary results, healthy participants showed an increase in reward learning from the first block to the second one, p<.05(Mdiff =-.25), and to the last block, p<.01 (Mdiff =-.32). However, there was no interaction effect between healthy and depressed participants, p=.54. Though not statistically significant, the mean performance on reward learning in the three groups was as expected: linear increase in healthy controls, (Mblock1 =.27, SDblock1 =.49, Mblock2 =.52, SDblock2 =.44, Mblock3 =.59, SDblock3 =.50), linear increase as well in severely depressed participants, but with a lower learning rate (Mblock1 =.19, SDblock1 =.37, A80 ABSTRACTS

Mblock2 =.31, SDblock2 =.59, Mblock3 =.37, SDblock3 =.61), and a non-linear trend in mildly depressed participants (Mblock1 =.33, SDblock1 =.23, Mblock2 =.54, SDblock2 =.63, Mblock3 =.37, SDblock3 =.59). Conclusions: The lack of significant results in mildly and severely depressed participants might be due to reduced sample sizes, as well as to individual variations on reinforcement learning. Further recruitment might prove the trends to be significant, therefore, corrective punishment might be one viable mechanism to induce reward responsiveness on short time intervals even in depressed participants. *This project is funded through an MRC Ph.D. studentship.

C24 DOES CHRONIC CITALOPRAM PRETREATMENT ENHANCE THE ABILITY OF ACUTE CITALOPRAM CHALLENGE TO MODULATE THE DEFAULT MODE NETWORK IN HEALTHY VOLUNTEERS? Dutta A, Neuroscience & Psychiatry Unit, Univ of Manchester, Stopford Building, Oxford Road Manchester M13 9PT [email protected] McKie S(1), Downey D(1), Anderson IM(1), Deakin JFW(1), Cross AJ(2), Davies KE(3), Williams SR(3), Juhasz G(4) (1) As presenting author; (2) AstraZeneca Neuroscience Innovative Medicines, Cambridge, MA, USA; (3) Imaging Science and Biomedical Engineering, Univ of Manchester, Stopford Building, Manchester, M13 9PT; (4) Neuroscience and Psychiatry Unit, Univ of Manchester, Stopford Building, Manchester, M13 9PT and MTA-SE-NAP B Genetic Brain Imaging Migraine Research Group, Hungarian Academy of Sciences, Semmelweis Univ, Hungary Introduction: The expression of the full antidepressant effects of SSRIs is thought to depend on 5-HT autoreceptor adaptation after repeated exposure leading to facilitated SSRI-induced 5-HT release. We sought functional correlates of this process in the human brain by comparing the CNS effects of acute intravenous citalopram infusion before, and after a subchronic course of oral citalopram or placebo. We predicted that the ability of acute citalopram to suppress the default mode resting state neural network (DMN), associated with rumination in depression, would be augmented after autoreceptor desensitisation induced by citalopram pre-treatment. Methods: 24 healthy volunteers were scanned on a Philips Intera 1.5T MR scanner whilst infused with 7.5mg intravenous citalopram. The phMRI sequence was 5 minute baseline continued for 20 minutes after the start of the infusion of citalopram for the first 7.5 minutes. Participants were then randomised to take either daily 20mg oral citalopram or placebo for 11 days followed by a 3 day washout to minimise ‘background’ citalopram concentration. A second citalopram phMRI scan was then carried out using the same protocol as the first scan. Images were preprocessed using SPM12. Independent Component Analysis was performed using the Group ICA for fMRI toolbox and the component showing the greatest spatial correlation to the DMN template was used. Each individual’s DMN component was then entered into a two way ANOVA to detect regions showing drug x visit interactions. Family wise error cluster level inference (pFWEc) was used at p<0.05 with a cluster forming height threshold of p=0.001. Results: Significant drug x visit interactions were noted in the cuneus, posterior cingulate cortex/precuneus, angular gyrus, medial frontal gyrus. Regional strength of the default mode during citalopram infusion was stable in those who received placebo between the first and second phMRI scans. In contrast, after subchronic oral citalopram treatment, iv citalopram was associated with a significantly reduced mean strength of the default mode in the cuneus and increased strength in the posterior cingulate/precuneus, left angular gyrus and medial frontal cortex. There were no associated changes in mood. Conclusion: Subchronic oral citalopram treatment changed the strength of the default mode during acute citalopram infusion, compatible with autoreceptor down regulation and supportive of 5-HT-modulation of the DMN. However, contrary to our hypothesis, an increased default mode strength in most hubs of the DMN was found. However, the effects in healthy volunteers might differ from those in depressed patients who are proposed to have overactivity of the DMN. Sponsorship: The project was funded by a strategic alliance between Imaging Science and Biomedical Engineering at the Univ of Manchester and AstraZeneca PLC. ABSTRACTS A81

C25 MAJOR DEPRESSION IS ASSOCIATED WITH UNSTABLE FUNCTIONAL CONNECTIVITY WITHIN THE DEFAULT MODE NETWORK Wise T, Centre for Affective Disorders, Dept of Psychological Medicine, Inst of Psychiatry, Psychology & Neuroscience, King’s College London, 103 Denmark Hill London SE5 8AF [email protected] Marwood L(1), Perkins AM(1), Herane-Vives A(1), Young AH(1), Cleare AJ(1), Arnone D(1), Luh W(2), Joules R(3), Lythgoe D(3), Williams SCR(3) (1) As presenting author; (2) Cornell MRI Facility, Cornell Univ, Ithaca, New York, United States; (3) Dept of Neuroimaging, Inst of Psychiatry, Psychology & Neuroscience, King’s College London, SE58AF Introduction: Major depression is associated with altered functional connectivity in a number of networks within the brain, most notably in the default mode network. However, to date no research has examined how connectivity between the medial prefrontal cortex and posterior cingulate cortex, key nodes in the default mode network, fluctuates over time in major depression. Methods: We calculated dynamic functional connectivity across the course of a single resting state fMRI scan between the medial prefrontal cortex and the posterior cingulate cortex. We examined the variability of connectivity strength between a group of 19 medication free patients with major depression and 19 healthy controls. We then assessed the reproducibility of this result in an independent sample of 19 medication free patients and 19 healthy controls. Results: We found that correlations between the medial prefrontal cortex and posterior cingulate cortex, key nodes in the default mode network, were significantly less stable in patients with major depression than those of matched healthy control subjects (p = .044). We were able to replicate these findings in an independent sample of individuals with depression (p = .045). Conclusions: Our results demonstrate that alterations within the default mode network go beyond changes in connectivity strength, and suggest that individuals with major depression are unable to maintain consistent connectivity within key functional circuits. We replicated this in an independent sample, indicating that this is a robust finding. These findings add a further dimension to theories that suggest altered functional connectivity may underlie symptoms of mood disorders, and may be associated with rumination in these conditions. Financial sponsorship: This work was supported by the Academy of Medical Sciences and the NIHR Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London.

C26 REDUCED CONNECTIVITY IN DISSOCIABLE FRONTO-STRIATAL CIRCUITS FOR IMPAIRED COGNITIVE FLEXIBILITY AND GOAL-DIRECTED PLANNING IN OBSESSIVE-COMPULSIVE DISORDER Vaghi MM, Dept of Psychology and Behavioural and Clinical Neuroscience Inst, Univ of Cambridge, Downing St., Cambridge CB2 3EB [email protected] van der Flier FE(1), Kundu P(2), Sule A(3), Vértes PE(4), Kitzbichler M(4), Apergis-Schoute AM(4), Zaman R(4), Voon V(4), Bullmore ET(4), Robbins TW(5), Fineberg NA(6) (1) Behavioural and Clinical Neurosciences Inst (BCNI), Univ of Cambridge, Cambridge; (2) Brain Imaging Center, Icahn Inst of Medicine at Mt. Sinai and Translational and Molecular Imaging Inst, Icahn Inst of Medicine at Mt. Sinai, NY, USA; (3) Cumbria Partnership NHS Foundation Trust, NHS, Penrith, Cumbria, UK; (4) Dept of Psychiatry, Univ of Cambridge, Cambridge; (5) Dept of Psychology, Univ of Cambridge, Cambridge; (6) Hertfordshire Partnership Univ NHS Foundation Trust and Univ of Hertfordshire, Hertfordshire Introduction: Recent hypotheses have suggested cognitive flexibility and goal-directed impoverished performance to represent core neuropsychological deficits in patients affected yb Obsessive Compulsive Disorder (OCD). This study was performed in order to test the direct contribution of impaired fronto- striatal connectivity to these deficits. Method: Multiecho resting state acquisition was used to obtain measures of functional connectivity in 43 OCD patients (27 medicated and 17 unmedicated) and 44 healthy comparison subjects. Cognitive flexibility and goal-directed performance were measured yb means of well-validated, standardized behavioural cognitive paradigms from the CANTAB battery; namely the Intra/Extra Dimensional Set Shifting and the Tower of London, respectively. We used two complementary analytical approaches. Firstly, we related cognitive flexibility and goal-directed performance to functional A82 ABSTRACTS connectivity strength from a priori identified striatal regions. Secondly, we applied unbiased brain network whole brain analysis to gain insight into fundamental network alterations. Results: OCD patients showed profound and marked impairment in cognitive flexibility (t=-2.649, df=84, p=0.01) and at the most difficult levels of goal-directed planning (t=-2.427, df=83, p=0.017) with no differences between medicated and unmedicated patients. Imaging data were consistent with a double dissociation such as whereas reduced functional connectivity between the caudate and the ventrolateral prefrontal cortex was selectively associated with reduced cognitive flexibility, goal-directed performance was selectively related to reduced functional connectivity between the putamen and the dorsolateral prefrontal cortex in OCD patients (p<0.01, FWE). Data-driven whole brain network analysis with Louvain modularity algorithm showed that in OCD patients nodes within the basal ganglia and cerebellum were most strongly intra- connected suggestive of an autonomous functional unit. Conclusions: These findings confirmed that OCD patients show marked impairment in cognitive flexibility and goal-directed performance, which previous studies demonstrated to have endophenotype qualities in OCD. The data extend neuropsychological models of OCD providing a direct link between reduced functional connectivity within fronto-striatal brain circuits and those cognitive processes underlying OCD symptoms possibly identified a valuable biomarker for this disorder. Supported by a Wellcome Trust Senior Investigator Award to TWR.

C27 THE RELATIONSHIP OF REPORTED CHILDHOOD ADVERSITY WITH ADULT HIPPOCAMPUS FUNCTION Grafton-Clarke D, Inst of Neuroscience, Newcastle Univ, Framlington Place, Newcastle upon Tyne NE2 4HH [email protected] Kotha RK(1), Thompson ST(1), Mullally SM(1), Watson SW(1) (1) As presenting author Adverse childhood experiences can increase the risk for depressive and anxiety disorders in adolescence and adulthood. Although the relationship between early adversity and depression is robust, the exact mechanism that underlies this is unclear. The aim of the present study was to investigate the possible mediating role of the hippocampus. Neuroimaging studies have provided evidence for hippocampal involvement, by demonstrating major depression is associated with atrophy of the hippocampus (Gradi & Pomi, 2008, Journal of Biological Physics, Vol.34(1), pp.107–120 ). As imagination and the ability to predict the future are hippocampal-dependent (Mullally & Maguire, 2013, The Neuroscientist, Vol.20(3), 220–234), it is possible this damage impairs the ability to visualise the future - an ability that has been shown to be protective against depression (Hamilton et al., 2014, Journal of Abnormal Child Psychology, 2015, Vol.43(3), pp.465–474). The current study investigated whether hippocampal function is associated with childhood adversity which was retrospectively assessed using the Emotional Neglect subscale of the Childhood Trauma Questionnaire (CTQ-EN). Four wellbeing questions were also included and respondents were asked for childhood postcode (to derive a deprivation index). The questionnaire was emailed to a convenience sample of 9985 students at Newcastle Univ and was completed by 1071 participants. Respondents were separated into high, medium and low EN and a subsample of 32 participants were randomly selected to perform additional tasks to assess hippocampal function. The tasks required participants to verbally describe memories, future events and imaginary scenarios. These descriptions were transcribed, segmented and classified according to whether they included spatial references, an entity presence, sensory descriptions and references to thoughts, emotions or actions. A composite score, the experiential index (EI), was generated for performance on each task. The subsample also completed Beck’s Depression Inventory and the State & Trait Anxiety Questionnaire. EN was negatively correlated with hippocampal function (r= -.84, p < 0.01) and associated with depression (r= .55, p = 0.003), trait anxiety (r= .43, p = 0.021) and low life satisfaction (r= -.56, p = 0.003). The relationship between EN and hippocampal function was not mediated by depression (p =.89) or anxiety (p =.85). The strong correlation between reported EN and a neuropsychological measure of hippocampal function in this sub-sample is of interest and supports the hypothesis that the hippocampus has a mediating role in the development of mood disorders. This research received no funding from any external agency in the public, commercial or not-for-profit sectors. ABSTRACTS A83

C28 WIDESPREAD INCREASED FUNCTIONAL VMPFC CONNECTIVITY IN OCD PATIENTS AND ITS RELATION TO COMPULSIVITY Bijleveld BA, Dept of Psychiatry, Univ of Cambridge, Herschel Smith Building, Robinson Way, Addenbrooke’s Hospital CB2 0SZ [email protected] Sahakian BJ(1), Apergis-Schoute AM(2), Robbins TW(3), Fineberg NA(4) (1) Dept of Psychiatry, Behavioural and Clinical Neuroscience Inst, Univ of Cambridge; (2) Dept of Psychiatry, Dept of Psychology, Behavioural and Clinical Neuroscience Inst, Univ of Cambridge; (3) Dept of Psychology, Behavioural and Clinical Neuroscience Inst, Univ of Cambridge; (4) South Essex Partnership Univ NHS Foundation Trust, Springhouse, Biggleswade Hospital, Bedfordshire UK. Dept of Psychiatry, QE II Hospital, Welwyn Garden City, Hertfordshire UK. Postgraduate Medical School, Univ of Hertfordshire, Hatfield UK Introduction: Recent neuroimaging studies have highlighted abnormalities in thalamo-cortico-striatal circuits in obsessive-compulsive disorder (OCD). The ventromedial prefrontal cortex (vmPFC) is believed to be a central hub in coordinating flexible behaviour and affective valuation. The present study builds on previously findings, showing that vmPFC hyperactivation in OCD patients predicts impaired use of safety signals in a pavlovian fear reversal paradigm. Altered connectivity between the vmPFC and areas associated with the salience, fronto-parietal and default mode networks, as well as the basal ganglia driving habitual behaviour, might play an important role in the aberrant behaviour seen in OCD. We used resting state analyses of a vmPFC seed that previously showed strong differences in safety signalling activation in the same group of OCD patients to gain insight in intrinsic alterations in vmPFC connectivity. Methods: We used a 10mm predefined seed in the vmPFC (-2, 26, -2) to compare functional resting state connectivity in 38 OCD patients to 33 matched healthy controls. Imaging data were analysed using SPM12. Each participant’s raw data was slice time corrected, spatially realigned to the first volume after which the mean image was spatially normalised to the Montreal neurological institute (MNI) brain template image and subsequently coregistered. We used contrasts OCD > Controls and Controls > OCD with a threshold of P < 0.001 uncorrected to look for whole brain group differences. Results: Whole brain connectivity from the vmPFC only showed significant increased connectivity in OCD. Patients with OCD exhibited significantly increased functional connectivity with the caudate (10, 12, 10), thalamus (-18, -20, -2), superior frontal gyrus (BA8) (12, 28, 42), fusiform (BA37) (48, -52, -8), inferior frontal gyrus (BA44) (-58, 6, 8), visual association cortex (BA19) (22, -68, -8) and cerebellum (20, -50, -32) (all p < 0.001, cluster threshold of 20 voxels). OCD patients also showed higher vmPFC auto-connectivity scaling linearly with ROI size. This autocorrelation predicted the severity of mental neutralisation (R2 = 0.2405, P = 0.0024) and washing (R2 = 0.1761, P = 0.0108). The vmPFC connectivity with the caudate, which was significantly higher in OCD patients (P = 0.00012), predicted Yale Brown Obsessive Compulsive Inventory scores (R2 = 0.1240, P = 0.0301). Conclusions: This study confirms that the vmPFC plays a central role in the altered thalamo-cortico-striatal loops found in OCD. The widespread vmPFC hyperconnectivity we found in OCD patients highlights dysfunction of the habit system, information processing, attention, self-awareness, mental imagery and inhibitory control. Keywords: OCD, vmPFC, resting state, fMRI, Seed based connectivity Acknowledgements: Wellcome Trust Senior Investigator Award to TW Robbins 104631/Z/14/Z and a joint award from the Medical Research Council and the Wellcome Trust supporting the Behavioural and Clinical Neuroscience Inst (G0001354).

C29 SUPPORT FOR THE ROLE OF THE GLUTAMATERGIC SYSTEM IN DEPRESSION: A 7T MRS STUDY Godlewska BR, Dept of Psychiatry, Univ of Oxford, Neurosciences Bldg, Warneford Lane, Oxford OX3 7JX [email protected] Lemke C(1), Emir U(1), Masaki C(2), Sharpley A(2), Cowen PJ(2) (1) Oxford Centre for Functional MRI of the Brain (FMRIB), John Radcliffe Hospital, Oxford OX3 9DU; (2) Univ Dept of Psychiatry, Univ of Oxford, Neurosciences Bldg, Oxford, OX3 7JX Introduction – The role of the glutamatergic system in the pathophysiology of depression has been gaining increasing attention. Over the past two decades, magnetic resonance spectroscopy (MRS) has A84 ABSTRACTS been used to study the pathophysiology of depression. Unfortunately, results have not been consistent. One limitation of previous studies was low field strengths unable to differentiate between glutamate and glutamine in a reliable manner. Therefore, to further probe the glutamatergic interactions associated with depression, we have acquired spectra at 7 T in two regions potentially relevant for depression (VOIs, volumes of interest), anterior cingulate cortex (ACC) and the putamen (Put), and correlated it with a number of clinical factors. Methods –Forty-one depressed patients (DSM-V criteria) and 28 age and gender matched healthy controls with no history or current psychiatric conditions were included. Severity of depression was measured by Hamilton Depression Scale (HAMD) and Beck Depression Inventory (BDI). Additional questionnaires measured anxiety (Spielberger’s Trait Anxiety Inventory), fatigue (Chalder Fatigue Questionnaire) and anhedonia (Snaith-Hamilton Anhedonia Scale). All participants were scanned on a 7T whole body MR system (Siemens, Erlangen) with signals acquired from two separate VOIs – in the ACC and in the Put. Metabolites were quantified with LCModel using the unsuppressed water signal as reference. Metabolite differences between groups and VOIs were assessed using a repeated measures ANOVA and post-hoc t-tests, and correlations were assessed by Pearson’s statistics. Results – Patients with depression had higher glutamine in the Put (p=0.011, 4.95 vs 4.22). The severity of depression as measured by both BDI and HAMD was correlated with the glutamate (Glu) and Glx (a composite measure of glutamate and glutamine) in the ACC (BDI: Glu -0.461, p=0.004, Glx -0.399, p=0.014, HAMD: Glu -0.455, p=0.005, Glx -0.337, p=0.041). Anxiety scores correlated with the levels of Glu in ACC (-0.294, p=0.040). There was no correlation between any of the measures and anhedonia or fatigue scores. Conclusion – This study supports the role of the glutamatergic system in the pathophysiology of depression. With the use of a 7T scanner it provides a more reliable measure of the metabolites, which may help in clarifying the inconsistencies between previous studies. Given the variability of the changes between VOIs studies, our findings suggest that the effect of depression on glutamate might have some regional specificity and may also relate to depression severity. Additional studies may be required in order to tease out additional insight into glutamatergic effects associated with depression. The study was funded by Medical Research Council (MRC).

C30 BLUNTED NEURAL RESPONSES TO ANTICIPATION EFFORT AND CONSUMMATION OF REWARD AND AVERSION IN MAJOR DEPRESSIVE DISORDER Rzepa E, School of Psychology and Clinical Language Sciences, Univ of Reading, Harry Pitt Building, Earley Gate, Reading RG6 7BE [email protected] CiaraMcCabe CM(1) (1) As presenting author Introduction: Previous studies have shown reduced responses to reward in patients with depression (Rizvi et al., 2016, Neuroscience and Biobehavioral Reviews, in press). Further we have shown that young people at increased risk of depression also have aberrant neural responses to reward and aversion (McCabe et al., 2012, Biological Psychiatry, 72, 588–97). However, how the different aspects of reward processing (anticipatory, effort, consummatory) are represented in the brains of young people with depression is as yet unknown. Methods: 25 currently depressed individuals (DA) and 25 age and gender matched healthy controls (HC) will be recruited. Our fMRI task examines the anticipation (reward and aversion cues) effort, (pressing buttons to win reward or avoid aversion) and consummation (the receipt of chocolate taste of unpleasant taste). Volunteers also rated wanting, pleasantness and intensity of the stimuli in the scanner. Results: Despite no differences in subjective pleasantness, wanting or intensity for the stimuli we found that the DA group took significantly longer (p=0.015) to complete the effort phase than the HC group and that this was only under the reward conditions. Our fMRI data revealed that the DA group compared to the HC group had reduced activity in the putamen, orbitofrontal cortex, precuneus, amygdala (all p<.001) and ventral striatum (p=.009 svc) for anticipation and consummation of reward and reduced activity in the posterior cingulate, precuneus, putamen and insula (p<.001) for anticipation and consummation of aversion. During the effort part we also found that the DA group compared to the HC group had reduced activity during the reward trials in the supplementary motor cortex, putamen and hippocampus and we found no group differences ABSTRACTS A85 during effort to avoid aversion. Conclusion: Our findings are the first to show that young people with depression have blunted anticipatory and consummatory neural responses to reward and aversion yet blunted neural response to effort only during reward trials. Further we find that young people with depression exert less effort when working for reward compared to controls. These results may indicate a mechanism by which individuals with depression experience anhedonia (reduced effort for reward) and fail to disengage from negative experiences which in turn can help maintain depression. Funded by the Medical Research Council PhD studentship.

C31 COGNITIVE BIASES IN EMOTIONAL MENTAL IMAGERY AND MOOD INSTABILITY IN BIPOLAR DISORDER Di Simplicio M, Cognition and Brain Sciences Unit, MRC, 15, Chaucer Road, Cambridge CB2 7EF [email protected] Renner F(1), Blackwell S(1), Mitchell H(1), Lau-Zhu A(1), Watson P(1), Holmes EA(2), Stratford H(3) (1) As presenting author; (2) MRC Cognition and Brain Sciences Unit, Cambridge, UK and Dept of Clinical Neurosciences, Karolinska Instt, Sweden; (3) Oxford Health NHS Foundation Trust, Warneford Hospital, Oxford OX3 7JX Introduction: Patients with bipolar disorder (BD) describe experiencing intrusive ‘flashforwards’ of negative future events, or playing movies of worst case scenarios in their mind that amplify anxiety and low mood. We have shown that targeting distressing emotion-laden mental imagery via a brief psychological intervention is able to reduce mood instability in BD (Holmes et al., 2016, Transl Psychiatry. Jan 26;6:e720). Based on this initial clinical evidence, we aimed to investigate the cognitive basis of imagery phenomenology in individuals with BD compared to non-clinical controls; and to determine the specificity of cognitive abnormalities in mental imagery processing in BD relative to depression and anxiety disorders. Methods: Participants (N=131) were: 27 BD depressed and 27 BD euthymic (BD combined group), 27 non-clinical controls, 26 individuals with unipolar depression and 25 individuals with anxiety disorders, matched on age, gender, ethnicity, levels of education and premorbid IQ. Experimental tasks assessed cognitive (objective, non-emotional) stages of mental imagery, including visual short-term memory and mental rotation. Questionnaires and experimental tasks assessed subjective domains of mental imagery including spontaneous imagery use, interpretation bias and emotional mental imagery. Results: Compared to non-clinical controls, (i) on experimental tasks assessing cognitive stages of mental imagery, the BD combined group showed no clear abnormalities; (ii) within the subjective domains of mental imagery, the BD combined group reported a greater impact of intrusive prospective imagery in daily life (p<0.001), more vivid and ‘real’ negative images on a prospective imagery task (p=0.013), and higher self-involvement on a picture-word task (p=0.04). No significant differences remained comparing BD depressed individuals to either the depressed or anxious clinical control group. Across all groups, imagery characteristics were associated with affective lability (affective lability predicted the impact of intrusive prospective imagery, p=0.001) and anxiety (anxiety predicted the impact of intrusive prospective imagery, p<0.001; and the vividness of negative future imagery, p=0.001). Conclusions: Compared to non-clinical controls, BD is characterised by cognitive biases in tasks of future emotional mental imagery only. These abnormalities are not specific to BD and may reflect transdiagnostic aspects of psychopathology associated with mood instability and anxiety comorbidity. Funding sources: EAH, HM: Wellcome Trust Clinical Fellowship [WT088217]; EAH, SEB and FR: MRC intramural programme [MC-A060-5PR50 to EAH]; MDS: MRC Career Development Fellowship; HJS: Oxford Inst of Training; ALZ: Cambridge Commonwealth, European & International Trust. A86 ABSTRACTS

C32 PSILOCYBIN-ASSISTED THERAPY INCREASES AMYGDALA RESPONSIVITY TO FEARFUL FACES, WHILE REDUCING DEPRESSIVE SYMPTOMS AFTER THERAPY IN TREATMENT-RESISTANT DEPRESSION Roseman L, Dept of Medicine, Imperial College London, Burlington Danes Bldg, Hammersmith Campus, 160 Du Cane Road, London W12 0NN [email protected] Nutt DJ(1), Carhart-Harris RL(1), Demetriou L(2), Wall MB(2) (1) As presenting author; (2) Imanova, Centre for Imaging Sciences, W12 0NN London, UK Introduction: In recent years psychedelic assisted therapy is experiencing a revival. This was the first study since the 60’s to treat depression with psilocybin. Depressed subjects are known to have a hyperactive right amygdala to fearful faces that is attenuated with SSRI’s (Ma, 2015, Molecular psychiatry, 20.3:311–319). We hypothesised that psilocybin-assisted therapy would change right amygdala response to fearful faces after the therapy. Methods: 20 subjects, diagnosed with moderate to severe, unipolar, treatment resistance depression, underwent psilocybin-assisted therapy in two sessions, 7 days apart, with psychological support before, during and after these session. The first session was with 10mg and the second with 25mg psilocybin PO. This was an open-label study. Subjects underwent fMRI scans before the first session and one day after second session. Neutral, fearful and happy were presented during the fMRI session and changes were measured in the right amygdala. Results: One subject did not complete the study due to adverse emotional reaction during the session. 12 out 19 subjects showed a decrease of more then 50% in BDI scores a week after treatment (out of them 11 were considered in remission (BDI<9)). Right amygdala showed increased response after treatment for fearful and happy faces but not to neutral faces compared to no-stimuli (cluster corrected, Z > 2.3, p < 0.05). Right amygdala showed increased response after treatment for fearful compared to neutral faces (cluster corrected, Z > 2.3, p < 0.05). Discussion: While treatment-response was robust, the effect on amygdala responsivity was in the opposite direction to effects of SSRI in the literature. Long-term antidepressant treatment leads to blunting of amygdala responses in the same emotional recognition task (Ma, 2015, Molecular psychiatry, 20.3:311–319), where our psilocybin assisted therapy enhanced amygdala responsivity. The difference in directionality of the amygdala changes might be related to the difference in the approach of both treatments: while the former attempts to ease negative emotions, the later attempts to confront and work through them. As the patients in this study showed rapid remission of symptoms, this suggests that blunting of emotional responsivity may not be necessary for therapeutic efficacy. Psychedelic assisted therapy may provide a treatment avenue that preserves, and potentially even enhances emotional receptiveness. Financial Support: Medical Research Council, UK.

C33 PSILOCYBIN REDUCES PESSIMISM BIAS IN PATIENTS WITH UNIPOLAR TREATMENT-RESISTANT DEPRESSION Lyons T, Centre for Neuropsychopharmacology, Division of Brain Sciences, Imperial College London, Burlington Danes Bldg, Hammersmith Hospital, Du Cane Rd, London W12 0NN [email protected] Carhart-Harris RL(1), Bolstridge M(1), Pilling S(2), Curran HV(2), Nutt DJ(1) (1) As presenting author; (2) Division of Psychology & Lang Sciences, Faculty of Brain Sciences, UCL, London WC1E 6BT The burden of depression is severe and increasing with many patients failing to respond to first line antidepressant treatments. There are reports of persistent improvements in wellbeing and optimism after a single exposure to psilocybin, a naturally occurring compound that is structurally similar to serotonin. This pilot study investigated the ability of psilocybin to reduce pessimism bias in patients with unipolar treatment-resistant depression. Methods: We recruited patients (N = 15) that met DSM-IV criteria for major depression of a moderate to severe degree (17+ on the 21-item Hamilton Depression Rating Scale) and have failed to clinically respond to at least two different classes of antidepressants within the current episode. The Prediction of Future Life Events (POFLE; Strunk et al. Behav Res Ther 2006; 44: 861-82) assessment was used to measure bias, which lists 40 life events balanced in terms of desirability that subjects must predict ABSTRACTS A87 the probability of occurring over the succeeding 30 days. The POFLE was split into two versions, A and B. All patients received both versions in counterbalanced order. Patients reported which events occurred 30 days after both versions. All patients received psilocybin in two oral doses: a low dose (10 mg), to give an impression of the drug effects at an attenuated intensity, and a subsequent high dose (25 mg) 7 days later. The POFLE assessments were carried out at screening and during the final follow-up visit. Results: There was no significant difference between the actual occurrence of positive outcomes (t(14)=1.38, p=0.19) or desirable (t(14)=1.22, p=0.24) and undesirable (t(14)=0.00, p=1.0) events between conditions. Therefore, predictions reflect true biases as opposed to differences in life circumstances. The mean of bias across all patients before psilocybin treatment was significantly lower than zero and indicated a tendency towards pessimism (t(14)=3.26, p=0.006). Following psilocybin treatment, the mean of bias was not significantly different from zero and indicated no cognitive biases (t(14)=0.768, p=0.455). Overall, patients showed a pessimism bias at screening that was significantly reduced following psilocybin treatment (t(14)=-2.71, p=0.017). When these data were broken down further, it was shown that psilocybin increased probability estimates given specifically for desirable events (t(14)=-2.125, p=0.05) and also reduced pessimism bias when predicting desirable (t(14)=-3.238, p=0.006), but not undesirable (t(14)=-1.238, p=0.22), events. Conclusions: These findings show that psilocybin improves depressed patients’ outlook for the future yb reducing pessimism when predicting the likelihood of good things happening to them. Psilocybin should be researched further as a potential intervention for treatment-resistant depression Funded by the MRC.

C34 PSILOCYBIN FOR TREATMENT RESISTANT DEPRESSION: SAFETY, EFFICACY AND MECHANISMS Carhart-Harris RL, Medicine, Imperial College London, Burlington Danes Building, Du Cane Rd W12 0NN [email protected] Bolstridge M(1), Rucker J(1), Day C(1), Roseman L(1), Kaelen M(1), Erritzoe D(1), Bloomfield M(1), Rickard J(1), Forbes B(1), Feilding A(1), Taylor D(1), Pilling S(1), Curran HV(1), Nutt DJ(1) (1) As presenting author Introduction: Psilocybin is a naturally occurring plant alkaloid found in the psilocybe species of mushrooms that have been ingested by some cultures for hundreds if not thousands of years in holistic “healing” ceremonies. Psilocybin is also the pro-drug of psilocin, a classic psychedelic drug and serotonin 2A receptor agonist. Psilocybin was used as an adjunct to psychotherapy in the 1950s and 60s and modern clinical trials have assessed its potential in the treatment of anxiety, obsessive compulsive disorder and addiction, with promising preliminary results. Here we assessed its safety and efficacy in patients with treatment resistant depression in an open-label feasibility study. Methods: Nineteen patients with treatment resistant depression (defined as failure of at least two different medications within the current episode) received two oral doses of psilocybin 7 days apart. A low dose (10mg) was given in the first dosing session and a high dose (25mg) in the second. Patients were psychologically prepared for their experiences, supported throughout and properly debriefed afterwards by a core team of mental health professionals. Patients underwent resting state fMRI at pre-treatment baseline and one day after their high-dose session. Main depression outcomes were the Quick Inventory of Depressive Symptoms (QIDS) and the Beck Depression Inventory (BDI). Results: The psilocybin was physiologically well tolerated by all of the patients, although some subjective anxiety was experienced by most at some point during their drug sessions, especially on the higher dose of psilocybin, and headaches lasting 1–2 days were reported by 50% of patients after the treatment sessions. Depressive symptoms were significantly reduced at 1-week post- treatment (mean (SD) QIDS and BDI decreases = 10.2 (5.3) and 22.3 (11.4), p < 0.01) and remained so for 3-months post-treatment (mean QIDS and BDI difference = 7.9 (7) and 16.9 (12.3), p < 0.01). Eleven (58%) patients met criteria for remission 1-week post-treatment (BDI score of <10) and 38% remained depression- free at the 3-month follow-up period. Altered brain network properties were observed post-treatment and these changes correlated with decreases in depressive symptoms (p < 0.05). Conclusion: The results of this feasibility study support the case for further research into the therapeutic efficacy of psilocybin as a treatment for major depression and suggest potential brain mechanisms underlying the treatment’s putative efficacy. Randomised control trials are now warranted to more thoroughly assess this novel treatment approach. Funding: This research was supported by the Medical Research Council, UK. A88 ABSTRACTS

C35 LURASIDONE IN BIPOLAR DISORDER: EARLY IMPROVEMENT AND CONDITIONAL PROBABILITY OF RESPONSE Warnock A, Sunovion Pharmaceuticals Europe Ltd, Southside, 105 Victoria St, London SW1E 6QT [email protected] Tsai J(1), Pikalov A(1), Loebel A(1) (1) 84 Waterford Drive, Marlborough, MA, USA, 01752 Introduction: The aim of this post-hoc analysis of a bipolar depression sample was to evaluate the conditional probability of responding to treatment with lurasidone, either as monotherapy, or adjunctive therapy, given lack of early improvement. Methods: Patients with bipolar I depression were randomised, in 2 double-blind trials, to 6 weeks of monotherapy with lurasidone (18.5–55.5 mg/d, N=161); or 74–111 mg/d, N=162) vs. placebo; or with lurasidone (18.5–111 mg/d) vs. placebo adjunctive with lithium or valproate. Lack of improvement was defined as a reduction of <25% in the Montgomery-Åsberg Depression Rating Scale (MADRS) total score. A completer analysis was conducted on the lurasidone treatment groups; in the adjunctive trial, separate analyses were performed for the lithium (N=90) and valproate (N=89) treatment groups. Week 6 response was defined as ≥50% improvement in the MADRS. Results: Week 6 response rates for lurasidone monotherapy at daily doses of 18.5–55.5 mg and 74–111 mg were 59.2% and 60.2%, respectively. For monotherapy treatment with lurasidone 18.5–55.5 mg/d and 74–111 mg/d, the proportion of patients with lack of improvement at weeks 1, 2, and 3 were 82.6% and 77.6%, 54.8% and 51.0%, and 39.7% and 34.7%, respectively. For monotherapy treatment with lurasidone 18.5–55.5 mg/d and 74–111 mg/d, the proportion of patients with lack of improvement at weeks 1, 2, and 3 who achieved a week 6 response were 54.6% and 52.6%, 43.5% and 36.7%, and 32.0% and 17.9%, respectively. Week 6 response rates on lurasidone adjunctive with lithium or valproate were 68.5% and 59.7%, respectively. For treatment with lurasidone adjunctive with lithium, the proportion of patients with lack of improvement at weeks 1, 2, and 3 was 71.1%, 43.7%, and 31.3%, respectively; and for lurasidone adjunctive with valproate, the proportion with lack of improvement at weeks 1, 2, and 3 was 76.4%, 57.1%, and 30.5%, respectively. For the adjunctive lithium group, lack of improvement on lurasidone at weeks 1, 2, and 3 was associated with week 6 response rates of 41.1%, 16.4%, and 5.5%, respectively. For the adjunctive valproate group, lack of improvement on lurasidone at weeks 1, 2, and 3 was associated with week 6 response rates of 38.9%, 23.6%, and 1.4%, respectively. Conclusions: The results of this post-hoc analysis found that the probability of achieving endpoint response was low in patients who did not show improvement after 3 weeks of treatment with lurasidone, whether as monotherapy, or adjunctive to lithium or valproate. Sponsored by Sunovion Pharmaceuticals Inc.

C36 LURASIDONE FOR THE TREATMENT OF BIPOLAR DEPRESSION: SYSTEMATIC REVIEW AND META- ANALYSIS Taylor MJ, Dept. Psychosis Studies, Inst of Psychiatry, Psychology & Neuroscience, London SE5 8AF [email protected] Introduction: Lurasidone is a new oral antipsychotic medication that has been proposed as a potential treatment for bipolar depression. Increasingly treatments are compared both on efficacy and acceptability using data from randomised controlled trials. Recent network meta-analyses of bipolar depression included two lurasidone studies. Here we aimed to identify all completed randomised controlled trials of lurasidone in the treatment of bipolar depression and synthesise their results using meta-analysis. Methods: Randomised controlled trials were identified by searching of electronic databases and from discussion with experts in the field. Data were extracted, and meta-analyses were performed using the meta package in R. Results: Three completed trials were identified in which over 1200 people with bipolar depression had been randomised between lurasidone and placebo. One further additional ongoing study was also identified. People randomised to receive lurasidone had a greater improvement in depression rating scale scores (SMD -0.26; 95% CI -0.39 to -0.14), and had a ABSTRACTS A89 higher response rate (OR 1.84; 95% CI 1.29 to 2.63). Acceptability was similar to placebo, with similar overall discontinuation rates (OR 0.95; 95% CI 0.72 to 1.26). Conclusions: Lurasidone appears to be a promising potential treatment for episodes of bipolar depression. Its overall efficacy and acceptability seem similar on current evidence to established treatments for this indication. Future studies could investigate longer-term acceptability and impact on relapse rates. Financial Sponsorship: None.

C37 A PRELIMINARY STUDY OF PLASMA NOCICEPTIN/ORPHANIN FQ IN PATIENTS WITH AFFECTIVE DISORDER Wang L, Dept. of Psychiatry, Univ. of Southampton, Academic Centre, College Keep 4–12 Terminus Terrace, Southampton SO14 3DT [email protected] Baldwin D(1), Hou R(1), Liu L(2), Qiao D(2) (1) As presenting author; (2) Shandong Mental Health Centre, Jinan shandong, China 250014 Introduction: The pathogenesis of affective disorder is not yet fully understood. One critical step for improving clinical outcome is to identify biomarkers for diagnosis and targeting of interventions. This study was aimed to investigate changes of plasma nociceptin/orphanin FQ(N/OFQ) level in patients with affective disorder. Methods: This is a cross-sectional case-control study. 45 patients with bipolar disorder current depressive episode or major depressive disorder (BD group), 26 patients with bipolar disorder current manic episode (BM group) and 31 healthy volunteers (control group) were recruited. BD patients were assessed by Hamilton Depression Scale (HAMD) and Montgomery and Asberg Depression Rating Scale (MADRS). BM patients were assessed by Bech-Rafaelsen Mania Rating Scale (BRMS). 2ml venous blood samples were collected at 7 am and plasma were kept at -70° after centrifuge. Plasma N/OFQ concentrations were measured by a commercial immunoassay kit. All statistical analyses were performed using SPSS 21 software. Independent samples t-test, analysis of variance and Spearman rank correlation analysis were conducted as appropriate. Results: The plasma N/OFQ concentrations were (18.58±4.12) ng/L in BD group, (11.29±3.28) ng/L in BM group and (13.92±4.53) ng/L in control group. Compared with the control group, the levels of plasma OFQ in BD group was higher (t=4.65, P<0.01), while the level in BM group was lower (t=-2.47, P<0.05). These differences were statistically significant. Correlation analysis between plasma N/OFQ levels and self- report questionnaires revealed that levels of plasma N/OFQ in BD group were positively correlated with total scores of HAMD (r=0.574, P<0.01) and MADRS (r=0.643, P<0.01); whereas there was a significantly negative correlation between N/OFQ levels and total scores of BRMS in BM group (r=-0.750, P<0.01). Conclusions: Findings from the current study demonstrate significant differences of plasma N/OFQ level among patients with bipolar disorder current depressive episode/major depressive disorder, bipolar disorder current manic episode, and matched heathy controls. In addition, plasma N/OFQ levels were significantly correlated with self-report measures of mood. Once confirmed and validated in further studies, plasma N/OFQ can potentially be used as biological markers for improving diagnosis of affective disorder. This study was sponsored by Shandong Mental Health Centre, the teaching hospital of Shandong Univ in China.

C38 CORTISOL LEVELS IN UNIPOLAR AND BIPOLAR DEPRESSION USING HAIR AND SALIVA SPECIMENS Herane Vives A, Psychological Medicine, Inst of Psychiatry, Psychology and Neurosciences, King´s College London, 103 Denmark Hill, London SE5 8AF [email protected] Papadopoulos A(1), Chin-kuo CH(1), Wise T(1), Young AH(1), Cleare A(1), Deangel V(2) (1) King´s college London; (2) Universidad de Chile Introduction: Early discrimination between unipolar and bipolar depression is a significant diagnostic challenge. One of the reasons is that at onset bipolar disorder (BD) often presents with symptoms undistinguishable from unipolar major depression (UD). Several line of research currently focus on understanding the biological underpinning of these common affective disorders to aid phenomenological syndromic recognition. The identification of a diversified and specific pattern of cortisol secretion might help develop biomarkers for these disorders. To date there is little agreement on their cortisol status. The A90 ABSTRACTS common use of specimens that reflect acute rather than chronic cortisol levels may partly explain this problem similarly to the potential diagnostic contamination of unipolar samples with bipolar cases. In this study, we investigated acute and chronic cortisol levels in patients with UD and BD. Methods: Fifty-nine participants with DSM IV UD, 12 with BD, age and sex matched with 40 healthy controls were recruited. Bipolarity was assessed with the HCL-33 and cortisol levels measured with saliva and hair specimens. Results: Hair cortisol concentration (HCC) was greater in BD vs. UD based on DSM diagnostic criteria. By using more sensitive HCL-33 bipolarity index criteria, HCC was also greater in BD vs. healthy controls. Measurement of the area under the curve with respect to the ground (AUCg) suggested lower cortisol in BD vs. healthy controls (p<0.05). Greater bipolarity index was associated with lower levels of AUCg cortisol. Conclusion: Our results support a different pattern of cortisol secretion in affective disorders which could aid diagnostic discrimination. Hypercortisolemia may be a chronic phenomenon only in BD although it is possible that this pattern may differ in relation to bipolarity subtype . The use of the bipolarity index as adjuvant screening tool in clinical diagnosis might help to correctly identify bipolar disorder when assessing depressive presentations. Dr Herane Vives was supported from the Bicentennial Fund for Human Capital Development (Becas Chile) and by the Psychiatric Research Trust.

C39 CORTISOL LEVELS IN MAJOR DEPRESSIVE EPISODE USING FINGERNAIL SPECIMENS Herane Vives A, Psychological Medicine, Inst of Psychiatry, Psychology and Neurosciences, King´s College London, 103 Denmark Hill, London SE5 8AF [email protected] Fischer S(1), Papadopoulos A(1), Wise T(1), Young AH (1), Cleare A(1), Deangel V(2) (1) King´s college London; (2) Universidad de Chile Background: hypercortisolemia may be a biomarker in mayor depressive episode (MDE), but not all studies have obtained this finding. Previous conflicting results may be partly explained by the use of different specimens that only assess acute cortisol levels. Recently, fingernails have been proposed to be a specimen for measuring chronic cortisol levels over a period of several weeks. But, cortisol levels using fingernails have not been measured in MDE yet. Methods: different clinical features and subtypes of depression in addition to cortisol levels that were measured using fingernails in a group of 26 subjects with DSM-5 MDE were compared with an age and gender matched group of 45 controls. Results: MDE subjects showed significantly higher cortisol levels measured in fingernails compared to control subjects. Mean=201.24 pg/ mg, (s.d=277.26 pg/mg) v/s 101.52 pg/mg, (s.d=90.54), respectively p=0.03. MDE subjects with non-atypical features exhibited higher cortisol levels than controls (p=0.03) Conclusion: the results showed elevated cortisol in MDD using an aggregate measure of cortisol over several weeks. MDE subjects has a subclinical chronic hypercortisolemia condition. Cortisol may have therefore a role in the frequent association between some common medical comorbidities such as hypertension and diabetes in subjects with MDE and subclinical hypercortisolemic condition. Hypercortisolemia may be a biomarker for MDE subjects with non-atypical features. Dr Herane Vives was supported from the Bicentennial Fund for Human Capital Development (Becas Chile) and by the Psychiatric Research Trust.

C40 A PRELIMINARY INVESTIGATION INTO THE EFFECTS OF A MINDFULNESS-BASED SMARTPHONE INTERVENTION ON ANXIETY AND SLEEP IN HEALTHY VOLUNTEERS Olney L, Dept of Psychiatry, Univ of Southampton, Academic Centre, College Keep, 4–12 Terminus Terrace, Southampton SO14 3DT [email protected] Garner M(1), Hou R(1) (1) As presenting author Introduction: Mindfulness is defined as a moment-to-moment awareness, cultivated by purposefully paying attention to the present experience with a non-judgmental attitude. Mindfulness interventions have been found to reduce anxiety as well as change the activity in key areas of the brain associated with learning, emotion and cognitive processing. However, the effect of delivering mindfulness interventions via a ABSTRACTS A91 smartphone application on anxiety in the general healthy population remains unclear. Aims: This study aims to investigate the effects of a mindfulness smartphone application on anxiety and sleep in healthy volunteers. Methods: A randomized controlled trial study design was employed: 37 healthy applicants were recruited and randomly allocated into an intervention group or an active control group. The members of the intervention group were given access to a mindfulness smartphone application called headspace and were instructed to complete a 10-minute guided meditation daily for a total of 20-days. The members of the active control group were given instructions to browse social media/news media on their smartphone for 10-minutes a day for a total of 20-days. Standard self-reported questionnaires were used to measure anxiety, sleep, mindfulness. The attention network task was used to measure cognitive processing. Measures were assessed at baseline and after 20 days of the intervention. Results: Independent samples t-test analyses on the pre and post change in the mean of outcome measures indicate that the intervention group had significantly reduced anxiety (p=0.046) and increased mindfulness (p=0.05) in comparison to the control group following the intervention. However, the study did not reveal any significant effects on either cognitive control (p=0.108) or sleep (p=0.872). Conclusions: The findings from this study are comparable to other studies that have used non-smartphone based mindfulness interventions, suggesting that the use of a smartphone-based mindfulness application could be a cost-effective alternative option and should be further investigated to offer a larger range of treatment options when considering the use of mindfulness in clinical population. Acknowledgements: The study was funded by the Univ of Southampton. The subscriptions of the mindfulness smartphone application were provided by Headspace.

C41 DIFFERENCES IN FUNCTIONING OF MINERALOCORTICOID RECEPTOR BETWEEN PATIENTS WITH UNIPOLAR AND BIPOLAR DEPRESSION Juruena MF, Psychological Medicine, Centre for Affective Disorders, IoPPN- King’s College London, PO72, De Crespigny Park, Denmark Hill, London SE5 8AF [email protected] Baes CVW(1), Burgos_de_Macedo B(1), Martins-Monteverde CMS(1), Bosaipo NB(1), Tofoli SMC(1), Pichiteli M(1), Umeoka E(1) (1) Dept. of Neurosciences and Behavior. School of Medicine of Ribeirao Preto. Univ of Sao Paulo, Brazil. Introduction: Favorable outcomes in the treatment of depression disorder deeply depend on differential diagnosis between Unipolar Major Depressive Disorder (MDD) and Bipolar Depression (BD), which must be done as soon as possible. However, methods to identify differences between these two diagnoses are still under development in neuroscience clinical studies. Biomarkers, such as the mineralocorticoid receptor (MR) function, are considered a promising strategy for such differentiation, seeing that malfunctioning of the HPA axis is one of the most common findings in psychiatry patients. Therefore, the aim of this study was to evaluate the different role of MR in the regulation of cortisol release in depressive patients diagnosed with MD or BD. Methods: Patients were diagnosed with BD (n=18) or MDD (n=8) by mean of the International Neuropsychiatric Interview (MINI-Plus), patients presenting scores ≥16 in Hamilton Depression Rating Scale (HAM-D21) were included in this study. Placebo, Fludrocortisone, and Spironolactone were administered at 22:00 and plasma cortisol assessment was performed at 09:00 next day. Each drug was administered with an interval of 72 h. Results: Two-Way ANOVA showed significant main effect for diagnostic (BD vs MD ; p=0.041; F(1, 24) = 4.622), treatment (Pla vs Flu vs Spi; P<0.001; F(2, 48)=17.61) as well as interaction between factors (p=0.004; F(2, 48)=5.942). Post hoc analyses indicated that BD patients show significant (p<0.05) lower levels of plasma cortisol than MDD patients after Placebo administration. After treatment with Fludrocortisone this is also significant (p<0.05) lower plasma cortisol levels in BD compared to MDD patients. There was no significant difference between BD and MDD patients after Spironolactone administration. Interestingly, plasma cortisol levels of BD patients after Fludrocortisone treatment were not different from Placebo-treated BD, however after Spironolactone treatment it was significantly higher (p<0.01) in BD. In the other hand, plasma cortisol concentration of MDD patients after Fludrocortisone treatment was significantly (p<0.05) lower from Placebo-treated MDD and after Spironolactone treatment it was not different (p>0.05). Conclusion: Our data indicate lower levels of plasma cortisol in BD patients. Furthermore, when compared to MDD patients, BD show higher cortisol suppression after MR agonist treatment and higher cortisol release after MR antagonist treatment, which A92 ABSTRACTS indicate an increase in MR activity in patients with BD. Thus, our data suggest that the MR dysregulation plays an important role into Bipolar Depressive physiopathology. Acknowledgments: The study was supported by Academy of Medical Sciences/ Royal Society-UK; CNPq, CAPES, FAEPA and FAPESP- Brazil.

C42 RELATIONSHIP BETWEEN BURNOUT SYNDROME, STRESS AND LIFESTYLE AMONG HEALTH CARE PROVIDERS IN A BRAZILIAN EMERGENCY DEPT Juruena MF, Psychological Medicine, Centre for Affective Disorders, IoPPN, PO72, De Crespigny Park Denmark Hill, London SE5 8AF [email protected] Pereira SS(1), Morero JAP(1), Santos MTS(1), Cardoso L(1), Pereira Júnior GA(2), Hegadoren K(3) (1) Ribeirao Preto College of Nursing, Univ of Sao Paulo, Brazil; (2) Ribeiro Preto Medical School, Univ of Sao Paulo, Brazil; (3) Univ of Alberta, Nursing Faculty, Edmonton, Alberta/ Canada Introduction: Burnout Syndrome is characterized as involving emotional exhaustion, depersonalization and low personal accomplishment. This occupational hazard is common and associated with increased absenteeism, disability and illness. The goal of this study is to assess levels of emotional exhaustion, depersonalization and personal accomplishment among health care providers in a Brazilian emergency Dept (ED) their relationships to stress symptoms and demographic and lifestyle variables. Method: This was a cross-sectional descriptive design involving health care providers in an urban public tertiary care hospital ED in the state of Sao Paulo, Brazil. A stratified random sample of 114 care providers was chosen from total population of these categories of workers. Data collection involved a demographic questionnaire, Maslach Burnout Inventory-Human Services Survey (MBI-HSS), Inventory of Stress Symptom (ISS) and “Lifestyle Fantastic” questionnaire. Descriptive analysis and bivariate correlations (with a significance level ≤ 0.05) were completed in SPSS. All ethical aspects were respected. Results: Demographic data included mean age (37± 7.9), gender (83.3% women), education (46.5% had high school education), marital status (49.1% married) with children (64.9%). Most of the participants were nursing technicians (56.1%), with experience from 1 to 35 years (mean = 9.7 ± 8.4), working 6.8 hour days and 32 hours per week. Burnout was identified in 13.2% of the participants. The scores on the MBI-HSS showed that 30.7% of participants had high levels of Emotional Exhaustion, 27.2% high levels of Depersonalization and 60.5% had lower Personal Accomplishment. Significant stress symptoms were identified in 43% of the sample and only 9.6% had a regular lifestyle. A positive correlation was found between Emotional Exhaustion and stress (p <0.000) and negatively correlated with life style variables (p <0.000). The correlation between Depersonalization and stress symptoms was also positive (p = 0.008). Lifestyle was negatively correlated with Depersonalization (p = 0.005), but positively correlated with lowered Personal Accomplishment (p <0.001). Conclusions: The prevalence of the Burnout Syndrome among health care providers in the ED was high, suggesting that this work environment can be chronically stressful for care providers. Almost 1/3 of the care providers reported high levels of Emotional Exhaustion and Depersonalization and 2/3 reported low personal accomplishment. Emotional Exhaustion and Depersonalization were directly related to stress. These data suggest that many of the participants in this study would struggle with providing optimal patient centered care, highlighting the need to develop and test for specific stress management strategies for individuals and for care providers in the ED as a group. Financial source: Sao Paulo Research Foundation – FAPESP.

C43 EMOTIONAL AND PHYSICAL ABUSE AS IMPACT FACTORS FOR SUICIDAL IDEATION IN BIPOLAR PATIENTS Juruena MF, Psychological Medicine, Centre for Affective Disorders, IoPPN, King’s College London, PO72, De Crespigny Park, Denmark Hill, London SE5 8AF [email protected] Menezes IC(1), Burgos_de_Macedo B(1), Baes CVW(1) (1) Dept. of Neurosci and Behavior. School of Medicine of Ribeirao Preto. Univ of Sao Paulo, Brazil. Introduction: Patients with bipolar disorder present higher suicide ideation than patients with other psychiatric disorders (MANN JJ, 2002, Ann Intern Med.; 136:302–311). Exposure to early life stress ABSTRACTS A93

(ELS) may lead to a greater susceptibility of developing depression and other comorbities (CARR CP et al., 2013, J Nerv Ment Dis.; 201(12):1007–1020). ELS present some subtypes that must be considered – emotional, physical and sexual abuses, and emotional and physical neglect. Aim: To assess if ELS and its subtypes may influence the level of suicidal ideation in bipolar patients, when compared to unipolar depressive patients. Methods: The sample was composed by two groups - unipolar depression (n=37 - without ELS= 19; with ELS= 18) and bipolar depression (n=59 – without ELS=26; with ELS=33). For diagnostic assessment, MINI International Neuropsychiatric Interview was used. To assess the intensity of depressive symptoms, GRID-Hamilton Depression Rating Scale was applied; to assess ELS, Childhood Trauma Questionnaire was applied; to assess suicidal ideation Beck Suicide Inventory was applied. Results: There was no difference in depressive symptoms between bipolar and unipolar groups, with or without ELS (p=0.538). Subjects with ELS presented higher levels of suicidal ideation (p=0.001), but there was no difference when considering unipolar or bipolar groups (p=0.236). Although, when considering subtypes of ELS, bipolar patients with ELS presented higher suicidal ideation than unipolar patients, when experienced emotional (p=0.038) and physical (p=0.002) abuses during childhood. Conclusion: Emotional and physical abuses are subtypes of ELS that increase the suicidal ideation in subjects with bipolar depression. Financial sources: Academy of Medical Sciences/Royal Society, UK; FAPESP, CNPq and CAPES, Brazil.

C44 UNLICENSED APPLICATIONS OF LICENSED PSYCHOTROPIC DRUGS WITHIN A TERTIARY CARE SPECIALIST SERVICE FOR AFFECTIVE DISORDERS: RETROSPECTIVE CASE-NOTE SURVEY Tiwari N, Psychiatry, Univ of Southampton, Antelope House, Brintons Terrace, Southampton SO14 0YG [email protected] Baldwin D(1) (1) Univ Dept of Psychiatry Academic Centre, College Keep 4–12 Terminus Terrace, Southampton SO14 3DT Introduction: Patients are often prescribed psychotropic medications outside the terms of the product licence in an attempt to improve clinical outcomes. It could be assumed that such ‘off- label’ prescriptions would be most common in patients with more severe and treatment resistant conditions, but the factors associated with unlicensed applications remain unclear. We wished to evaluate the extent of unlicensed prescriptions for patients within a single tertiary care Mood and Anxiety Disorder Service, and hypothesised there would be no association of unlicensed prescriptions with proxy markers of the burden of illness. Methods: A retrospective examination of characteristics of all patients referred to a regional specialist tertiary referral service for affective disorders over a five-year period. We determined the number of unlicensed and licenced applications in three broad diagnostic groups (unipolar depressive disorders, bipolar disorder, anxiety disorders), and in the sub- group of patients with comorbid depressive and anxiety disorders. The group of patients in which at least one treatment recommendation was for an unlicensed application was compared to the patient group in which all treatments were within the terms of the product licence. Proxy markers of overall illness burden included a history of long-term mental health problems, of receiving electroconvulsive therapy, non-fatal self-harm, of psychosis, and of psychiatric inpatient admission. Results: The study sample comprised 144 patients. In patients with unipolar depressive disorders, similar proportions were recommended to receive an unlicensed treatment, or to receive only licenced treatments (33 and 32 patients, respectively). Similar proportions were seen in groups with bipolar disorder (24 and 26 patients, respectively) and anxiety disorder (15 and 14 patients, respectively). There were no significant differences between ‘licensed’ and ‘unlicensed’ groups in any proxy marker of overall illness burden. The p- values comparing the 2 groups along the markers of illness burden for unipolar depressive disorder namely history of ECT, history of inpatient admission,history of psychosis,long term illness and history of self harm ( in this order for all groups) were ,847,.105,.048,.321,.480 . For Bipolar disorder: .381,.571,.549,.666 and .666 and for anxiety disorders : .517,.741,.NA as very small figure,..272 and .517 respectively. Conclusion: Limitations include retrospective examination of notes recorded for other purposes within a single service, treatment recommendations which might reflect idiosyncratic personal A94 ABSTRACTS preference, and use of proxy markers which do provide a fully comprehensive summary of overall illness burden. But this retrospective case-note study accords with the findings of previous evaluations and indicates that ‘off-label prescribing’ is common in patients with treatment-resistant affective disorders. In this group, we were unable to identify any factors which were significantly associated with treatment recommendations involving an unlicensed application, which suggest that other influences must be important. No funding was sought or provided for this study.

C45 FOR THE TREATMENT OF BIPOLAR DISORDER: META-ANALYSIS Taylor MJ, Dept. Psychosis Studies, Inst of Psychiatry, Psychology & Neuroscience, London SE5 8AF [email protected] Palacios J(1), Young A(1), Yildiz A(2) (1) Dept Psychological Medicine, IOPPN, London; (2) Dokuz Eylül Univ Introduction: Tamoxifen is an oral medication that has been proposed as a potential treatment for bipolar disorder. Tamoxifen acts to inhibit the intracellular action of Protein Kinase C, which is also an action of well-established treatments such as lithium and valproate. Here we aimed to identify randomised controlled trials of tamoxifen in the treatment of bipolar disorder and synthesise their results using meta- analysis. Methods: Randomised controlled trials were identified yb searching of electronic databases and from discussion with experts in the field. Data were extracted, and meta-analyses were performed using the meta package in R. Results: Five placebo-controlled randomised controlled trials of tamoxifen in the treatment of acute mania were identified. There were no trials in the treatment of episodes of bipolar depression, or for relapse prevention. The studies of mania treatment were of between three and six weeks duration. Tamoxifen was studied either as monotherapy (two trials) or as augmentation of lithium or valproate (three trials). Change in mania scale scores favoured tamoxifen over placebo: SMD -2.14 (95% CI -3.39 to -.89; 4 trials). Response rates were also higher: RR 4.35 (1.99 to 9.50; 4 trials). Acceptability was similar to placebo: RR 1.03 (0.94 to 1.13; 5 trials). Conclusions: Tamoxifen appears to be a promising potential treatment for episodes of mania. Future studies could investigate its effects as an adjunct to antipsychotic medications for improved anti-manic efficacy, and establish its longer term effects on mood and relapse. Financial Sponsorship: None.

C46 THE NEURAL EFFECTS OF PSYCHOTHERAPY IN DEPRESSION AND ANXIETY DISORDERS: A META- ANALYSIS Marwood L, Centre for Affective Disorders, Dept of Psychological Medicine, IoPPN, KCL, 103 Denmark Hill, P074, London SE58AF [email protected] Wise T(1), Perkins A(1), Cleare A(1) (1) Centre for Affective Disorders, IoPPN, KCL, 103 Denmark Hill, PO74, London, SE58AF Introduction: A better understanding of the neural mechanisms of psychotherapy in depression and anxiety disorders could aid better understanding of the recovery process and help target more effective treatments. Typically, pre-post treatment changes in brain activation are found in areas involved in emotional processing and self-regulation (i.e. limbic, striatal and prefrontal brain regions), but no clear pattern has yet emerged in this rapidly growing field of research. We report a meta-analysis of changes in brain activation after receiving evidence-based psychotherapy for depression and anxiety. Methods: A comprehensive literature search was conducted up to May 2015 to identify functional neuroimaging studies that scanned patients with anxiety or depression before and after evidence-based psychotherapy. Obsessive compulsive disorder and specific phobia were excluded due to purported differences with these disorders. A meta-analysis of the neural changes accompanying psychotherapy was conducted on suitable studies using seed-based d mapping. To ensure only the most robust results were reported and to limit the influence of outliers, a jackknife sensitivity analysis was conducted to assess the contribution of individual studies to the overall results. Publication bias was assessed by visual inspection of funnel plots and Egger ABSTRACTS A95 regression tests. Results: 14 studies (including 218 patients) met our inclusion criteria: 5 panic disorder; 4 post-traumatic stress disorder; 3 major depression; and 2 social anxiety disorder studies. A significant decrease in clusters with peak coordinates in the left anterior cingulate / paracingulate gyrus (Brodmann area; BA 10, Z-score = -2.46, p<.001, 1500 voxels); left (BA 45, Z = -2.41, p<.001, 900 voxels) and right (BA 44, Z= -1.63, p<.001, 127 voxels) inferior frontal gyrus; and left insula (BA 10, Z= -1.62, p<.001, 254 voxels) were found after therapy, compared with before. No significant increases in activation were found post- compared to pre-therapy. Conclusions: The results are consistent with neural models of improved self- and emotional-regulation following psychotherapy in depression and anxiety as evidenced by decreased activity within prefrontal, anterior cingulate and insular cortices. This meta-analysis suggests that there are consistent, trans-diagnostic, brain changes with psychotherapy in depression and anxiety disorders. Sources of financial sponsorship: LM was supported by a joint MRC-IoPPN studentship.

C47 MOOD INSTABILITY CAPTURED BY DAILY REMOTE MONITORING IN PARTICIPANTS VULNERABLE TO BIPOLAR DISORDER Panchal P, Dept of Psychiatry, Univ of Oxford, Oxford Centre for Human Brain Activity, Warneford Hospital, Oxford OX3 7JX [email protected] Nelissen N(1), Nobre AC(1), Scholl J(2), Rushworth MFS(2), Darby D(3), Saunders KEA(4), Harrison PJ(4), Harmer CJ(4) (1) As presenting author; (2) Dept of Experimental Psychology, Univ of Oxford, Oxford, OX1 3UD; (3) Florey Inst of Neuroscience and Mental Health, The Univ of Melbourne, Victoria, Australia; (4) Univ Dept of Psychiatry, Warneford Hospital, Oxford, OX3 7JX Introduction: Mood instability is a common feature of Bipolar Disorder (BD). It is a risk factor for the development of BD and is associated with worse functional outcomes. The emergence of new technologies and the ubiquity of internet access enable prospective monitoring of daily mood and concurrent cognitive states. This allows the interaction between mood and cognition to be studied longitudinally, and enables us to better understand how this contributes to mood instability. In this study we explore whether remote monitoring can capture mood instability longitudinally, and whether instability can be detected at shorter temporal scales in response to cognitive test outcomes. Methods: Remote mood and cognitive monitoring over 10 weeks, as part of the ongoing Cognition and Mood Evolution across Time study. We analysed data from 13 participants scoring >7 on the Mood Disorder Questionnaire (MDQ), thus showing increased vulnerability to BD, and 7 matched controls (MDQ<5). Mood was assessed using the PANAS scale. Participants completed an iPad gambling task, where they could gain or lose money. Before and after gambling, participants scored their happiness on a 21-point scale (from unhappy (-10), neutral (0), to happy (10)). The Root Mean Square of Successive Differences (RMSSD) was used to analyse variability in PANAS scores across the 10-week period. Further correlational analyses (Spearman) were employed to test the relationships between PANAS and happiness scale ratings, and the change in happiness ratings as a consequence of gambling outcome. Results: The high MDQ group showed greater variability in PANAS scores (Mdn=4.47) compared to the controls (Mdn=3.23) (p<0.05), but no difference in average PANAS scores (high MDQ Mdn=4.27, control Mdn=5.77, p=0.16). Daily PANAS scores were strongly associated with average daily happiness ratings (rs=0.56, p<0.00). Change in happiness ratings correlated with gambling outcome, such that happiness decreased after losing, and increased after winning, for both high MDQ (rs=-0.50, p<0.01) and control (rs=-0.44, p<0.01) participants. Conclusions: Daily longitudinal remote mood monitoring captures mood instability in participants vulnerable to BD. On shorter time scales we can detect mood fluctuations linked to the outcome of a gambling task. Findings add to the increasing evidence for mood instability as a marker for BD, and suggest the efficacy of remote technology in assessing instability at short time scales. This will aid in the search for biomarkers that can be used in the development of novel therapeutic targets and in determining the effectiveness of current treatments. Wellcome Trust funded. A96 ABSTRACTS

C48 VALPROATE FOR BIPOLAR DISORDER: THE QUALITY OF PRESCRIBING PRACTICE IN THE UK Paton C, Prescribing Observatory for Mental Health, Royal College of Psychiatrists, 21 Prescot St., London E1 8BB [email protected] Bhatti SF(1), Barnes TRE(2), Cookson J(3), Ferrier IN(4) (1) As presenting author; (2) Centre for Mental Health, Imperial College London W12 0NN; (3) East London Foundation NHS Trust; (4) Inst of Neuroscience, Newcastle Univ The National Institute for health and Care Excellence (NICE) and the British Association for Psychopharmacology (BAP) have published evidence-based guidelines for the management of bipolar disorder. Valproate is a known human teratogen: when taken during pregnancy, the absolute risk of giving birth to a baby with a neural tube defect such as spina bifida is as high as 2%, a substantial increase over population norms. The incidence/prevalence of neurodevelopmental disorders such as autism in children exposed to valproate in-utero is also increased several fold over population norms and the average IQ of these children is considerably lower. In 2015, the MHRA reiterated its recommendations to healthcare professionals that women of childbearing age should be given appropriate information about these risks when valproate treatment is being considered. POMH-UK initiated a quality improvement programme focussing on the prescription of valproate for people with bipolar disorder. Fifty-five mental health Trusts/ healthcare organisations participated in the baseline audit in 2016, submitting data on prescribing practice for bipolar disorder in relation to evidence-based practice standards. Data were submitted on 6,705 patients with a diagnosis of bipolar disorder under the care of adult services, of whom over a third (36%) were prescribed valproate. Valproate was more commonly used than lithium, which was prescribed for a quarter (25%) of patients in the sample. More than three-quarters (79%) of those prescribed valproate were also prescribed an antipsychotic, most commonly quetiapine or olanzapine. Of the 2,364 women of child-bearing potential (50 years of age or younger), a quarter (24%) were prescribed valproate. Of the 74 in whom valproate had been started in the past six months, there was no documentation of any discussion about the potential benefits and side effects of the newly-initiated valproate in just over a quarter (27%). There was no documented discussion about the need for contraception in almost half (45%) and for a similar proportion (50%) there was no evidence that they had been informed about the potential teratogenic effects of this medication. Benchmarked data on Trust and individual clinical team performance against the practice standards relating to prescribing will be fed back to the participating Trusts/healthcare organisations to prompt local review and reflection on practice followed by the generation of action plans to address areas where practice falls short of the standards, which is clearly the case in relation to the use of valproate in women of child-bearing age.

C49 A STUDY OF MENTAL HEALTH STATUS OF PARENTS OF CHILDREN AND ADOLESCENTS WITH MENTAL DISORDERS Cheng X, Dept of Psychiatry, Univ of Southampton, Academic Centre, College Keep 4–12 Terminus Terrace, Southampton SO14 3DT [email protected] Baldwin D(1), Hou R(1), Liu J(2), Yang Y(2) (1) As presenting author; (2) Shandong Mental Health Centre, Jinan Shandong, China 250014 Introduction: The current study was to investigate mental health status and related factors of parents of children and adolescents who were admitted to psychiatric hospital due to mental disorders. Methods: Three study groups were recruited, including Group 1: 162 parents of hospitalized children and adolescents with mental disorders; Group 2: 158 parents of hospitalized children and adolescents with physical illnesses; and Group 3: 166 parents of healthy children and adolescents. There were no significant differences between groups in terms of age, gender, profession, and family income (P>0.05). All participants completed a series of questionnaires including a basic family information questionnaire, Symptom Checklist 90 (SCL-90), Self-Rating Depression Scale (SDS), Self-Rating Anxiety Scale (SAS), Social Support Rating Scale (SSRS). Data were analysed using SPSS21. Results: Total scores of SAS, SDS and the ABSTRACTS A97 total scores and the positive symptom distress level of SCL-90 in groups 1 and 2 were significantly higher than the control group (P<0.05). SCL-90 factor scores of somatization, depression, and anxiety between groups were also significantly different (P<0.05). Moreover, SCL-90 factor scores of anxiety, hostility, photic anxiety and obsession of group 1 were significantly higher than the other two groups (P<0.05). Correlation analyses showed that mental health status of parents was positively correlated to their educational background, their perception of mental disorders, social support environment, and attitude of health care professionals (P<0.05). Conclusions: Mental health status of parents of children and adolescents with mental disorders was significantly poorer than the healthy control and physical illness groups, and in particular, they had higher levels of anxiety, depression, and hostility. The study revealed that factors affecting their mental health status were their educational background, their perception of the mental disorders, social support environment, and attitude of health care professionals. Interventions targeting these risk factors may help to improve the mental health status of these parents. This study was sponsored by Shandong Mental Health Centre, the teaching hospital of Shandong Univ in China.

C50 SCREENING FOR PHYSICAL HEALTH PROBLEMS IN MOOD DISORDER SERVICE OUTPATIENTS: A REPLICATION Kinch AJ, Univ of Southampton Medical School, Dept of Psychiatry Academic Centre, College Keep 4–12 Terminus Terrace, Southampton SO14 3DT [email protected] Holt RIG(1), Baldwin DS(2) (1) The Inst of Developmental Sciences (IDS Bldg), MP887, Univ of Southampton, Southampton General Hospital, Tremona Rd, Southampton SO16 6YD ; (2) Univ Dept of Psychiatry Academic Centre, College Keep 4–12 Terminus Terrace, Southampton SO14 3DT Background: Severe mental illness is associated with an increased risk of cardiovascular disease which is partly explained by an increased prevalence of known risk factors such as smoking and diabetes. A previous survey conducted within our service (Lack D et al. Ther. Adv. Psychopharmacol. 2015; 5: 22–25) demonstrated that physical health monitoring prior to referral was sub-optimal, with many missed opportunities for screening. We repeated the investigation, to ascertain the proportion of patients with affective disorders who had been monitored for metabolic and other abnormalities in accordance with current guidelines, and to see whether screening had been optimised. Method: A retrospective case- note study of electronic and paper records of consecutive newly referred patients attending a regional specialist tertiary referral service for complex, severe and treatment-resistant affective disorders, between April 2013 and September 2015. We searched for documented evidence of glucose testing, body weight measurement, recording of blood pressure, and estimation of lipid profile and serum prolactin in each patient. The number of contacts with healthcare services was calculated to provide an estimate of the opportunities for monitoring. Results: The notes of 94 newly referred patients (28 men, 66 women: mean aged 48.3 years, range 18–74 years) were examined: retrospective data for the previous 12 months were available for 60 patients (20 men, 40 women: mean aged 48.3 years, range 18–74 years). In this sub-group, the mean number of contacts where monitoring had been possible in the 12 months before assessment was 11.15 (median 6, range 1–65): documented evidence of physical health monitoring was seen in 34 (56.7%) patients. Monitoring was more likely in patients who were currently undergoing antipsychotic treatment or who recently underwent psychiatric inpatient care. When compared to the previous study, there was a substantial increase in the proportion of patients with evidence of physical health monitoring (56.7% vs. 29.2%, p <.01). Conclusion: As in the previous study, screening for risk factors in this vulnerable population is documented variably. Physical health monitoring is more likely in patients who are receiving antipsychotic treatment or who have recently undergone psychiatric inpatient care, and may have improved, but in a substantial minority of patients monitoring remains sub-optimal. Many had multiple contacts with healthcare professionals where monitoring would have been possible, but many opportunities still appear to be being missed. No funding was sought for this project. A98 ABSTRACTS

C51 IS CHRONIC FATIGUE SYNDROME AN INFLAMMATORY DISORDER? A SYSTEMATIC REVIEW AND META-ANALYSIS Sartor ML, Centre for Affective Disorders, Dept of Psychological Medicine & Dept of Medicine, IoPPN, King’s College London & Univ of Padova, De Crespigny Park, Denmark Hill, London & via Giustiniani, 2, Padova, SE5 8AF & 35128 [email protected] Starwbridge R(1), Young AH(1), Cleare A(1) (1) Centre for Affective Disorders, Dept of Psychological Medicine, Inst of Psychiatry, Psychology & Neuroscience, King’s College London, De Crespigny Park, Denmark Hill, London, SE5 8AF & The NIHR Biomedical Research Centre, South London and Maudsley NHS Foundation Trust and the Inst of Psychiatry, Psychology and Neuroscience, King’s College London Aims: Chronic fatigue syndrome (CFS) is a poorly-understood disorder causing prolonged, disabling physical and mental fatigue. CFS shares high comorbidity with mood disorders, and some symptomatic resemblances to “sickness behavior” as seen in inflammatory states. Supporting evidence of immunological dysfunction in some CFS cohorts leads to the possibility that a meaningful biomarker could be identified to ultimately improve diagnosis and intervention. eW aimed to examine the existing literature and undertake the first meta-analyses comparing inflammatory cytokines between people with and without CFS. Methods: A systematic review of the literature comparing levels of inflammatory biomarkers in CFS patients and controls was conducted. Studies were included which measured circulating cytokines in vivo, in adults with a current CFS diagnosis and controls without CFS. From the included studies, meta-analyses were conducted for biomarkers that had been measured in at least four studies. Results: The systematic review yielded 48 studies; within these, 13 inflammatory markers were examined in meta-analyses. We found the following biomarkers to be significantly higher in those with CFS than controls: tumor necrosis factor-alpha (TNFa), p=0.03; transforming growth factor-beta (TGFb), p=0.01; c-reactive protein (CRP), p=0.008; and interleukins 2 (IL-2), p=0.01 and 4 (IL-4), p=0.02. IL-1b and IL-12 showed a trend to be higher in this group (p=0.05 and p=0.08, respectively). The remaining biomarkers (IFNy, IL-1a, IFNa, IL-8, IL-10 and IL-6) were not significantly different between the two populations. Conclusions: These results support an inflammatory component in CFS. The relationship between depression, chronic fatigue and inflammation requires further investigation, as does the role of inflammation in the prognosis of CFS. We were not able to determine the underlying cause of inflammatory changes in CFS, nor whether they are a primary or secondary feature. Nevertheless, this research may lead to an improved understanding of inflammatory functions in the etiology and/or maintenance of CFS, and the development of possible illness markers or treatment targets. This abstract presents independent research partly funded by the National Institute for Health Research (NIHR) and the Biomedical Research Centre for Mental Health at South London & Maudsley NHS Trust and the Inst of Psychiatry, Psychology & Neuroscience (King’s College London).

C52 THE PROFILE OF P2X7 RECEPTOR EXPRESSION AND PROPERTIES IN POLARIZED PRIMARY MOUSE MICROGLIA CULTURES Jackson MG, Pharmacy and Pharmacology, Univ of Bath, Claverton Down Rd, Bath, North East Somerset BA2 7AY [email protected] Wickens RA(1), MacKenzie AB(1), Bailey SJ(1) (1) As presenting author Adenosine triphosphate (ATP)-gated P2X7 receptors (P2X7R) play an important role in neuro-inflammation and have been implicated in mood disorders (Sperlágh and Illes, 2014, Trends Pharmacol Sci, 35, 537–547). Microglia mediate brain inflammation where they possess the ability to polarize into M1/M2 phenotypes based on cytokine and toll-like receptor activation. Although a functional role for P2X7R has been demonstrated in lipopolysaccharide (LPS)-treated primary rat microglial cells (Bianco et al., 2006, Journal of Neurochemistry, 99, 745–758), little is known about their expression or function in activated mouse primary microglial cells. This study aims to investigate P2X7R expression and function in M1 polarized, ABSTRACTS A99 pro-inflammatory, mouse microglia. Primary microglia were obtained from cortices of mixed-gender C57BI/6 mouse pups (P0-P2) using the low trypsinization method (Saura et al., 2003, Glia, 44, 183–189). Cells were plated at 300,000 cells/well; after 21 DIV, astrocytes were removed by incubating with diluted trypsin (0.0625%) and experiments began 24h later. BV-2 cells, (immortalized murine microglia; Blasi et al., 1990, Journal of Neuroimmunology, 27 229–237) and primary microglial cells were treated with either 100 ng/ ml LPS (24h) or 100ng/ml LPS plus interferon (IFN)γ (20 ng/ml) (24h) to obtain M1 polarization, confirmed by expression of pro-IL-1β detected by quantitative fluorescent western blot (LICOR) (n=3). P2X7R protein expression was also evaluated by quantitative western blot and compared to β-actin expression. P2X7R activation was measured by uptake of fluorescent ethidium in BV-2 microglia. M1 polarization with LPS+IFNγ down-regulated P2X7R expression in both primary microglia (63.2±29.1% n=3) and BV-2 microglia (87.5±32.8 % n=5) compared with non-polarized controls (M0). LPS alone down-regulated P2X7R expression in primary microglia (62.0± 33.7% n=3) but not BV-2 cells (110.5±36.4% n=5). Extracellular ATP-mediated (>1mM) ethidium uptake in BV-2 M0 microglia (n=3) was inhibited by the selective P2X7R antagonist A740003 (30μM) confirming functional expression of P2X7R. LPS priming +/- IFNγ (24h) did not alter ATP-mediated (1–5 mM) ethidium uptake in BV-2 microglia compared to control conditions (n=3). Understanding changes in microglia leading to pro-inflammatory activated states is important in developing novel treatments that target neuro-inflammation. While there is a trend for down-regulation of P2X7R in activated, polarized M1 mouse primary microglia, there is little evidence of LPS-induced change in P2X7R expression or function in BV-2 cells. This contrasts to findings in rat primary microglia. We are investigating functional activation of P2X7R in primary mouse microglia. RAW is funded by an MRC Industrial Case PhD Studentship.

D01 INCREASED RELIANCE ON PAVLOVIAN HEURISTICS IN PATHOLOGICAL ANXIETY Robinson OJ, UCL ICN, 17 Queen Square, London WC1N 3AR [email protected] Aylward J(1), Mkrtchian A(1), Roiser JP(1), Dayan P(2) (1) As presenting author; (2) Gatsby Computational Neuroscience Unit, 25 Howland Street, London, W1T 4JG Introduction: Pathological feelings of anxiety are the most common psychiatric symptoms, affecting close to 15% of adults in the UK. Unfortunately our understanding of the neurocognitive mechanisms of anxiety is limited, contributing to poor treatment outcomes. Numerous studies have linked anxiety to negative affective biases, and at the same time parallel work has linked anxiety to inhibitory control. These process are, however, entangled: behavioural inhibition is a prepotent Pavlovian reaction to aversive outcomes (e.g. freezing in response to threats). One plausible hypothesis, therefore, is that anxiety increases reliance on Pavlovian heuristics as a computationally efficient means of avoiding potential harm. In this study we tested this hypothesis by exploring the effect of anxiety (both induced and pathological) on a valenced inhibition task designed to probe instrumental control over Pavlovian biases. Methods: A sample of healthy controls (N=59) and unmedicated individuals meeting criteria for mood and anxiety disorders (N=43) completed a task (adapted from Guitart-Masip et al. 2012,NeuroImage,62:154–166) requiring go responses for reward and inhibitory nogo responses to avoid punishment (Pavlovian heuristics) as well as crossed trials in which instrumental control was required to overcome these biases (nogo for rewards or go to avoid punishment). The task was completed at baseline and under threat of unpredictable shock - a back translated anxiety-induction paradigm. Expectation maximization was used to fit a range of modified Rescorla Wagner learning models to trial-by-trial choice behaviour (blind to treatment or condition). Critically, some models included a free Pavlovian parameter which increased the probability of inhibitory responses in the face of punishments. Results: Model selection revealed that the winning model included this Pavlovian parameter. Critically, pathological (p=0.012), but not induced (p=0.2), anxiety was associated with increased reliance on this parameter. Induced anxiety, by contrast, increased punishment sensitivity (p=0.003) and reduced learning rate (p=0.009), parameters which were not influenced by pathological anxiety (p>0.05). Conclusions: Pathological anxiety was associated with increased reliance on Pavlovian heuristics indicating it may hold value as a potential target for intervention. Indeed, SSRIs have been shown to modulate this Pavlovian parameter (Guitart-Masip et al. 2014, Psychopharm.,231:955–966), a process which may contribute to their anxiolytic effect. Future work will explore the potential of A100 ABSTRACTS behavioural training to counteract these biases as a potential psychological intervention for anxiety disorders. Funded by a Medical Research Foundation Equipment Competition Grant (C0497, Principal Investigator OJR), and a Medical Research Council Career Development Award to OJR (MR/K024280/1).

D02 THREAT OF SHOCK AND AVERSIVE INHIBITION: INDUCED ANXIETY MODULATES PAVLOVIAN- INSTRUMENTAL INTERACTIONS Mkrtchian A, ICN UCL, 17 Queen Square, London WC1N 3AR [email protected] Roiser JP(1), Robinson OJ(1) (1) As presenting author Introduction: To understand anxiety disorders, it is crucial to understand how adaptive anxiety affects behaviour. This may be examined by experimentally inducing anxiety in healthy individuals using the translational threat of shock paradigm. Accumulating work with this paradigm suggests that anxiety promotes harm-avoidant behaviours, possibly through enhanced inhibitory behaviour. However, the specific underlying mechanism is unknown as behaviour involves both Pavlovian (“hard-wired” invigoration towards rewards/inhibition towards punishments) and instrumental (flexible modification of actions to optimise rewards/minimise punishments) systems. This study examined how threat of shock impacts aversive and appetitive Pavlovian-instrumental interactions in healthy individuals. Methods: Healthy participants (N=62) performed a Reinforced go/no-go task (Crockett et al., 2009, Journal of Neuroscience, 29(38), 11993-11999), which assesses aversive and appetitive Pavlovian-instrumental interactions by response vigour in four different conditions where ‘action’ (go, no-go) and ‘valence’ (reward, punishment) is manipulated. Critically, in the ‘go to win’ (GW) and the ‘no-go to avoid’ (NGA) conditions, the Pavlovian and instrumental systems cooperate (evidenced by faster responses) by activating the same actions towards outcomes. In the ‘no-go to win’ (NGW) and the ‘go to win’ (GW) conditions, the two systems compete against each other (evidenced by slowing responses) as they activate opposing preferred actions towards outcomes. The task was performed under alternating threat (risk of an unpredictable shock) and safe blocks. Results: A repeated measures ANOVA revealed a significant three-way interaction between threat, action and valence (F(1,61)=4.83, p=0.032). This was driven by threat, compared with safe, significantly slowing responses (increased inhibition) during the GA condition (F(1,61)=4.39, p=0.04) but there was no significant difference between threat and safe for the other three conditions (GW, NGW, NGA), p>0.05 for all. Conclusions: Threat of shock promotes reliance on Pavlovian aversive biases, as evidenced by threat slowing responses selectively when participants had to speed responses to avoid a punishment (the aversive Pavlovian-instrumental conflict condition), whilst leaving appetitive Pavlovian-instrumental interactions unperturbed. Critically, this may drive adaptive harm-avoidant behaviour in threatening situations where Pavlovian and instrumental processes are aligned, but may result in maladaptive behaviours when instrumental control is preferable. These findings highlight a possible neurocognitive mechanism underlying anxiety, which may inform both diagnosis and treatment. Specifically, treatments may seek to target instrumental over-ride of Pavlovian control and future studies should therefore aim to explore these effects in patient populations. Funding: Medical Research Foundation Equipment Competition Grant (C0497, Principal Investigator OJR), and a Medical Research Council Career Development Award to OJR (MR/K024280/1).

D03 NITROUS OXIDE SPEEDS THE REDUCTION OF DISTRESSING INTRUSIVE MEMORIES IN AN EXPERIMENTAL MODEL OF PSYCHOLOGICAL TRAUMA Das RK, Clinical Psychopharmacology Unit, UCL, 1–19 Torrington Place London WC1E 7HB [email protected] Tamman A(1), Nikolova V(1), Freeman TP(1), Kamboj SK(1), Lazzarino AI(2), Bisby JA(3) (1) As presenting author; (2) Dept of Epidemiology and Public Health, 1–19 Torringtone Place, UCL, London WC1E 7HB; (3) Inst of Cognitive Neuroscience, 17–19 Queen Square, London WC1N 3AR Background: Post-traumatic stress disorder (PTSD) involves maladaptive long-term memory formation which underlies involuntary intrusive thoughts about the trauma. Such intrusions are one of the cardinal ABSTRACTS A101 and most disabling symptoms of PTSD. Preventing the development of such maladaptive memory is a key aim in preventing the development of PTSD. We examined whether the N-methyl D-aspartate receptor (NMDAR) antagonist gas nitrous oxide (N2O) could reduce the frequency of intrusive memories by inhibiting NMDAR-dependent memory consolidation in a laboratory analogue of psychological trauma. Method: Participants were randomised to inhale N2O (N = 25) or medical air (N = 25) for thirty minutes after viewing a negatively-valenced emotional film clip (‘trauma film’). Participants subsequently completed a daily diary assessing frequency of intrusive thoughts relating to the film clip. A week later, participants completed an explicit memory recall task related to the film. Psychophysiological and subjective measures of acute responses to the film and drugs were also collected. Results: Post-encoding N2O sped the reduction in intrusive memory frequency [F (6,315) = 2.382, p = 0.029), with a significant reduction by the next day in the N2O group (β = −1.33, t(315) = −2.767, p = 0.007, 95% CI −2.32 to −0.341) compared to 4 days later in the Air group. Importantly, N2O did not interfere with putatively adaptive, verbally accessible encoding of the film [t(48) = 0.696, p = 0.49, r = .1]. N2O interacted with post-film dissociation, producing increased intrusion frequency in those who were highly dissociated at baseline. Sleep length and quality the night after viewing the film did not differ between the groups. Groups did not differ in psychophysiological or subjective responses to the trauma film. Conclusion: N2O speeds the reduction of intrusive analogue trauma memory in a time-dependent manner, consistent with sleep- dependent long-term consolidation disruption. These results are parsimoniously described by disruption of the early phase of a synaptic tag-and-capture model of consolidation. Further research with N2O is warranted to determine its potential to inoculate against enduring effects of psychological trauma; however, caution is also warranted in dissociated individuals where N2O may aggravate PTSD-like symptomatology. Given that N2O is already used on the NHS as a pre-hospital anaesthetic, it may be having unforeseen prophylactic consequences on the development of trauma-like symptomatology. This research was funded entirely through internal sources.

D04 RISK AND LOSS AVERSION IN CLINICAL ANXIETY Charpentier CJ, Inst of Cognitive Neuroscience, Univ College London, 17 Queen Square, London WC1N 3AZ [email protected] Roiser JP(1), Robinson OJ(1) (1) As presenting author Introduction: Anxiety disorders are associated with both biased emotional processing and disrupted decision-making. In healthy individuals these processes are known to strongly interact, such that decisions are influenced by emotional cues and states and vice versa. However, whether these interactions are altered in individuals with anxiety disorders remains an unresolved and important question, with the potential to provide a better understanding of the cognitive processes that drive clinical anxiety. Methods: To address this question, patients with Generalized Anxiety Disorder (GAD: N=25) and matched healthy controls (N=23) completed a gambling task, featuring a decision between a risky (gamble) and a safe (certain) option on every trial. Each decision was preceded by happy, fearful, or neutral faces, or object primes. One type of gamble featured only wins (“win-only”), allowing us to assess risk aversion; the other type involved weighing a potential win against a potential loss (“mixed”), allowing us to assess loss aversion. Results: Relative to healthy controls, GAD patients exhibited enhanced risk aversion (t(46)=2.491, P=0.016), but similar levels of loss aversion (t(46)=0.141, P=0.889). However, both risk and loss aversion seemed robust to priming by emotional cues. Conclusions: Patients with anxiety often report difficulties making decisions. These findings suggest that this may be driven by a reduced propensity to take risks, but not by a stronger aversion to losses. By focusing on reducing risk sensitivity, rather than sensitivity to negative outcomes, psychological interventions may be able to target this symptom of anxiety. This work was funded by a UCL Grand Challenge Studentship. A102 ABSTRACTS

D05 AN INVESTIGATION OF PERIPHERAL INFLAMMATORY CYTOKINES IN GENERALISED ANXIETY DISORDER Hou R, Dep. of Psychiatry, Univ of Southampton, Academic Centre, College Keep 4–12, Terminus Terrace, Southampton SO14 3DT [email protected] Holmes C(1), Garner M(1), Osmond C(1), Baldwin D(1) (1) As presenting author Introduction: Greater understanding of the role of cytokines in communications between the nervous and immune systems has led to integrative and explanatory models for neuropsychiatric disorders. The primary aim of this study was to examine peripheral pro-inflammatory and anti-inflammatory cytokines and their balance in generalised anxiety disorder (GAD). Methods: A cross-sectional study using between group comparisons and regression analyses to investigate associations among psychological, cognitive and biological variables. 54 patients with GAD and 64 healthy controls were recruited. Participants completed self-report measures of anxiety, depression, cognition, and an objective test of attention and executive function. A panel of pro-inflammatory and anti-inflammatory cytokines were measured using multiplex technology. Results: Case-control logistic regression analyses revealed significant differences in serum levels of IL-10, TNF-α, and IFN-γ (p<0.01) between GAD and control groups after adjusting for age, gender and BMI: these group differences were independent from the presence and degree of depression. Comparison of pro- and anti-inflammatory cytokine ratios indicated that there were significantly higher ratios of TNF-α /IL10, TNF-α /IL4, IFN-γ /IL10, and IFN-γ /IL4 in the GAD group (p<0.05) when compared to the control group. There were no significant associations between cytokine levels and cognitive measures (p>0.05). Conclusions: The current study is the first to investigate both pro- and anti-inflammatory cytokines and their balance in patients with GAD in comparison to healthy controls. The findings indicate an increased pro-inflammatory response and a decreased anti-inflammatory response. More importantly, the changes in serum cytokine levels in GAD are independent of depression. The balance between type 1 helper cells (Th1) and type 2 helper cells (Th2) is an essential determinant in containing the inflammatory response and a delicate balance is required for normal regulation of neuropsychiatric functioning. Findings of the study provide the first demonstration of the Th1/Th2 cytokine imbalance in GAD. The study was funded by the Univ of Southampton.

D06 ELEVATED CORTISOL AWAKENING RESPONSE ASSOCIATED WITH EARLY LIFE STRESS AND IMPAIRED EXECUTIVE FUNCTION IN HEALTHY ADULT MALES Butler K, School of Psychology, Univ of Lincoln, Brayford Pool, Lincoln LN6 7TS [email protected] Klaus K(1), Pennington K(1) (1) As presenting author Introduction: Hypothalamic-pituitary-adrenal (HPA) axis dysfunction may play a role in the aetiology of mental health disorders. More specifically, an increased cortisol awakening response (CAR; an index of HPA-axis function) may predict subsequent depression (Adam et al. 2010, Psychoneuroendocrinology 35: 921–31) whereas blunted CAR has been reported in individuals with familial risk for psychosis (Cullen et al. 2014, Psychoneuroendocrinology 46: 1–13). Both higher and lower CAR has also been shown to be associated with specific risk factors for depression and psychosis such as early life stress (ELS) and impaired executive function. This study aimed to investigate further the association of CAR with both ELS and executive function in healthy adult males. Methods: Our sample consisted of 106 healthy adult males (mean age 34.5 years, range 21 to 63 years) with no current psychiatric diagnosis (self-reported) who were recruited from Lincoln, UK, and surrounding areas. Saliva was collected for cortisol analysis at awakening (T0) and 30 minutes post-awakening (T30) on two consecutive days prior to a test session that included the Childhood Traumatic Events Scale (CTES; Pennebaker and Susman, 1988, Sos. Sci. Med. 26: 327–32). CTES considers exposure to 6 types of childhood traumatic experience (death, divorce or parental separation, traumatic sexual experience, violence, illness or injury and other) up to 17 years old. Executive ABSTRACTS A103 function was assessed with the Spatial Working Memory (SWM), Intra-Extra Dimensional Set Shift (IED) and One Touch Stockings of Cambridge (OTS) tasks (CANTAB, Cambridge Cognition Ltd.). Results: Across all participants T30 cortisol (mean: 9.03nmol/L, SE: 0.52) was significantly greater than T0 cortisol (mean: 7.51nmol/L, SE: 0.42; p<.001). CAR (range: -11.48–18.25nmol/L) was significantly greater in those exposed to ELS (n=39, mean: 2.68nmol/L SE: 0.71) compared to those who were not (n=67, mean: 0.85nmol/L SE: 0.44; p<.05). CAR significantly negatively correlated with OTS performance (r = -0.271, p=.005) but not with performance in the other executive function measures. Conclusions: This study supports previous research which has found that occurrence of ELS is associated with an increased CAR. In addition, it is the first study to show that increased CAR is significantly associated with impaired planning and problem solving in healthy adult males. The CAR may reflect a biological mechanism by which ELS and executive function act as risk factors for psychopathology and its use as a potential biomarker, in studies assessing cognitive function in different at risk populations, should be explored in future work. Sources of financial sponsorship: This work was funded by the Univ of Lincoln’s School of Psychology, the College of Social Science Research Fund and the Academic Return to Research Fund (R2F).

D07 RELATIONSHIP BETWEEN HEART RATE VARIABILITY, PERCEIVED STRESS AND MENTAL DISORDER IN BRAZILIAN HEALTH PROFESSIONALS IN EMERGENCY SERVICES Juruena MF, Psychological Medicine, Centre for Affective Disorders, IoPPN- King’s College London, PO72, De Crespigny Park, Denmark Hill, London SE5 8AF [email protected] Cardoso L(1), Pereira SS(1), Papathanassoglou E(2), Hegadoren K(2), Pereira Júnior GA(3) (1) Ribeirao Preto College of Nursing, Univ of Sao Paulo, Brazil; (2) Univ of Alberta, Nursing Faculty, Edmonton, Alberta/ Canada; (3) Univ of São Paulo, Ribeirão Preto Medical School Introduction: Heart rate variability (HRV) is a marker of cardiac autonomic function and changes in HRV have been linked to cardiovascular disease and chronic stress. Chronic occupational stress has been reported in those that work in emergency Depts (ED). The goal of this study was to assess the relationships among levels of perceived stress, mental disorders and HRV measures in Brazilian health professionals in the ED. Methods: This was a cross-sectional descriptive study in a stratified random sample of 53 health care providers (registered nurses, nursing assistants and physicians) in an urban tertiary care hospital in the state of Sao Paulo, Brazil. A minimum of 4-hour RR variability was analyzed (2 hours at beginning and 2 hours at the end of their shift) for each professional with a validated device (Polar V800). All data were systematically post-processed, and the standard deviation of all normal RR Intervals (SDNN) and Low Frequency/high Frequency ratio (LF/HF) were used as markers of HRV and sympathovagal balance, respectively. Current level of stress was assessed by the Perceived Stress Scale (PSS) and the Self-reporting Questionnaire (SRQ20) was used to detect mental disorders. Descriptive analysis and bivariate correlations were performed in SPSS (alpha level < 0.05). All ethical principles were respected. Results: Preliminary demographic data included: type of care provider (42% nursing assistants, 13.2% nurses, 7.6% physicians), age in years (mean = 36.1± 8.3), gender (86.8% women), duration of service in years (mean= 8.5± 7.3) and length of shift in hours (7.5). Mean SDNN values in the first two hours of each shift were 40.67ms±20.7 and LF/HF mean 4.4±2.4. The mean SDNN values did not change in the last 2 hours of recording (39.32ms±16.2), as opposed to LF/HF values that were lower (M=3.7±2.4). The average PSS score was 22.40±10.7 and 28.3% of these professionals meet criteria for a mental disorder. Perceived stress scores were strongly associated with presence of a mental disorder (r=0.754; p=0,000) and weakly with LF/HF in the last two hours of the shift (r=0,241; p=0,041). Conclusion: Although the physiologic interpretation of LF/HF is not clear-cut these results provide evidence of higher than normal LF/HF, which may indicate heightened sympathetic activity. Moreover, these findings provide preliminary evidence that level of perceived stress may correlate with LF/HF in care providers in ED. More work is needed to explore if this relationship is causal and potential mediating factors. Financial source: Sao Paulo Research Foundation – FAPESP. A104 ABSTRACTS

D08 PARENTAL CHILDHOOD ABUSE AND NEONATAL HPA AXIS FUNCTION: MECHANISMS OF RISK TRANSMISSION Sethna V, Division of Psychological Medicine, Section of Stress, Psychiatry and Immunology Lab & Perinatal Psychiatry, King’s College London, IOPPN, The Maurice Wohl Clinical Neuroscience Inst Cutcombe Road, Brixton, London SE5 9RT [email protected] Conroy S(1), Pawlby S(1), Pariante CM(1) (1) As presenting author Introduction: Parental exposure to child abuse may have transgenerational effects, with offspring of abuse victims showing dysfunctional neuroendocrine profiles. Although the time periods and mechanisms of risk transmission remain unclear, dysregulated maternal cortisol levels during pregnancy and poor behavioural regulation in neonates may be involved. Infants born to abused women show blunted cortisol levels and poor birth outcomes – male infants may be particularly responsive to maternal HPA axis dysregulation. Therefore, the current study investigates, for the first time, whether maternal HPA axis functioning in pregnancy and neonate behaviour post-birth mediate the association between maternal childhood abuse and neonate HPA axis function. A further aim was to examine the role of infant gender in the transmission of risk. Methods: Maternal childhood abuse exposure was assessed retrospectively using the Childhood Experience of Care and Abuse questionnaire (abuse present: n=70, 38 males; abuse absent: n=81, 46 males). At 25 weeks gestation maternal cortisol awakening response (CAR) was computed from cortisol samples at wakeup, wake +15 min, wake +30 min and wake +60 min. At 6 days post delivery neonatal behavioral adjustment was examined using the Neonatal Behavioral Assessment Scale and cortisol was measured before (pre) and immediately after (post). Results: Using regression based path analysis (controlling for socio-demographic functioning, maternal depression and birth outcomes) neonatal cortisol functioning was not directly explained by maternal childhood abuse but indirectly via mothers CAR at 25 weeks and via infant’s autonomic stability at 6 days - these pathways were evident in male infants only (β = -3.40, p <0.05 and β = 2.58, p<0.05 respectively). In contrast, 6-day cortisol levels in infant females were not directly explained by maternal childhood abuse but indirectly via infant social interactive functioning (β = 4.23, p < 0.05). Furthermore, there was no evidence of mediation by maternal CAR during pregnancy in female infants. Conclusion: These findings link maternal childhood abuse exposure with neonatal stress physiology and suggest differential pathways of intergenerational transmission for males and females. Our results also have implications for understanding the relative vulnerability of males to developmental problems which may originate in the uterine environment. Financial sponsorship: None.

D09 IS SOMATIC ANXIETY IN PREGNANCY ASSOCIATED WITH INATTENTION IN CHILDREN? Bolea Alamanac BM, School of Social and Community Medicine, Univ of Bristol, Oakfield House, Oakfield Grove Bristol BS8 2BN [email protected] Davies SJC(1) (1) Centre for Addiction and Mental Health, 80 Workman Way Toronto, Ontario M6J 1H4 Introduction- It is possible that stress hormones produced by anxious mothers during pregnancy could interfere with neural migration and dendritic growth in key areas of the foetal brain. Some studies have indicated that these neurodevelopmental abnormalities could alter the balance between inhibitory and excitatory neural circuits, increasing the severity of ADHD symptomatology in the child. Methods -The Avon Longitudinal Study of Parents and Children (ALSPAC) collected data of mental state during pregnancy of 14541 women employing the Crown Crisp Experiential Index (CCEI). Using confirmatory factor analysis a somatic anxiety factor was extracted from this scale. It contained the following items: dizziness / shortness of breath, feeling faint, feeling sick, tingling sensations and excess sweating. It was converted to a binary measure with the cutoff being the top 15% of scores. At age 8.5 years children of these mothers undertook three subtests of the Test of Everyday Attention for Children (TEA-CH): Sky Search, Sky Search Dual Test and Opposite Worlds. Each of these is representative of an attentional skill: sustained attention, selective attention and attentional control respectively. Scores were adjusted for motor speed. A linear regression model was fitted to evaluate the association between maternal somatic anxiety during ABSTRACTS A105 pregnancy and TEA-CH scores. Sociodemographic variables were included as covariates. Results- A total of 4198 children for the Sky Search subtest, 3845 for the Sky Search Dual Test and 4202 for the Opposite Worlds subtest had complete data and were included in the analyses. Children exposed to high maternal somatic anxiety did not have significantly poorer scores in any of the TEA-CH subtests than non-exposed children either before or after adjusting for confounders [Sky Search: OR=1.00, (0.98–1.01, p=0.821), Sky Search Dual Test: OR=0.94, (0.82–1.07, p=0.357), Opposite Worlds: OR= 1.00 (0.95–1.05, p=0.850), all analyses fully adjusted]. There were no differences by gender. Imputation for missing data did not change the results. Conclusions- No evidence was found for an association between somatic anxiety during pregnancy and inattention symptoms in children at age 8.5 years measured by three subtests of the TEA-CH. The lack of association for somatic anxiety contrasts markedly with existing evidence on depression, it remains possible that the attentional tests used were not sufficiently sensitive to detect differences at this age. The study conforms to the ethical standards established by the ALSPAC ethics committee as stated in http:// www.bristol.ac.uk/alspac/researchers/data-access/ethics/. No funding was required.

D10 WHAT IS KNOWN ABOUT ADULT SEPARATION ANXIETY DISORDER? A SYSTEMATIC REVIEW Gordon RP, Univ Dept of Psychiatry, Univ of Southampton, College Keep, 4–12 Terminus Terrace, Southampton SO14 3DT [email protected] Huneke N(1), Baldwin DS(1) (1) Univ Dept of Psychiatry, College Keep, Univ of Southampton Introduction: The DSM-5 includes the category of ‘separation anxiety disorder’ among the anxiety disorders diagnosable in adults (previously it was regarded exclusively as a childhood-onset condition), even though little is known about the condition. We therefore undertook a systematic review of its clinical features, prevalence, comorbidity, course, aetiology and treatment. Method: Search of electronic data bases (‘Medline’, ‘PsychInfo’) for relevant English-language papers on adult separation anxiety disorder (ASAD) to March 2016. Papers with a focus on childhood separation anxiety disorder only were excluded. Eligibility assessment was carried out independently by two reviewers (RG, NH). Disagreements between reviewers was resolved by consensus. Results: The database search yielded 375 citations, of which 108 papers met our inclusion criteria. Study heterogeneity precluded meta-analysis. Epidemiological studies (n=2) estimate the lifetime prevalence of ASAD to be between 4.8–6.6% and 12 month prevalence as 1.9%; between 43.1% and 75.2% of ‘cases’ in adults had a reported onset in adulthood. Psychopathological investigations (n=9) emphasise that distinction of ASAD from other anxiety disorders (such as agoraphobia, panic disorder and generalised anxiety disorder) is a challenge. Aetiological studies (n=9) include aberrant attachment styles, polymorphisms in oxytocin receptor genes, and hypersensitivity to inhalation of CO2. Clinical outcome is negatively influenced by the presence of co-morbid separation anxiety. There are no published randomised controlled trials of pharmacological or psychological treatment approaches for patients with ASAD. Conclusions: There is considerable scope for extending current understanding of ASAD. Prospective studies of prognosis and randomised controlled trials are needed to aid the development of new treatment guidelines. RG and NH are NIHR Clinical Academic Fellows with supportive grants from the Research Management Committee of the Faculty of Medicine at the Univ of Southampton.

D11 DECLINE IN PRESCRIPTIONS TO THE OVER 65S IN THE ONTARIO DRUG BENEFIT DATABASE FROM 1998 TO 2013, AND IMPACT OF LEGISLATIVE CHANGES ON BENZODIAZEPINE PRESCRIBING Davies SJC, Geriatric Psychiatry Division, CAMH and Dept of Psychiatry, Univ of Toronto, 6th Floor, 80 Workman Way, Toronto, ON, Canada M6J1H4 [email protected] Jacob B(1), Rudoler D(1), Zaheer J(1), deOliveira C(1), Kurdyak P(1) (1) CAMH/Univ of Toronto Introduction: The Ontario Drug Benefit (ODB) database captures all provincially funded prescriptions A106 ABSTRACTS dispensed to Ontario Residents. Since most drugs, including , are funded for residents aged >=65, data for this group are essentially population-based. Benzodiazepine prescribing patterns from 1998 – 2013 were described. Method: ODB data were abstracted for each year for residents aged >= 65. We examined prescription rates (percentage of population having at least one prescription in the calendar year) for all benzodiazepines, anxiolytics vs hypnotics and individual benzodiazepines. Results were stratified by age and gender. We examined impact on prescribing rates of provincial legislative changes tightening benzodiazepine prescription requirements from November 2011. Impact on monthly benzodiazepine prescribing rate was explored by constructing an ARMAX model. Data prior to November 2011 were used to forecast prescribing rates to December 2013. These were compared with actual observed prescribing rates. The effect of legislative changes was estimated using a step function. Results: Benzodiazepines were prescribed to 23.2% of the Ontario population aged over 65 (n=1,412,638) in 1998, declining to 14.9% in 2013 (n=2,057,899) (p<0.001 for trend). Hypnotic benzodiazepines declined more rapidly than anxiolytic. Lorazepam was the most prescribed benzodiazepine throughout, but rates declined from 11.4% in 1998 to 8.5% in 2013. rates fell from 2.3% in 1998 to 0.7% in 2013. Oxazepam, alprazolam and nitrazepam also declined substantially. By contrast, prescription rates increased consistently until 2011, 1.7 fold overall. After the changes of November 2011, exaggerated declines were observed both in total prescription rates and for each drug individually. Monthly data for total benzodiazepine prescribing rates revealed that after the changes, observed rates for most months were below the lower 95% confidence limit predicted from data prior to November 2011. The step function, conditional on covariates, suggested benzodiazepine rates were 0.3% (p<0.001) below forecast rates. Prescribing rates were significantly greater throughout the observation period in older age groups (p<0.001) and females (p<0.001). Discussion: Benzodiazepine prescribing rates declined markedly from 1998 to 2013. The introduction of z-drugs may explain the greater reduction in prescription of hypnotic benzodiazepines compared with anxiolytics. However, zopiclone, the sole ‘z-drug’ available in Ontario, is funded only in special circumstances thus zopiclone prescribing trends could not be examined. Reasons for increasing clonazepam prescribing rates, while all others declined, require further elucidation. Benzodiazepine prescribing rates declined markedly from 1998 to 2013. The introduction of z-drugs may explain the greater reduction in prescription of hypnotic benzodiazepines compared with anxiolytics. However, zopiclone, the sole ‘z-drug’ available in Ontario, is funded only in special circumstances thus zopiclone prescribing trends could not be examined. Reasons for increasing clonazepam prescribing rates, while all others declined, require further elucidation. The exaggerated decline in prescription rates after November 2011 suggests that targeted legislation may reduce benzodiazepine prescribing, but the effect, while statistically significant, was small. No External Funding.

E01 MATERNAL IMMUNE ACTIVATION IS ASSOCIATED WITH EARLY NEURODEVELOPMENTAL, GROWTH AND PLACENTAL ABNORMALITIES IN WISTAR RATS Edye ME, Manchester Pharmacy School, Univ of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT [email protected] Murray KN(1), Dennison J(1), Harte MK(1), Neill JC(1), Knuesel I(2), Prinssen E(2) (1) As presenting author; (2) Roche Pharmaceutical Research and Early Development, Roche Innovation Center Basel, 124 Grenzacherstrasse, Basel, CH 4070, Switzerland Maternal immune activation (mIA) is a key model for neurodevelopmental disorders (NDDs) such as autism spectrum disorder and schizophrenia. However, mIA studies predominantly use mice despite the benefits of rats in cognitively-demanding or social behavioural tasks and neuroimaging read-outs. Moreover, experimental studies often only use adult males. Thus we aimed to validate a rat model of mIA by conducting a comprehensive dose-response study in three strains of rats before investigating the morphological and molecular consequences in both male and female offspring at early developmental time points. Following strain (Wistar, Hooded Lister, Sprague Dawley) and dose (2.5–15mg/kg) optimisation, mIA was induced in pregnant Wistar rats with 10mg/kg poly(I:C) i.p. at gestational day (GD)15 (n=2–3 per timepoint). Offspring brain and body weight were assessed in males and females ABSTRACTS A107 at GD21 (n=17–23) and postnatal day (PD)21 (n=8–12). Brains were harvested for quantitative RT-PCR to measure changes in gene expression. Statistical analysis was performed using Kruskal-Wallis test followed by Dunn’s multiple comparison test. A nested ANOVA was performed for comparisons between offspring. Mann-Whitney test was used when parametric analysis was not suitable. Poly(I:C) at 10mg/ kg induced the most robust immune response in Wistar rats, thus this dose and strain were selected for future experiments. Poly(I:C) at 10mg/kg on GD15 reduced body weight (P<0.001) and placenta weight (P<0.001) in pups at GD21 and this reduction in body weight was maintained at PD21 (P<0.001). Reduced expression of myelin basic protein (1.6 fold males, 1.8 fold females; P<0.01), major facilitator superfamily domain containing protein 2a (8.1 fold males, 4.2 fold females; P<0.01) and glial fibrillary acidic protein (1.9 fold males; P<0.01) was observed in the frontal cortex of offspring of poly(I:C)-treated dams at GD21. Myelin basic protein expression was elevated at PD21 (9.0 fold males; 10.9 fold females; P<0.01), alongside elevated expression of myocyte enhancer factor-2 (2.0 fold males, 2.3 fold females; P<0.01) and semaphorin 3a (4.4 fold males, 4.1 fold females; P<0.001), and reduced Shank3 (2.6 fold males, 2.4 fold females; P<0.05). Similar results were observed for male and female offspring. This study provides a robust parametric assessment of poly(I:C)-induced mIA in rats and explores the consequences of this immune challenge on morphology and gene expression in male and female offspring. Our data show that poly(I:C) results in smaller offspring with altered gene expression linked to impaired synaptic function. This work provides a framework to study the developmental trajectory of disease-relevant, sex- specific phenotypic changes in rats. Sources of financial sponsorship: This work is supported by Roche and the Univ of Manchester MRC Confidence in Concept scheme.

E02 THIS ABSTRACT HAS BEEN WITHDRAWN

E03 CAN NEUROINFLAMMATION INDUCED BY INTRA-NASAL LIPOPOLYSACCHARIDES INJECTIONS CAUSE ALZHEIMER’S LIKE PATHOLOGY IN AGED WILD TYPE MALE MICE? Seroiska R, School of Life Sciences, Univ of Nottingham, Queen’s Medical Centre, Derby Rd, Nottingham NG7 2UH [email protected] *Carneiro MSF(1), Agostini A(1), Barron M(1), Pardon MC(1) (1) As presenting author Introduction: Alzheimer’s disease (AD), the commonest cause of dementia, generally presents with memory loss and cognitive impairment. AD is characterized by the presence of amyloid plaques, formed by extracellular aggregates of beta-amyloid peptides and neurofibrillary tangles, composed of intracellular phosphorylated tau. Recently, neuroinflammation emerged as a possible mechanism leading to AD. Mice models of AD exist, although none fully recapitulates the disease, with most being transgenic models based on the rare familial subtype of AD, rather than the common sporadic subtype, which has age as the most well known risk factor. To test whether neuroinflammation causes AD-like pathology, we used intra-nasal lipopolysaccharides (LPS) injections to, non-invasively, overcome the blood-brain-barrier. We hypothesize that intra-nasal LPS injections in aged wild type mice will trigger neuroinflammatory responses causing AD-like behavioural and neuropathological features. Methods: 20 months old wild type male mice; (average weight: 48±10g) received 3 intranasal injections of either vehicle phosphate buffered saline (PBS, n=5) or LPS (10ug each; n=5) over a period of 24 hours. For behavioural assessment we used Food Burrowing test, as a measure of activity of daily living and Spontaneous Alternation (Y-maze) for locomotor activity, spatial working memory and sickness syndrome. Immune responses in the hippocampus were assessed by immunostaining of glial markers (IBA-1 for microglia and GFAP for astrocytes). Data analysis (Student’s t-test and Two-Way ANOVA) was done using Graphpad Prism. Immunostaining of amyloid plaques and western immunoblotting of total and phosphorylated tau will be performed. Results: Food Burrowing performance was not significantly altered by LPS (PBS: 21.5±8.3g; LPS: 20.0±13.6g, p=0.78). Spatial working memory was also unaffected (alternation rate: PBS: 64.4±4.8%; LPS: 64.0±5.3%, p=0.58). There was no evidence of astrogliosis in the hippocampus. PBS- and LPS-treated A108 ABSTRACTS mice did not differ for the percentage area stained by GFAP (PBS: 5.2±0.2%; LPS: 5.5±0.2%, p = 0.54), or with IBA-1 (PBS: 3.8±0.2%; LPS: 3.9±0.2%; p = 0.67). Microglial clusters were identified, which may imply amyloid-beta deposition. Conclusions: preliminary data suggests that intranasal LPS injections has no major behavioural effect, possibly due to lack of astrogliosis in the hippocampus. One possibility is that the transit of LPS via the olfactory nerve pathway does not effectively reach the hippocampus. Thus, other brain areas alongside this pathway will need to be analysed. Authors received scholarships granted by The Ministry of Science, Technology and Innovation of Brazil.

E04 THIS ABSTRACT HAS BEEN WITHDRAWN

E05 N-3 FATTY ACIDS INTAKE ASSOCIATED WITH DELAY AVERSION AND TEMPORAL PROCESSING, BUT NOT INHIBITORY CONTROL, IN CHILDREN WITH ATTENTION DEFICIT HYPERACTIVITY DISORDER (ADHD) Chang JP, Psychological Medicine, IOPPN, King’s College London, The Maurice Wohl Clinical Neuroscience Inst, Cutcombe Road, London SE5 9RT [email protected] Mondelli V(1), Pariante CM(1), Lu YJ(2), Su KP(3), Jingling L(4), Huang YT(4) (1) As presenting author; (2) Bao-Bei Child Rehabilitation Clinic, Taoyuan, Taiwan; (3) Dept of Psychiatry and Mind-Body Interface Lab (MBI Lab), China Medical Univ Hospital, Taichung, Taiwan; (4) Graduate Inst of Cognitive and Neural Sciences, China Medical Univ, Taichung, Taiwan Introduction: Omega-3 polyunsaturated fatty acids (n3-PUFAs) deficiency has been associated with the manifestations of attention deficit hyperactivity disorder (ADHD) (Antalis et al, 2006, Prostaglandins Leukotrienes and Essential Fatty Acids, 75, 299–308), and it has also been associated with cognitive performance (van der Wurff et al, 2016, Nutrients, 8, piiE13.) and emotional regulation (Lin et al, 2010, Biological Psychiatry, 68, 140–7). The aim of this study is to investigate the association between n3-PUFAs intake and neurocognitive functions in children with ADHD. Methods: Forty-two subjects were enrolled in the study, where 21 of them were diagnosed with DSM-IV ADHD, and 21 of them had no life-time diagnosis of ADHD. The n-3 PUFAs intake was recorded by Food Frequency Questionnaire. Essential fatty acid (EFA) deficiency severity was defined by EFA deficiency scale, and ADHD symptom severity was evaluated by SNAP-IV. The children were also assessed by Go-No-Go Task, Delayed Reaction Time Task, and Finger Tapping Task for inhibitory control, delay aversion, and temporal processing. Results: The ADHD group had a greater EFA deficiency severity (7.24 + 4.56, p= 0.02) and performed more poorly in delay aversion (-177.88+280.40, p= 0.02) and temporal processing (85.34+10.96, p< 0.001) tasks. Moreover, ADHD symptom severity was positively correlated with EFA deficiency severity. In addition, EFA deficiency severity was associated with a higher delay aversion (p< 0.001). Conclusion: Children with ADHD had a higher EFA deficiency, and the EFA deficiency was associated with a greater severity in ADHD symptomatology and delay aversion. Funding:This work was supported by the following grants: MOST 103-2320-B-039-030- from Ministry of science and Technology in Taiwan, DMR100-IRB-160 and CMUH104-REC2-058 from China Medical Univ in Taiwan.

E06 FRACTIONATING APATHY – EFFORT, REWARD AND IMPULSIVITY MECHANISMS IN PARKINSON’S AND HUNTINGTON’S DISEASE O’Callaghan C, Behavioural and Clinical Neuroscience Inst, Univ of Cambridge, Sir William Hardy Building, Downing Street, Cambridge CB2 3EB [email protected] Mason SL(1), Barker RA(1), Sahakian BJ(2), Robbins TW(3) (1) Dept of Clinical Neurosciences, Univ of Cambridge; (2) Dept of Psychiatry, Univ of Cambridge; (3) Dept of Psychology, Univ of Cambridge Introduction: Apathy is pervasive across neuropsychiatric conditions. Defined as a loss of motivation, currently little progress has been made to identify the cognitive and neural substrates that underpin this ABSTRACTS A109 broad construct. Establishing objective measures to assess apathy and identifying its sub-processes will be critical to advancing our knowledge of the neural systems involved and to promote novel drug therapies. Here we use neurodegenerative diseases as model to identify objective measures to assess apathy and parcellate the construct into sub-components. Methods: In a cohort of Parkinson’s disease (PD) and Huntington’s disease (HD) patients we administered a battery of novel behavioural tasks to measure causal components involved in apathy, including reward sensitivity, willingness to exert effort and reinforcement learning. Questionnaire assessment of everyday apathy, mood and impulsivity was also included. Results: Key preliminary results include strong correlations between extent of everyday apathy severity in HD and reduced effort on a novel translational Progressive Ratio task (r=-.55, p<.001). In PD patients, higher levels of apathy were associated with increased loss aversion on a reinforcement learning task (r=.5, p<.05) and higher apathy severity in daily life was associated with a more conservative strategy on a novel betting task (r=-53, p<.05) consistent with reduced sensitivity to positive reward. Extensions to the current analyses include computational modelling of the data and factor analysis to determine apathy sub-types. Conclusions: We identify relationships between novel, objective apathy assessments and apathy severity in daily life. These preliminary findings show that specific cognitive processes are related to everyday manifestations of apathy. Specifically, that increased apathy is accompanied by alterations in reward processing and effort-maintenance, highlighting these as potential mechanisms that might contribute to the development of apathy in clinical conditions. These findings have implications beyond Huntington’s and Parkinson’s disease and can inform the processes underlying apathy in other neuropsychiatric conditions. Funding statement: CO is supported by an Australian National Health and Medical Research Council Neil Hamilton Fairley Fellowship (GNT1091310); The Behavioural and Clinical Neuroscience Inst is supported by a joint award from the Medical Research Council and Wellcome Trust.

E07 SYMPTOMS OF ATTENTION DEFICIT HYPERACTIVITY DISORDER (ADHD) AND DISORDERED EATING: A DUAL PATTERN ASSOCIATION Kaisari P., The School of Psychology, Univ of Birmingham, Edgbaston, Birmingham B15 2TT [email protected] Dourish C.T. (1), Higgs S. (2) (1) 2P1vital, Wallingford, Oxfordshire, UK; (2) School of Psychology, Univ of Birmingham, Birmingham, UK Background/Aims: Attention Deficit Hyperactivity Disorder (ADHD) has been associated with disordered eating. However, whether core symptoms of ADHD are specifically associated with subtypes of disordered eating, remains unclear. In addition, there has been little investigation of mediating factors that may have important implications for treatment. This study investigated the potential relationships between ADHD symptomatology and binge eating versus restrictive eating and assessed whether depression and/or attention to hunger/satiety signals mediate any associations that were identified. Methods: The participants were 256 adults (27.0% males) aged 18–60 years (M=26.9) who were invited to take part in an online study. Symptoms of ADHD were assessed using the Conner’s Adult ADHD Rating Scale Self Report-Screening Version. The Binge Eating Scale (BES), the Bulimic Investigatory Test, Edinburgh (BITE), the Loss of Control over Eating Scale (Brief version) the Dutch Eating Behaviour Questionnaire and the Eating Attitudes Test (EAT-26) were used to assess disordered eating. Levels of current depression were assessed using the Hospital Anxiety and Depression scale (HADS). The Reliance on Hunger/Satiety cues sub-scale of the Intuitive Eating Scale was used to assess eating in response to internal hunger and satiety signals. Results: Principal components analysis reduced the disordered eating measures to two components: ‘’binge/disinhibited eating’’ and ‘’restrictive eating’’. Regression analyses controlled for important confounds (age, gender, Body Mass Index, socio-economic status, alcohol use and ADHD medication) revealed that inattentive but not hyperactive/impulsive symptoms of ADHD predicted binge/disinhibited eating (p<0.05). In contrast, hyperactive/impulsive but not inattentive symptoms of ADHD predicted restrictive eating (p<0.05). Depression was found to be a significant predictor for both binge/ disinhibited eating (p<0.001) and restrictive eating (p<0.001). However, the relationship between symptoms of ADHD and eating pathology was not mediated by depression. Reliance on hunger/satiety cues was a significant mediator between inattentive symptoms and binge/disinhibited eating. Conclusions: Unique associations were identified between inattentive symptoms of ADHD and binge/disinhibited eating and hyperactive/impulsive A110 ABSTRACTS symptoms of ADHD and restrictive eating, suggesting for the first time a double dissociation between specific ADHD symptoms and disordered eating. Neither relationship was explained by depression associated with ADHD symptoms. The results suggest that further investigation of the role of inattentive symptoms in binge eating may be helpful in developing novel treatments for both ADHD and binge eating disorder. Financial support from a BBSRC CASE studentship in collaboration with P1vital.

E08 COMPARATIVE MULTIMODAL META-ANALYSIS OF STRUCTURAL AND FUNCTIONAL BRAIN ABNORMALITIES IN AUTISM SPECTRUM DISORDER AND OBSESSIVE-COMPULSIVE DISORDER Carlisi CO, Child and Adolescent Psychiatry, Inst of Psychiatry, Psychology and Neuroscience, King’s College London, 16 DeCrespigny Park, London SE5 8AF [email protected] Norman LJ(1), Lukito SS(1), Rubia K(1), Mataix-Cols D(2), Radua J(3) (1) As presenting author; (2) Dept of Clinical Neurosci, Karolinska Inst SE-171 77 Stockholm, Sweden; (3) FIDMAG Germanes Hospitalà ries, CIBERSAM, Avda. Jorda, 8, 08035, Barcelona, Spain Objective: Autism spectrum disorder (ASD) and obsessive-compulsive disorder (OCD) share cognitive control deficits which may underlie poor control over stereotyped/repetitive and compulsive behaviours, respectively. However, it is unclear whether these respective symptom profiles are mediated by common or distinct neural profiles. This comparative multimodal meta-analysis assessed shared and disorder-specific neuroanatomy and neurofunction of cognitive control. Methods: A comparative meta-analysis of 62 voxel- based morphometry (VBM) and 26 functional magnetic resonance imaging (fMRI) studies of cognitive control was conducted using anisotropic seed-based d mapping (AES-SDM) comparing regional grey matter volume (GMV) and activation abnormalities between ASD (sMRI:911; fMRI:188) and OCD (sMRI:928; fMRI:247) patients relative to controls. Multimodal meta-analytic comparisons compared groups across both VBM and fMRI. Results: Both disorders shared reduced function and structure in rostral and dorsomedial prefrontal cortex including anterior cingulate. OCD patients had disorder-specific increase in structure and function of left basal ganglia (BG)/insula relative to controls and ASD, who had reduced right BG/insula volumes relative to OCD. In fMRI, ASD patients showed disorder-specific reduced left dorsolateral-prefrontal activation and reduced posterior cingulate deactivation, while OCD patients showed temporo-parietal underactivation. Conclusions: The multimodal comparative meta-analysis shows both shared and disorder-specific brain abnormalities. While rostro-dorsal medial prefrontal cortex was smaller in structure and function in both disorders, this was concomitant with increased structure and function in basal ganglia and insula in OCD, but a reduction in ASD, presumably reflecting a disorder-specific fronto-striato-insular dysregulation in OCD and a fronto-striato-insular maldevelopment in ASD. Disorder-differential mechanisms thus appear to drive overlapping phenotypes of cognitive control abnormalities in ASD and OCD. CC is supported by a National Institute for Health Research (NIHR) Biomedical Research Centre PhD studentship and a King’s College London overseas graduate award. LN and SL are supported by Medical Research Council (MRC) and Inst of Psychiatry, Psychology and Neuroscience Ph.D Excellence awards.

E09 THE QUALITY OF PRESCRIBING OF PSYCHOTROPIC MEDICINES FOR PEOPLE WITH INTELLECTUAL DISABILITY UNDER THE CARE OF UK MENTAL HEALTH SERVICES Paton C, Prescribing Observatory for Mental Health, Royal College of Psychiatrists, 21 Prescot St., London E1 8BB [email protected] Bhatti SF(1), Purandare K(2), Barnes TRE(3), Roy A(4) (1) As presenting author; (2) Central & North West London NHS Foundation Trust, London NW9 9QY; (3) Centre for Mental Health, Imperial College London, W12 0NN; (4) Coventry & Warwickshire Partnership Trust Concerns have been raised about the inappropriate, off-label use of psychotropic medicines in people with intellectual disability (ID) especially in primary care settings. Following reports on care at a private hospital, Winterbourne View, published in 2012 and 2014, NICE published a guideline entitled ‘Challenging ABSTRACTS A111 behaviour and learning disabilities: prevention and interventions for people with learning disability whose behaviour challenges’ (NG11, May 2015). This guideline includes recommendations that treatment for co- morbid mental illness should be optimised and that antipsychotic medication should only be considered for the management of ‘behaviour that challenges’ if other strategies have failed and/or the risk to the person or others is severe, for example violence, aggression or self-injury. In 2009, the Prescribing Observatory for Mental Health (POMH-UK) invited NHS mental health Trusts/healthcare organisations in the UK to participate in an audit-based, quality improvement programme focusing on the prevalence and quality of prescribing of psychotropic medication for people with ID. National clinical audits against NICE-derived practice standards were conducted in 2009, 2011 and 2015. Three hundred and thirty-eight clinical teams from 54 specialist mental health Trusts participated in the 2015 audit, submitting data for 5,654 patients with ID. Antidepressant medication was prescribed for over a third (37%), just over half (54%) of whom had diagnosed depression or anxiety. Nearly two-thirds (64%) of the total national sample were prescribed antipsychotic medication, of whom almost half (49%) had a diagnosis of schizophrenia-spectrum or affective disorder. The clinical reason for prescribing an antipsychotic was violence, aggression or self-injury in just over a third (36%) of cases. The reasons for prescribing antipsychotic medication were clear in almost all cases and medication had been clinically reviewed within the last year in nearly everybody (97%). This suggests that clinical decisions regarding the initiation of antipsychotic medication were considered and that subsequent review of the impact on target symptoms was systematic. However, for 1 in 5 (20%) of the total sample there was no documented evidence of any side-effect assessment over the past year in the clinical records. The monitoring of extrapyramidal and metabolic side effects was inconsistent. The data confirm that psychotropic medicines are commonly prescribed for people with ID. The majority of prescriptions for antipsychotics were for diagnosed comorbid mental illness or behaviours recognised by NICE as potentially legitimate treatment targets. While the quality of prescribing decisions was generally good in specialist service settings, the monitoring of side effects remained inadequate.

F01 THE EFFECTS OF SELECTIVE GLUN2B ANTAGONISM ON PAIRED ASSOCIATIONS LEARNING Kotoula V, Dept. of Neuroimaging, Centre for Neurimaging Sciences, IoPPN King’s College London, 16 de Crespigny Park , London SE5 8AF [email protected] Doyle OM(1), DeSimoni S(1), Williams S(1), Mehta MA(1) (1) As presenting author Introduction: NMDA receptors play a crucial role in cognitive processes. Many agents targeting the GluN2B subunit of the NMDA receptor have demonstrated pro-cognitive properties and thus may benefit various brain disorders. In this BOLD fMRI study, the PAL (Paired Associates Learning) task was used to assess the effect of EVT-101, a selective GluN2B antagonist, on brain regions involved in paired associations learning. Based on i) the brain distribution and density of GluN2B receptors ii) the known role of the hippocampus during the PAL and iii) the results of preclinical research on the effects of selective NMDA receptor antagonism on episodic memory we hypothesised that EVT-101 would, dose dependently, increase activation in brain regions important for the PAL, specifically the hippocampus and parahippocampal gyrus (PHG). Methods: A lower (8mg) and a higher (15mg) dose of EVT-101 or placebo were administered in a double- blind, placebo-controlled, three way cross over design to 21 healthy volunteers who were scanned while performing the PAL task. During the task three repeated phases of active encoding and retrieval of visual spatial associations are followed by the control condition for which no active encoding or retrieval is required. The data were analysed using a standard univariate approach and a flexible factorial design was created in order to examine the effect of the different doses of the drug or placebo on the contrasts of interest. Results: Whole brain analysis in the placebo group revealed that a widespread network of brain regions were activated during the learning phases of the task compared to the control condition. For active encoding compared to control encoding, ROI analysis in the hippocampus and PHG bilaterally, revealed a significant increase in the activation of that region (x=-32 y=-28 z=-23, p<0.05). During retrieval, activation in the PHG significantly decreased for placebo as the phases of the task progressed but this did not occur for either the lower (x =-21 y=-45 z=-11, p<0.05) or the higher dose of the drug (x=-22 y=-49 z=-11, p<0.05). Discussion: Higher doses of EVT-101 significantly altered brain activation in areas that play an important A112 ABSTRACTS role in the encoding and retrieval of visual spatial associations, in the absence of performance changes. The observed imaging changes indicate that potential minimum dose levels were used for this study and further investigation would be required towards a better understanding of the optimal doses of the drug which could produce significant and clinically relevant effects on episodic memory. This work has been supported by Evotec Neurosciences GmbH and the NIHR Biomedical Research Centre for Mental Health.

F02 PHYSIOLOGICAL AND PSYCHOLOGICAL EFFECTS OF THE MONTREAL IMAGING STRESS TASK Martens MAG, Dept of Psychiatry, Univ of Oxford, Neurosciences Bldg, Warneford Hospital, Oxford OX3 7JX [email protected] Tunbridge EM(1), Harrison P(1) (1) Neurosciences Bldg, Univ Dept of Psychiatry, Warneford Hosp, Oxford, OX3 7JX Stress induces a range of physiological effects and is implicated in a number of psychiatric disorders. Acute stress elevates cortical catecholamines and affects cognitive function, including working memory and cognitive flexibility. We investigated the physiological and psychological effects of a laboratory-based stressor, the Montreal Imaging Stress Task (MIST). Non-smoking, healthy men (n=24, mean age = 25.4years) attended the laboratory and, following a one-hour baseline period, were administered either the control or stressor versions of the MIST, before completing a working memory test (the N-back task). Physiological measures of arousal, including salivary-cortisol, heart rate, blood pressure and skin conductance together with subjective stress and mood ratings, assessed using visual analogue scales (VAS), were recorded at 20 minutes intervals. This resulted in 9 data points over time (3 baseline measures, 3 taken during testing (during MIST, post-MIST and N-back), and 3 post-testing). Data were analysed using repeated-measures ANOVAs with time (9 levels) as the repeated-factor and condition (stress or control) as the between-subjects factor, with LSD-posthoc tests. The relationship between physiological and psychological variables during the MIST was explored using partial correlations partialling for condition. Analysis of skin conductance, heart rate variability and N-back data is ongoing. Salivary-cortisol levels were increased in the stress group, compared to controls (F1,22=7.47; p=0.012, ηp2=0.254), until 80 minutes after commencing the MIST. Stress-related increases in pulse and blood pressure measures were seen during administration of the MIST, reflected as interactive effects of time and condition (systolic: F8,160=2.056; p=0.043; diastolic: F8,160=2.713; p=0.016; pulse [trend level only]: F8,160=2.02; p=0.092). Subjective stress ratings were greater in the stress group, compared with controls, during (F1,21=6.4; p=0.020) and immediately after (F1,21=5.5; p=0.029) the MIST. However, there were no correlations between subjective stress ratings (obtained during the MIST) and any of the concurrently- measured physiological variables (p’s>0.16; except for pulse rate: p=0.069). In summary, administration of a laboratory stressor led to robust increases in physiological and psychological stress measures. However, there was little evidence that these two types of ratings correlated with one another. Our on-going studies are investigating the effect of the MIST on working memory performance and emotional processing. We are also comparing the physiological and psychological effects of the MIST with those of a newly-designed virtual reality environment stressor. Ethical approval for this study was granted by the Central Univ Research Ethics Committee of the Univ of Oxford. This research was supported by a grant awarded by the MRC.

F03 EVIDENCE FOR A COMPUTATIONAL ACCOUNT OF STRIATAL DOPAMINE 2 RECEPTOR FUNCTION Adams RA, Inst of Cognitive Neuroscience & Div of Psychiatry, UCL, 6th Floor Maple House, 149 Tottenham Court Road, London W1T 7NF [email protected] Roiser JP(1), Illingworth BJG(2), Nour M(3), Dahoun T(3), Howes OD(3), Moutoussis M(4), Schwartenbeck P(4), Friston KJ(4) (1) Inst of Cognitive Neuroscience, UCL, London WC1N 3AZ; (2) Medical School, UCL, London WC1E 6BT; (3) MRC CSC, Imperial College, London W12 0NN; (4) Wellcome Trust Centre for Neuroimaging, UCL, London WC1N 3BG Introduction: Reinforcement learning (and other) models of decision-making propose that dopamine ABSTRACTS A113 signalling in the striatum encodes stimulus salience or reward (and aversive) prediction error, and motivation or average rate of reward. ‘Active inference’ is an alternative decision-making model based on the premise that the brain (and hence perception and action) maximises Bayesian model evidence – or minimises prediction error. Under active inference, dopamine is thought to encode the precision (inverse variance) of the agent’s beliefs that its policies will achieve its goals. This precision plays a similar role to the inverse temperature parameter in standard (softmax) decision models – i.e. it affects the stochasticity of decision-making – but, unlike inverse temperature, it is updated dynamically according to the context. In this way, when the agent is not confident that any policy will achieve its goal, its choice of a policy becomes more stochastic. Like striatal dopamine responses, the signalling of precision (of policies) has both tonic and phasic aspects: tonic signalling encodes prior (or posterior) beliefs about precision, and phasic signalling encodes updates to these beliefs. Tonic dopamine signalling in the striatum largely activates dopamine 2 receptors (D2Rs) as they are more sensitive to low dopamine concentrations. On this basis, we hypothesised that: i) Prior precision and an inverse temperature parameter (estimated using three cognitive tasks in each subject) would correlate within subjects; ii)Prior precision would correlate with D2R availability in the striatum. Methods: 20 healthy participants performed three cognitive tasks: a ‘draws to decision’ beads task, a Go NoGo task and the Limited Offer task. Participants’ prior precisions in the latter two tasks were estimated using active inference models, and an inverse temperature parameter was estimated from the former task. 19 participants also underwent [11C]PHNO positron emission tomography (PET) scanning to measure striatal D2R availability (BPnd). Results: Prior precisions correlated with each other (r=0.22, p=0.02) and one correlated with inverse temperature (r=0.27, p=0.01); the other did not (r=0.07, p=0.14) unless an outlier was removed. Prior precision (Go NoGo task) had a quadratic relationship with striatal BPnd (r2=0.34, p=0.04) in an ROI analysis, confirmed using voxel-wise non-linear regression. Prior precision (Limited Offer task) had a quadratic relationship with striatal BPnd (r2=0.50, p=0.009) in an ROI analysis, confirmed using voxel-wise non-linear regression. Conclusions: These results indicate precision (of policies) is a valid construct and that tonic dopaminergic signalling in the striatum may encode prior beliefs about this precision. This study was funded by an Academy of Medical Sciences Starter Grant for Clinical Lecturers and a UCLH Biomedical Research Centre Neurosciences Grant to Dr Rick Adams, and Medical Research Council-UK (no. MC-A656-5QD30), Maudsley Charity (no. 666), Brain and Behavior Research Foundation, and Wellcome Trust (no. 094849/Z/10/Z) grants to Dr Howes and the National Institute for Health Research (NIHR) Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London.

F04 MICE HAPLOINSUFFICIENT FOR MAP2K7, A SCHIZOPHRENIA RISK GENE, SHOW ALTERED ACQUISITION AND PERFORMANCE ON A RODENT GAMBLING TASK Openshaw RL, Inst of Neuroscience and Psychology, Univ of Glasgow, 306c West Medical Building, Univ Avenue, Glasgow G12 8QQ [email protected] Morris BJ(1), Thomson DM(2), Pratt JA(2) (1) As presenting author; (2) Strathclyde Inst of Pharmacy and Biomedical Sciences, Glasgow Developing effective therapies to treat psychiatric disorders relies on animal models with appropriate face and construct validity. We investigated mice haploinsufficient for Map2k7 (Map2k7+/-), a gene functionally associated with schizophrenia (Winchester et al., 2012, Human Molecular Genetics, 21:4910–4921), in a rodent gambling task (rGT) that closely resembles the human Iowa Gambling Task (IGT). In the rGT, mice choose between four options that differ in their magnitude and ratios of reward/punishment possibilities. Options associated with large rewards also lead to disproportionately greater punishment than those associated with smaller wins. Therefore, to obtain optimum amounts of reward in the time available, mice should select smaller, frequent rewards rather than large, infrequent rewards. Impairment in decision making is frequently observed in the IGT in patients with schizophrenia, where they make more disadvantageous choices (Sevy et al., 2007, Schizophrenia Research, 92(1–3):74–84). Mice (12 Map2k7(+/-)(6 male), 12 WT (6 male)) were trained on the rGT using touchscreen apparatus (Campden Instruments) until they reached stable performance. Data A114 ABSTRACTS were analysed by repeated measures ANOVA. Map2k7+/- mice showed enhanced accuracy during task acquisition (F(1,88)=96.42; p<0.05) and in the number of trials to reach criteria (F(1,22)=6.07; p<0.05). Following task acquisition, Map2k7+/- mice displayed altered preference for the “high-risk, high-reward” option (F(1,100)=39.38; p<0.001), and in % optimal choices made on average (which is the % choice of the two options that give smaller rewards and shorter, infrequent punishments) (F(1,100)=22.96; p<0.001). Additionally, Map2k7+/- mice showed hyperactivity in latency measures: reward collection latency (F(1,100)=9.32; p<0.05), correct response latency (F(1,100)=40.13; p<0.001) and in the number of beam breaks made during each daily session (F(1,100)=94.33; p<0.001). Overall, Map2k7+/- mice display alterations in gambling task acquisition and performance and show general hyperactivity. These results suggest MAP2K7 and/or other signalling components in its pathway have a role in decision making. We previously reported that Map2k7+/- mice show hyperactivity in the open field, and an attentional deficit in the 5-CSRTT. These data support and extend these findings, showing promise for Map2k7+/- mice being a translatable model of cognitive dysfunction in psychiatric disorders, with potential for aiding future development of improved therapies. This work is financially supported by the Medical Research Council and we thank Prof. Penninger (IMBA, Vienna, Austria) for the gift of the mice. All work was conducted under the Animals (Scientific Procedures) Act 1986.

F05 PARSING THE DIFFERENTIAL CONTRIBUTIONS OF PRIMATE ORBITOFRONTAL CORTEX AND PERIGENUAL ANTERIOR CINGULATE CORTEX TO CONTINGENCY LEARNING Horst NK, Dept of Psychology; Behavioural and Clinical Neuroscience Inst, Univ of Cambridge, Downing Street, Cambridge CB2 3EB [email protected] Jackson SAW(1), Horiguchi N(1), Robbins TW(1), Roberts AC(2), Cardinal RN(3) (1) As presenting author; (2) Dept of Physiology, Development and Neuroscience, Univ of Cambridge, Downing Street, Cambridge, CB2 3DY; Behavioural and Clinical Neuroscience Inst, Downing Street, Univ of Cambridge, CB2 3EB; (3) Dept of Psychiatry, Univ of Cambridge, Herchel Smith Buidling for Brain & Mind Sciences, Forvie Site, Robinson Way, Cambridge CB2 0SZ; Behavioural and Clinical Neuroscience Inst, Downing Street, Univ of Cambridge, CB2 3EB The control of instrumental learning, whereby organisms come to link actions with their outcomes, is supported by both a stimulus-response ‘habit’ system and a goal-directed action-outcome mechanism (Dickinson, 1985, Philos Trans R Soc Lond, 308:67–78). Disruption to the balance between these systems is suggested to be responsible for the pathology of obsessive-compulsive disorder (OCD, Gillan & Robbins, 2014, Philos Trans R Soc B, 369:1–11), a neuropsychiatric illness involving dysfunction in fronto- striatal circuitry. Goal-directed actions have been shown to be dependent upon contingency, defined as the difference between the probability of reinforcement given a response, and the probability of reinforcement in the absence of that response (Hammond, 1980, J Exp Anal Behav, 34:297–304). We have previously reported that marmoset monkeys with excitotoxic lesions of either orbitofrontal cortex (OFC) or perigenual anterior cingulate cortex (pgACC) fail to detect when one of two trained action-outcome contingencies is degraded by the non-contingent delivery of one of the potential outcomes (Jackson et al., Cerebral Cortex, in press). We sought to extend these findings using localised reversible inactivations in marmosets. Adult common marmosets (Callithrix jacchus; males = 2, females = 2) were trained to make increasing numbers of responses to a stimulus on the left side of a touchscreen for delivery of reward A or, in separate sessions, to the same stimulus on the right side of the screen for reward B. Once marmosets reliably made ~10 responses per reward, they were implanted with chronic indwelling cannulae in the OFC (n = 2), the pgACC (n = 1), or both (n = 1). The GABA-A/GABA-B agonists muscimol/baclofen were infused to reversibly inactivate either the OFC or pgACC during contingency degradation probe sessions. In the ‘degraded’ probe session, the reward for which the marmoset was working was also delivered non-contingently on an independent schedule, thereby degrading the association between action and outcome. In the ‘non-degraded’ session, the non-contingent reward differed from the contingent reward, thus leaving the action-outcome contingency intact. Under control conditions (n=5), subjects reduced responding more in ‘degraded’ sessions vs. ‘non-degraded’ sessions [F(1,4)=39.1, p<0.01]. Preliminary OFC ABSTRACTS A115 inactivation data (n=3) indicate a near-significant trend toward diminished impact of non-contingent reward on responding overall. Inactivation of pgACC (n=2), by contrast, appeared to specifically disrupt the action-outcome association. By using a novel contingency degradation paradigm that allows multiple acute pharmacological manipulations, we have revealed differences in the nature of the contribution of the OFC and pgACC to the use of action-outcome contingency information in response selection. These results have broader implications for dysregulation of instrumental learning in OCD. This study was funded by a Wellcome Trust Senior Investigator Award104631/Z/14/Z (to TWR) and conducted within the Univ of Cambridge Behavioural and Clinical Neuroscience Inst, supported by a joint award from the MRC and the Wellcome Trust. SAWJ was supported by a BCNI-MRC studentship.This study was funded by a Wellcome Trust Senior Investigator Award104631/Z/14/Z (to TWR) and conducted within the Univ of Cambridge Behavioural and Clinical Neuroscience Inst, supported by a joint award from the MRC and the Wellcome Trust. SAWJ was supported by a BCNI-MRC studentship.

F06 OSCILLATIONS IN COGNITION: ATTENTION-PROMOTING DOSE OF METHYLPHENIDATE AUGMENTS ALPHA RHYTHM IN THE DORSAL ATTENTION NETWORK Hayward A, Manchester Pharmacy School, Univ of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT [email protected] Gigg J(1), Neill JC(2) (1) Faculty of Life Sciences, Stopford Building, Oxford Road, Univ of Manchester, Manchester, M13 9PT, UK; (2) Manchester Pharmacy School, Stopford Building, Oxford Road, Univ of Manchester, Manchester, M13 9PT, UK Introduction: Methylphenidate (Ritalin) is the first line treatment for attention deficit hyperactivity disorder. However, the abuse liability of methylphenidate limits its clinical use. To develop new improved treatments, a comprehensive understanding of its mechanism of action to enhance attention is critical. We have shown previously that methylphenidate (1 mg/kg) improves attention in low attentive rats (Tomlinson et al.2014.European Neuropsych;24:1371–1380). Considering the putative neural target for its pro-attentive effects, methylphenidate increases fMRI BOLD signal in the dorsal attention network, especially the medial prefrontal and posterior parietal cortices (mPFC and PPC; Tomasi et al.2011.Neuroimage;54:3101–10).Our aim here is to investigate further the mechanism of action of methylphenidate by measuring oscillations of extracellular local field potentials (LFP/focal EEG) in mPFC and PPC, regions known to control attentional processing. Methods: Multi-electrode arrays recorded LFP signals from mPFC and PPC of urethane-anaesthetised (1.3g/kg i.p.) adult female Lister Hooded rats (n=5). Each animal received a saline injection (1ml/kg, i.p.) followed by LFP recording for 30min. Animals then received methylphenidate (1mg/kg, i.p.) and LFPs were recorded for a further 60min. Methylphenidate and saline LFPs were analysed for power spectral density (PSD) and methylphenidate recordings were normalised to the average PSD in the saline recording. Resultant relative PSDs were averaged for the EEG bands delta (1–4 Hz), theta (4–8 Hz), alpha (8–15 Hz), beta (15–30 Hz), low gamma (30–50 Hz) and high gamma (50–90 Hz). Data were analysed using two-way ANOVA with animal ID as a blocking factor to produce a pseudo repeated measures design, LSD planned comparisons were used to compare time points. Results: There was a significant suppression of delta (p<0.001) and increase in alpha (p<0.05) for methylphenidate PSDs in both mPFC and PPC. A significant reduction in theta was specific to the mPFC (p<0.01). Power in beta and gamma bands was generally higher in mPFC under methylphenidate challenge but these increases did not reach statistical significance. Conclusion: The main effect of methylphenidate in mPFC and PPC on PSD is to promote an increase in alpha whilst attenuating lower frequency bands. Alpha power is linked to performance in tasks of attention, particularly selective attention. Therefore, the changes seen here may represent the neural basis for the pro-attentive effects of methylphenidate. These observations support the hypothesis that a major action of methylphenidate is to promote a shift in EEG state towards alpha rhythm within the mPFC and PPC, thereby increasing selective attention by enhancing stimulus selection. Declaration of interest: Jo Neill has received expenses to attend conferences and fees for lecturing and consultancy work (including attending advisory boards) from the manufacturers of various psychiatric drugs. Funding Source: The work was funded by B-Neuro and the Univ of Manchester Pharmacy School. A116 ABSTRACTS

F07 ALTERED D2/3 RECEPTOR AVAILABILITY AND FUNCTIONAL CONNECTIVITY IN RATS SELECTED FOR EXTREME HIGH AND LOW CHOICE IMPULSIVE BEHAVIOUR Barlow RL, CNS Diseases, Boehringer Ingelheim Pharma GmbH & Co. KG, Div. Research, Birkendorfer Strasse 65, Biberach an der Riss, Germany 88397 [email protected] Wearn A(1), Gorodetskaya N(1), Stiller D(1), Nicholson J(1), Pekcec A(1), Gorges M(2), Kassubek J(2) (1) As presenting author; (2) Dept of Neurology, Univ of Ulm, RKU, Oberer Eselsberg 45, D-89081 Ulm, Germany Impulsivity is a multi-faceted behavioural construct which can broadly be defined as a tendency to act prematurely and without foresight. It is symptomatic of several neuropsychiatric disorders including schizophrenia, ADHD and substance abuse disorders, and can be observed behaviourally as impaired response inhibition (motor impulsivity) or the preferential choice of risky or immediate rewards (choice impulsivity). Maladaptive impulsivity may arise due to dysfunctional interactions between the prefrontal cortex and ventral striatum. Previous work has shown that rats expressing high levels of motor impulsivity have reduced dopamine D2/3 receptor availability in the ventral striatum, as compared to low impulsive rats (Dalley et al., 2007. Science 315: 1267–1270). Choice impulsivity was assessed in a cohort of adult male Lister-hooded rats (n=96; 300–350g), using the delay discounting task. Rats were ranked for their level of impulsivity based on their preference for a small immediate reward over a larger, delayed reward. In this study, we investigated whether high levels of choice impulsivity were also associated with decreased D2/3 receptor availability, using [18F]fallypride PET imaging. Ex vivo receptor autoradiography with [3H]raclopride was then used to further investigate possible sub-region specific differences in D2/3 receptor availability. Additionally, we assessed the functional connectivity between pre-defined regions of interest, using resting-state MRI. A negative correlation was observed between impulsive choice and [18F]fallypride binding potential in both the left and right hemisphere of the ventral striatum in high (n=11) but not low (n=10) impulsive rats (rs=0.79, p=0.01; rs=0.71, p=0.03). [3H]Raclopride receptor autoradiography then revealed a dissociation in D2/3 binding between the core and shell subregions of the nucleus accumbens; with increased binding observed in the shell (F= 6.132, p=0.02), and decreased binding in the core region of high impulsive rats (F= 6.689, p=0.018). Additionally, increased functional connectivity was observed between the cingulate cortex and the striatum of high impulsive rats (t=3.0; p=0.007), although no significant differences in connectivity were observed between the orbitofrontal cortex and striatum. These results indicate that increased connectivity between the cingulate cortex and striatum may drive increased levels of choice impulsivity. Additionally, D2/3 receptor availability in the ventral striatum, specifically at the levels of the nucleus accumbens core, appears to influence trait-like choice impulsivity, potentially as a consequence of altered dopaminergic tone. Complete Financial Sponsorship by Boehringer Ingelheim Pharma GmbH & Co. KG, Div. Research Germany, Birkendorfer Strasse 65, 88397, Biberach an der Riss, Germany.

G01 TO GO OR NOT TO GO: BRAIN TRAINING IN BULIMIA NERVOSA Turton RT, Section of Eating Disorders, Dept of Psychological Medicine,. King’s College London, IoPPN, 103 Denmark Hill, Camberwell, London SE5 8AF [email protected] Nazar BN(1), Burgess EB(1), Cardi VC(1), Hirsch* Joint last author CH(1), Treasure* Joint last author JT(1), Lawrence NL(2) (1) Dept of Psychological Medicine, King�s College London, Inst of Psychiatry, London; (2) Washington Singer Labs, Univ of Exeter, Prince of Wales Rd, Exeter Introduction: Targeting impulsive food behaviours through strengthening inhibitory processes may be a beneficial approach in obesity, Bulimia Nervosa (BN), and binge eating disorder. Food-specific inhibition training involves the use of novel computerised go/no go paradigms to help train individuals to inhibit their automatic impulses towards highly palatable foods. However, this approach has not been tested in clinical populations to date. The aim of this study was to examine whether a single ABSTRACTS A117 session of food-specific go/ no go training can help to reduce the consumption of binge-foods in BN. Methods: 15 participants with BN and 25 Healthy Controls (HCs) have taken part. Participants are given one session of food-specific go/no go training and one control session, using a within-subjects AB/ BA crossover design. The primary outcome measure is food consumption on a taste test following the training. The study was approved by a local ethics committee (reference: 14/L0/2166). Results: For HCs, there were no significant differences in food consumption on the taste test between the experimental and control conditions (p > .05). In BN, a Wilcoxon signed-rank test showed that chocolate consumption was significantly lower in the experimental (Mdn = 44kcal) versus control condition (Mdn = 143kcal), z = -2.073, p = .038, r = .38. Conclusions: Preliminary results suggest that go/no-go training is helpful for BN in reducing the consumption of chocolate, therefore; go/ no-go training could be useful as an adjunct treatment enhancer to help increase control over binge-foods. However, further studies with more sessions of training and longer-term follow-ups are needed. Acknowledgement: This research is part funded by the Medical Research Council and the Psychiatry Research Trust.

G02 THE ACCEPTABILITY AND EFFECTIVENESS OF ONLINE FOOD RESPONSE INHIBITION TRAINING FOR WEIGHT LOSS IN A LARGE PUBLIC SAMPLE Lawrence NS, Psychology, Univ of Exeter, Perry Road, Exeter EX4 4QG [email protected] Javaid M(1), Verbruggen F(1), Chambers CD(2), VanBeurden S(3), Greaves CJ(3) (1) As presenting author; (2) Psychology, Cardiff Univ, Park Place, Cardiff CF10 3AT; (3) Univ of Exeter Medical School, Magdalen Road, Exeter EX1 2LU Over half of UK adults (64%) are overweight or obese, resulting in 30,000 deaths and over £5 billion in costs to the NHS per year. Over-eating is a key contributor to obesity. Research suggests that computerised food response inhibition training, which requires people to repeatedly inhibit motor responses (keyboard presses) to food pictures, reduces food intake, decreases food liking and facilitates weight loss (Lawrence et al., 2015, Appetite, 95, 17–28). Furthermore this low-cost web-based intervention is easy to complete and highly acceptable to participants. Here we conducted an open pragmatic study of food response inhibition training delivered online and in an unsupported manner to the general public to further examine its acceptability and effectiveness in reducing food intake and weight. We also conducted a thematic analysis of comments about subjective training effects. Online recruitment and data collection took place over 6 months (July – December 2015) following media dissemination of our pilot trial that included a link to the study website (http://foodtraining.exeter. ac.uk/). The website included a statement of informed consent, brief instructions, and a link to an (optional) online pre-training survey to measure baseline variables. Participants were encouraged to complete four daily 10-minute training sessions, consisting of a go/no-go task in which energy-dense foods were always paired with a no-go signal. They were emailed a post-training survey 4–6 weeks after completing pre-training. The pre- and post-training surveys measured self-reported weight and frequency of intake of energy-dense snack foods over the past month. Of the 18,725 participants who tried the training, 7763 provided pre-training measures and 933 also provided post-training measures. Comparison of valid (non-missing) data at pre- and post-intervention suggested significant weight loss (M = -0.91 kg, SD = 2.56; t(793) = 10.01, p<.001, Cohen’s dz=0.35) and reduced snacking frequency (M = -19.78%, SD 28.8; t(817) = 19.63, p<.001, Cohen’s dz=0.68). There was a weak association between training ‘dose’ and the reduction in snacking frequency (r(817)=0.11, p=.002). This study suggests a high public demand for this online weight-loss intervention, although recruiting and retaining people into an unsupported study is challenging. Study participants reported reduced food intake and weight, with the majority reporting that the training may have helped them to reduce their intake (74%) and that they would/might recommend the intervention to a friend (82%). These promising findings support our earlier pilot trial and await validation in large-scale RCTs with objective measures. This work was supported by the Univ of Exeter. A118 ABSTRACTS

G03 COPING VS. BINGE-EATING OR FOOD ADDICTION PREDICTS ATTRITION FROM HIGH INTENSITY WEIGHT-LOSS TREATMENT Burgess EE, Psychology, Univ of Alabama at Birmingham, CH 201 1300 Univ Blvd | Birmingham, AL 35294–1170 USA, 35294-1170 [email protected] Sylvester MD(1), Boggiano MM(1), Soleymani T(2), Daniel S(2) (1) As presenting author; (2) Dept of Nutrition Sciences Introduction: More than 50% of patients drop-out from weight loss treatment. The identification of factors that put an individual at high risk of attrition is important to inform strategies aimed at increasing retention rates in treatment. Methods: N=298 participants (F=242, M=56) pursuing one of three weight-loss treatments were recruited. Programs included a high intensity low calorie diet (Optifast©), low intensity lifestyle modification program (Risk Reduction: RR), and a high intensity low carb diet. The following baseline variables were tested for ability to predict drop-out from treatment at 3 months: psychological eating motives (Social, Coping, Enhancement, Reward), binge-eating symptoms, food addiction symptoms, and program type. Data was analysed using a logistic regression, while controlling for age, sex, and ethnicity. Results: Due to program type significantly increasing the odds of drop-out from treatment (p<0.001), a separate logistic regression was tested for each program type. Patients with high binge-eating symptoms were nearly 10% more likely to drop-out from low intensity lifestyle modification (N=166, p<0.05). Patients with high motivation to use food as a means to cope were nearly 4.5X more likely to drop out of a high intensity, low calorie, diet intervention (N=67, p<0.05). Similarly, patients with high coping motives were 6X as likely to drop-out from high intensity, low carb, intervention (N=65, p<0.05). Conclusions: Binge-eating only seems to pose risk for treatment drop-out when not appropriately addressed by a high intensity intervention plan. High motivation to use food as a means to cope can pose a barrier to treatment for individuals in high intensity intervention programs and should be addressed as part of the behavioural treatment program. This study was not supported by any financial funding.

G04 POTENTIAL NEW TREATMENT APPROACHES FOR BINGE EATING DISORDER (BED): SUPPRESSIVE EFFECTS OF TRANSCRANIAL DIRECT CURRENT STIMULATION (TDCS) ON FOOD INTAKE AND FOOD CRAVING IN BED Burgess EE, Psychology, Univ of Alabama at Birmingham, UAB | The Univ of Alabama at Birmingham Dept of Psychology CH 201 or 415| 1300 Univ Blvd | Birmingham, AL 35294-1170 USA, 35294 [email protected] Sylvester MD(1), Morse KE(1), Amthor FR(1), Mrug S(1), Lokken KL(1), Boggiano MM(1), Soleymani T(2) (1) As presenting author; (2) Dept. of Nutrition Sciences Introduction: This study investigated the effects of a single administration of real and sham transcranial direct current stimulation (tDCS) on food craving, in-lab food intake, and desire to binge-eat in individuals with binge eating disorder (BED) and sub-threshold BED (sub-BED). Potential mechanisms of suppressive effects on food craving and food intake were also explored. Methods: 30 participants (F=20, M=10) with BED (N=19) or sub-BED (N=11) were recruited. Participants were administered both a sham and a real session of tDCS, where current intensity was 2 mA for twenty minutes. The anode was placed over the right dorsolateral prefrontal cortex (DLPFC) and cathode over the left DLPFC. Total calories eaten from three foods offered in-lab (Oreo cookies, M&M’s, and potato chips) were calculated. Food cravings in response to food images were assessed pre-post tDCS (Likert scale), and participants’ rated their desire to binge-eat (Likert scale) later that evening. This was a repeated measures design, where all analyses report within subject differences across the real vs. sham session. Results: tDCS significantly decreased food intake (F(1, 29) = 4.35, p = 0.046; partial eta squared = 0.13) with the greatest suppression on the preferred food (F(1,29) = 5.35, p = 0.03, partial eta squared = 0.16). Males had greater reduction in overall food craving (F(1,28) = 4.09, p = 0.05, partial eta squared= 0.13), dessert craving (F(1,28) = 4.99, p = 0.03, partial eta squared = 0.15), and marginally greater reduction for carbohydrate craving ABSTRACTS A119

(F(1,28) = 3.79, p = 0.06, partial eta squared = 0.12). tDCS decreased desire to binge-eat only in males (F(1,28) = 4.77, p = 0.04, partial eta squared = 0.19). Increased intent to restrict food intake significantly predicted greater reduction in food intake by tDCS (Beta = -0.39, p = 0.048, R2 = 0.19, Cohen’s f2 = 0.23). There was no correlation between suppression of food intake and suppression of food craving (p = n.s.). Conclusions: Multiple tDCS session-induced neuroplasticity may hold promise as a safe and longer- lasting treatment adjunct for BED, particularly for males with BED. Right DLPFC anode placement suggests enhanced cognitive control as a functional mechanism for the anti-binge effects of tDCS. This study was funded by a UAB Faculty Development Grant to MMB.

H01 ACUTE EFFECTS OF MELISSA OFFICINALIS ESSENTIAL OIL AS A TREATMENT FOR INSOMNIA IN THOSE WITH MILD COGNITIVE IMPAIRMENT AND DEMENTIA Alford C, Health and Social Sciences, UWE, Coldharbour Lane, Bristol BS16 1QY [email protected] Austin L(1) (1) Dept for Health,Univ of Bath, Bath BA2 7AY Introduction: Disturbed sleep affects older adults including those with dementia and mild cognitive impairment (MCI). Given the unwanted side effects noted with standard treatment hypnotics that can be a greater problem for older people e.g. falls, and reports of successful interventions with aromatherapy in dementia reducing restlessness (Ballard et al. 2002 J Clin Psychiatry 63, 553–558), two studies examined the effects of melissa essential oil on sleep. Method: Twelve older adults (mean age 63 years) and fourteen participants with mild to moderate dementia or MCI (mean age 74 years), all with mild to moderate insomnia, as screened by the Pittsburgh Sleep Quality Index, and including both females and males who took part in the two separate studies which employed single-blind, balanced crossover designs. Measures included wrist actiwatches (CamNtech(R)), Leeds Sleep Evaluation Questionnaires and for the older adult study, Line Analogue Rating Scales (LARS). Treatment was with 1 drop (0.29ml) of melissa oil administered to bed clothes following our established protocol, at home over 2 nights either self-administered or by the partner or carer, compared to massage base oil (placebo) and included a 2 day washout and a no treatment control. Results: In the older adult study, melissa oil significantly increased actigraphic sleep by 42 minutes (p=0.003), compared to the no treatment control. In the dementia study, there was a significant increase in actual sleep time of 45 minutes with melissa compared to the no treatment control (p=0.015) and 38 minutes with placebo (p=0.009), as well as a reduced wake time of 39 minutes compared to control. No subjective, next day residual impairment was found. Older adults reported that melissa oil significantly reduced drowsiness following treatment, (p=0.041), as measured by LARS, although the MCI and dementia patients were unable to successfully complete the subjective ratings. Conclusion: Melissa was well tolerated and found to be a safe and beneficial treatment for mild to moderate sleep disturbance for older adults, and those with dementia and mild cognitive impairment living in the community, following 2 sequential treatment nights. Future studies should investigate longer treatment periods to evaluate whether treatment effects are maintained and provide a practical improvement to sleep management in those with MCI or dementia. No financial sponsorship was received for this study.

H02 INVESTIGATING THE EFFECTS OF TYPHOID VACCINE ON SLEEP IN HEALTHY PARTICIPANTS: A RANDOMISED, DOUBLE- BLIND, PLACEBO-CONTROLLED, CROSS-OVER STUDY Sharpley AL, Dept of Psychiatry, Univ of Oxford, Neurosciences Building,Warneford Hospital, Oxford OX3 7JX [email protected] Cooper CM(1), Williams C(1), Godlewska BR(1), Cowen PJ(1) (1) As presenting author Introduction: An increasing body of evidence links the occurrence of sleep continuity disturbances with increased inflammation and both insomnia and chronic inflammatory illnesses are associated A120 ABSTRACTS with high rates of co morbid depression. Typhoid vaccination results in a mild inflammatory response that significantly increases levels of the pro-inflammatory cytokine, interleukin (IL)-6. Studying healthy participants enables investigation of the effects of inflammatory challenge in individuals who do not currently have an inflammatory condition, or alterations to their sleep patterns due to a sleep disorder or as part of a depressive condition. Methods: Sixteen healthy participants, (9 female, 7 male; mean age 26.6 years; range 20–38 years; BMI 23; range 18.8–26.4), were randomly allocated in a double-blind, placebo-controlled, cross-over design to receive typhoid vaccination and placebo injections 1–2 weeks apart. Each participant was given a single 0.5mL injection of either typhoid polysaccharide vaccine (Typhim Vi®) (Sanofi Pasteur MSD Limited) or placebo (0.9% sodium chloride saline solution) into the non-dominant deltoid muscle in the arm at 1500h and completed mood questionnaires hourly for 4 hours. They were fitted with sleep EEG equipment (Embla Titanium) and 2 hours post vaccination a blood sample was taken and the serum frozen for analysis of IL-6. Participants returned home to sleep and completed the Leeds Sleep Evaluation Questionnaire (LSEQ) the following morning. Sleep was analysed using the REMLogic software and manually edited. Results: The typhoid vaccine was well tolerated with mild side effects. IL-6 levels (pg/ml) were significantly increased two hours post typhoid vaccine (0.90 ± 0.61) compared to the saline injection (0.53 ± 0.15) (mean ± SD); Paired t test P≤0.026. Relative to placebo, typhoid vaccination produced several changes in sleep continuity parameters. Specifically, total sleep time (TST) (min) (P≤0.005) and sleep efficiency (SE) (%) (P≤0.007) were significantly decreased; consistent with this, there were significant increases in wake after sleep onset (WASO) (min) (P≤0.007), total wake (min) (P≤0.005), sleep stage transitions (P≤0.026), number of awakenings (P≤0.007) and awakening index (P≤0.005) following typhoid vaccine compared to the saline injection. All subjective state measures were unaffected by typhoid administration. Conclusions: Inflammatory mechanisms may underlie the impairment in sleep efficiency which is a hallmark of major depression. Because impaired sleep is also a predictor of major depression there may be a role for suitable anti-inflammatory approaches in strategies designed to prevent the onset of depression. This work was supported by Oxford Univ Deptal funds.

H03 SLEEP AND CIRCADIAN PHENOTYPES IN A COHORT OF PEOPLE WITH BIPOLAR DISORDER McAllister-Williams RH, Academic Psychiatry, Inst of Neuroscience, Newcastle Univ, Wolfson Research Centre, Campus for Aging and Vitality, Westgate Road, Newcastle upon Tyne NE4 5PL [email protected] Bradley AJ(1), Gallagher P(1), Anderson K(2) (1) As presenting author; (2) Dept of Neurology, Royal Victoria Infirmary, Queen Victoria Rd, Newcastle upon Tyne, NE1 4LP Introduction: Sleep disturbance is a common feature of all stages of bipolar disorder (BD) including euthymia. It is hypothesised to be associated with reduced function, quality of life and cognitive impairment and is therefore an important treatment target. However, there is a lack of data describing the nature of sleep in BD using objective and subjective measures of sleep variables and circadian function and objectively assessing participants for sleep apnoea. Methods: BD patients in any mood state (n=46) and age matched healthy controls (n=42) were assessed with 21 days of actigraphy, sleep diaries and mood and sleep questionnaires. Circadian rhythm was assessed by measuring urinary secretion of 6-sulphatoxymelatonin over two 48 hour periods. Sleep apnoea was assessed with one night of portable oximetry. Sleep variables were examined and used to identify participants sleep phenotype based on the International Classification of Sleep Disorders 3rd edition (ICSD-3) criteria. These phenotypes included short sleepers (< 6 hours sleep per night), long sleepers (> 10 hours sleep per 24 hour period), and circadian rhythm disorders including irregular sleep wake rhythm and delayed sleep phase. Results: BD participants had poorer self-reported sleep on the Pittsburgh Sleep Quality Index (p<0.001) and higher levels of daytime sleepiness on the Epworth Sleepiness Scale (P=0.004). Significantly more BD participants than controls (29.2% vs 10%) were identified with sleep apnoea (p=0.029). Actigraphy demonstrated that BD participants had more variable wake times (p=0.023) had longer (p=0.051), though more variable (p=0.001) mean nocturnal sleep time, and lower sleep efficiency (p=0.03) than controls. ABSTRACTS A121

There were no differences in the proportions of short sleepers in each group (BD 8.7%, controls 4.8%, p=0.678). BD patients had lower overall activity levels (p<0.001) and more BD patients spent > 10 hours in bed (BD 13%, controls 0%; p=0.027) and had > 10 hours sleep over each 24 hour period (BD 30.4%, controls 7.1%; p=0.006). A greater number of BD patients had a sleep pattern indicative of a circadian rhythm disorder (BD 30.4%, controls 9.5%; p=0.015). Conclusions: Patients with BD are less active and on average have longer and more variable sleep patterns than healthy controls. There is much variability in the sleep phenotypes within BD including both short and long sleepers and those with irregular sleep patterns resembling circadian rhythm disorders. Sleep apnoea was more common in the BD group which may have a significant influence on the sleep pattern and quality in those patients, as well as potential adverse effects on their physical and mental health. AJ Bradley’s PhD of which this study is a part is funded by Eli Lilly & Co Ltd.

H04 ANTIDEPRESSANTS FOR TREATMENT OF ‘PRIMARY’ INSOMNIA: RESULTS OF A COCHRANE SYSTEMATIC REVIEW Baldwin DS, Clincial and Experimental Sciences, Univ. of Southampton Faculty of Medicine, College Keep, 4–12 Terminus Terrace, Southampton SO14 3DT [email protected] Manson C(1), Wilson S(2), Lipinska G(3), Malizia A(4), Stuart B(5), Mayers A(6) (1) As presenting author; (2) Centre for Neuropsychopharmacology, Imperial College, London; (3) Dept. Psychology, Univ of Cape Town, Rondebosch 7701.; (4) Neurosurgery, North Bristol NHS Trust; (5) Primary Medical Care, Univ. of Southampton; (6) Univ. of Bournemouth, Poole BH12 5BB Introduction: Although insomnia is a common and burdensome condition there are no ideal hypnotic drugs. Concern about potential adverse effects with benzodiazepines has led many GPs to prescribe antidepressant drugs for relief of sleep disturbance, but they are not licensed for that indication. We aimed to ascertain the extent and level of evidence through performing a systematic review (and where possible meta-analysis). Method: We followed Cochrane systematic review procedures, searching OVID MEDLINE, EMBASE, PsycINFO and the Cochrane Register (CENTRAL). We included randomised controlled trials in adults with a primary diagnosis of insomnia in which an antidepressant as monotherapy was compared to placebo, insomnia medications (e.g. benzodiazepines or ‘Z’ drugs), another antidepressant, or ‘treatment as usual’. The primary outcome measure was subjective improvement in sleep, secondary outcomes including objective sleep measures, tolerability, daytime symptoms and functioning. Results: A total of 4245 studies were identified: 20 could be included. Three studies (total n=135) compared an SSRI with placebo: two paroxetine studies found significant improvements in subjective sleep measures at 6 (p=0.03) and 12 weeks (p<0.001), but no difference was seen in the fluoxetine study. One study (n=60) compared paroxetine with alprazolam and found a significantly greater improvement in subjective sleep with paroxetine. Three studies (total n=489) which compared an SSRI with another antidepressant (agomelatine, nefazodone) found no significant differences in sleep quality (SMD 0.97, 95% CI -0.91 to 2.85) though heterogeneity was high. Six studies (total n=812) compared a TCA (5 with doxepin, 1 with trimipramine) with placebo: data could be pooled from 4 studies (n=518) and showed a significantly greater improvement in subjective sleep quality over placebo (SMD -0.39, 95% CI -0.56 to -0.21). A single study compared doxepin with lormetazepam, and another compared doxepin with , but neither revealed significant differences. Six studies compared other antidepressants with placebo (1 mianserin; 5 trazodone): three trazodone studies (total n=370) provided extractable data indicating a greater improvement in subjective sleep outcomes for trazodone over placebo (SMD -0.34, 95% CI -0.66 to -0.02. Conclusions: The evidence base for antidepressants in primary insomnia is reliant on few and mostly small studies of short term administration, with substantial design limitations. Some data support the short-term use of some antidepressants, but there is currently no evidence to support the use of amitriptyline (despite its common use in clinical practice), or for long-term prescription of an antidepressant. Hence available evidence does not support common current practice. Funding: Partial funding through small grants from NIHR School of Primary Care Research and Primary Care and Population Sciences Unit at Univ of Southampton. A122 ABSTRACTS

GL1 PARSING THE SEROTONIN SYSTEM INTO DISTINCT FUNCTIONAL UNITS Gaspar P, Inst du Fer à Moulin Inserm UMRS-839, Univ Pierre et Marie Curie.17 rue du Fer à Moulin 75005, , France [email protected] Scotto-Lomassese S(1), Fernandez S(1) (1) As presenting author Serotonin signalling has been implicated in a multitude of functions in the brain and the periphery. While some of the functional specificity of serotonin’s actions can be imparted by a wide diversity of 5-HT receptors, it has become increasingly clear that the source of cells that produce or release 5-HT is equally important. In the brain all 5-HT is produced in the raphe neurons of the brain stem. These neurons share common neurochemical identities, but can be distinguished by their developmental trajectories, anatomical, molecular and physiological properties (Kiyasova et al. 2011, Gaspar and Lillesaar 2012). We will summarize recent data from our and different laboratories that lead to a more comprehensive parsing of the serotonin raphe cell groups (Fernandez et al.2015, Okaty et al. 2015, Muzerelle et al. 2016, Commons 2016). We will also show our work in progress indicating the role of guidance molecules to orient the dorsal and medial raphe neurons to different targets and how these different raphe subnuclei are involved in the control of distinctive functions such as learning or the onset of maternal behaviour. Funders : INSERM, FP7, FRM, ANR, Labex BIOPSY, ENP.

PD01 TOWARDS A BETTER UNDERSTANDING OF THE GENETIC BASIS OF PSYCHIATRIC DISORDERS: DEVELOPMENT OF NEW TRANSLATIONAL MODELS OF BEHAVIOUR Thomson DM, SIPBS, Univ of Strathclyde, 161 Cathedral Street, Glasgow G4 0RE [email protected] Winchester CL(1), Pratt JA(1), Dawson N(2), Hughes ZA(3), Dunlop J(3), Brandon NJ(3), McVie A(4), Morris BJ(4) (1) As presenting author; (2) Div of Biomedical and Life Sciences, Lancaster Univ, Lancaster; (3) Pfizer Global Research and Development, Cambridge, MA 02139; (4) Univ of Glasgow, Glasgow, UK Introduction: Understanding the neurobiological impact of genetic risk factors for neuropsychiatric disorders is necessary for the development of new and improved treatments. In particular, assessment of behavioural domains underpinned by analogous neural substrates to clinically relevant behavioural constructs, offer improved opportunities for translational drug discovery (Pratt et.al, 2012, Nature Rev Drug Discovery, 11-7, 560.). Development of a genetically manipulated (Disc1Tr hemi) mouse model which expresses a truncated form of the Disc1 gene linked to human psychiatric disease has allowed for neurobiological assessment of the effects of the mutation. We have previously shown that Disc1Tr hemi mice exhibit prefrontal cortex hypometabolism and disrupted hippocampal-PFC connectivity (Dawson et al, 2015, Transl Psychiatry, 5, e569), suggestive of cognitive dysfunction. Here we sought to examine a range of cognitive domains in these mice using a widely used task of attention and a novel task of spatial working memory. Methods: To this end we employed the 5-choice serial reaction time task in order to measure attentional processing where we manipulated response withholding and attentional loads. In addition we used a variant of the n-back task in the 8-arm radial maze to examine working memory function under normal conditions and those of NMDA receptor antagonism. Results: The 5-Choice serial reaction time task ANOVA revealed impaired inhibitory control in the Disc1Tr mice under conditions of increased inhibitory load (P<0.05), reflected as increased levels of anticipatory errors. Under normal task conditions and where we manipulated the attentional load no deficits in attention or motivation were observed. In the radial arm maze performance of the wild type and Disc1Tr mice was not different; however we observed a differential effect of ketamine (10mg/kg) on working memory performance. Wild type mice were impaired in the first two levels of the radial maze task (P<0.05 Fisher post-hoc), whereas Disc1Tr hemi mice failed to show reduced performance levels, suggesting a reduced impact of ketamine on working memory function. Conclusion: This series of experiments has demonstrated that for the 5-choice serial reaction time task, conditions of increased response withholding load are necessary to reveal subtle phenotypic differences in the Disc1Tr hemi mouse. In conjunction with this finding we suggest ABSTRACTS A123 that NMDA antagonism in the n-back radial arm maze task may have utility in examining subtle working memory differences in genetically modified mouse models with otherwise normal response profiles. This work was supported by an award (NS-GU-170) from the Translational Medicine Research Collaboration.

PD02 AUGMENTATION OF HUMAN NEOCORTICAL GAMMA OSCILLATIONS WITH A NOVEL KV3 POSITIVE MODULATOR IN VITRO Modebadze T, Inst of Neuroscience, Newcastle Univ, The Medical School, Newcastle upon Tyne NE2 4HH [email protected] Gillougley C(1), LeBeau FE(1), Cunningham MO(1), Alvaro GS(2), Large CH(2) (1) As presenting author; (2) Autifony Therapeutics Ltd, Medicines Research Centre, Via Fleming 4, 37135 Verona, Italy The failure of translation of biological effects from preclinical animal models to humans is a major barrier to the development of new and effective therapies for CNS disorders, particularly in psychiatry. One of the possible reasons for this failure is the complexity of human cortical microcircuits that are likely to exhibit different physiology and pharmacology to rodent neuronal circuits. Performing research in rodent systems has significant limitations and to reduce the risk of failure in the clinic it would be highly preferable to perform basic research in adult human brain tissue to eliminate species difference confounds and validate the efficacy of medicines in assays that are directly derived from the target organ. Synchronous gamma frequency oscillations (30–80Hz) are critical for processing and integrating cognitive modalities and have been shown to be impaired in patients suffering from schizophrenia. Kv3 channels are selectively expressed in neocortical PV+ interneurons which fire accurately at high frequencies to orchestrate rhythmic activity of neocortical networks. Targeting Kv3 channels, and enhancing the activity of PV+ interneurons, has potential as a pharmacological treatment for patients suffering from schizophrenia. We have examined the effect of a novel Kv3 modulator (AUT00206) in slices of resected human neocortical tissue. Additionally, we tested the effect of AUT00206 in slices from rats treated subchronically with psychotomimetic agent phencyclidine (PCP). Persistent gamma frequency oscillations were elicited by the bath application of kainate and carbachol. In order to acutely model network alterations associated with schizophrenia we tested the impact of AUT00206 on human cortical gamma activity in the presence of PCP. In these slices, AUT00206 (10 μM) produced a significant increase (50%) in gamma oscillation power (n=6 slices). Additionally, in frontal cortical slices obtained from a patient suffering from chronic schizophrenia, AUT00206 alone was able to significantly increase (47%) the power of pharmacologically induced oscillations (n=3 slices). Similar effect was observed in slices of prefrontal cortex from subchronically PCP treated rats where AUT00206 augmented gamma power by 30%. These results demonstrate, for the first time using ex vivo live human brain tissue, that AUT00206 significantly increases the power of human neocortical network oscillations in slices pre-exposed to PCP and slices from a patient with chronic schizophrenia. The findings are supported by studies in the animal model proving this twin-track approach to be highly beneficial. We show that modulation of Kv3 channels may have the potential to correct disruptions in neuronal synchronization in schizophrenic patients.

PD03 MATERNAL SEPARATION INCREASES SUSCEPTIBILITY TO ACUTE STRESS-INDUCED NEGATIVE AFFECTIVE BIAS IN ADULT RATS Stuart SA, Physiology, Pharmacology and Neuroscience, Univ of Bristol, Biomedical Sciences Bldg, Univ Walk, Bristol BS8 1TD [email protected] Robinson ESJ(1) (1) As presenting author Cognitive mechanisms are thought to underlie the development and perpetuation of depressive states (Harmer et al., 2009, Br. J. Psychiatry, 195:102–108; Roiser et al., 2011, Neuropsychopharmacol., 37:117– 136), and our previous work in rodents indicates that stress-induced negative affective bias may be a contributory factor (Stuart et al, 2013, Neuropsychopharmacol., 38:1625–1635). We now hypothesise that A124 ABSTRACTS early life stress, a risk factor for depression that increases stress-responsivity in adulthood, may enhance susceptibility to these biases. In this study we investigate the effects of maternal separation on both negative and positive affective biases in adulthood using the rodent affective bias test (ABT). Sprague Dawley rat pups were separated from their mothers for 180mins/day from day 1–14 (MS180), whilst control pups remained unhandled. 16 males (300–350g, n=8/group) underwent ABT training from day 60. The ABT uses a bowl-digging task where rats encounter two independent positive experiences (finding food reward in a specific digging substrate). Treatment or control is administered prior to the experience, and the absolute reward value is kept consistent across all sessions. Affective bias is quantified in a preference test where both rewarded substrates are presented together and the rats’ choices recorded over 30 randomly reinforced trials. Animals underwent ABT pairing sessions where they received either corticosterone (0.0–10mg/kg, s.c.) versus vehicle to induce a negative affective bias, or the antidepressant venlafaxine (0.0–10mg/kg, i.p.) to induce a positive bias. The effect of maternal separation on processing of absolute reward value was tested using a 2 pellet vs. 1 pellet discrimination study, and hedonic effects were assessed in a sucrose preference test. Corticosterone induced a dose-dependent negative affective bias in both groups (2-way ANOVA, DRUG: F2,28=22.8, p<0.001, GROUP: F1,14=10.7, p=0.006). However, at the lower dose, MS180 rats showed a significantly greater bias than controls (unpaired ttest: t14=2.95, p=0.01). Venlafaxine induced a dose-dependent positive affective bias with no group difference (DRUG: F2,28=16.6, p<0.0001, GROUP: F1,14=0.21, p=0.66). While control rats showed a significant positive bias for an increase in absolute reward value, MS180 rats did not (unpaired ttest: t14=5.4, p<0.0001). There was no group difference in sucrose preference. These findings suggest that maternal separation in rats increases susceptibility to negative affective biases, and impairs animals’ ability to learn different values of reward, independent of a consummatory anhedonic phenotype. However, sensitivity to antidepressant-induced positive biases remains unaffected. These findings may reveal a possible cognitive mechanism for early life stress as a risk factor for adult depression. SAS is supported by the BBSRC.

PD04 REFINING TECHNIQUES IN ATTENTIONAL SET-SHIFTING Tait DS, School of Psychology and Neuroscience, Univ of St Andrews, St Mary’s Quad, South St, St Andrews Fife KY16 9JP [email protected] Brown VJ(1) (1) Univ of St Andrews, School of Psychology and Neuroscience, St Mary’s Quad, South St, St Andrews, Fife, KY16 9JP Introduction: Intradimensional/Extradimensional (ID/ED) attentional set-shifting tasks (ASST) are a well- established method of assessing executive function in humans, monkeys, and rodents. Performance at the ED stage, which challenges subjects to shift to a previously irrelevant aspect of a stimulus, is affected by frontal cortex integrity, with impairments arising in human disorders such as schizophrenia and Parkinson’s disease, and rodents with induced medial prefrontal cortex (mPFC) dysfunction. Yet not all forms of ‘cognitive inflexibility’ are the same: similar-appearing deficits in ED shift performance yma be mediated by different psychological mechanisms, dependent on a specific disorder’s pathology. To assess those underlying processes requires refinement of the ASST, with additional or altered stages designed to test specific hypotheses – analysing trials to criterion, and errors where appropriate. Methods: We have previously modified the rodent ASST to investigate set-shifting deficits after mPFC manipulation (using a ‘semi-novel’ ED shift – where a previously irrelevant stimulus becomes rewarded – and therefore attention to that stimulus when it was designated irrelevant should facilitate the ED shift), and set-formation/ reversal learning deficits after orbital PFC manipulation (altering reversal learning stages to explore errors arising from perseveration and learned non-reward; and three additional ID stages to explore slowed set- formation). In rats where four IDs do not result in set-formation, we wanted to know what they were doing instead of forming set. We have therefore added a ‘probe’ stage (where only stimuli from the irrelevant dimension are changed), and a bi-conditional stage (where configurations of stimuli – e.g. if odour A, then medium B – predict reward). In set-deficient subjects, probe stages are more difficult (because novel stimuli disrupt established learning if they aren’t treated as irrelevant), whereas bi-conditional stages are easier (because trying to parse relevant from irrelevant disrupts learning when a configuration of ABSTRACTS A125 stimuli from both dimensions is rewarded). Results: Rats with an inactivated subthalamic nucleus are set- deficient, and do not show evidence of an attentional set after multiple ID stages. They are also impaired at probe stages, yet are improved relative to control performance at bi-conditional stages – suggesting a mechanism for solving discriminations when relevant cannot be distinguished from irrelevant is to learn the stimuli configurally. Conclusions: Modifying the rodent ASST to identify and assess specific psychological mechanisms is an important investigative tool for establishing cognition in both normal and impaired subjects. Treatment for patients with neurological disorders can only be improved by more precise understanding of the cognitive processes to target. DS Tait is funded by the Univ of St Andrews.

PW1 THE ROLE OF ARC IN MEMORY AND PSYCHIATRIC DISORDERS Trent S, Neuroscience & Mental Health Research Institute, 3.34 Hadyn Ellis Building, Maindy Road, Cardiff, CF24 4HQ [email protected] Arc is an activity-inducible immediate-early gene whose mRNA is dendritically targeted and accumulates in the spines and post-synaptic densities of activated synapses (Steward, O, et al 1998 Neuron 21(4):741– 51). Arc protein supports all forms of synaptic plasticity (Shepherd, J. & Bear, MF, 2011 Nat Neurosci 14(3): 279–84) and associative forms of memory consolidation including contextual fear memory (Guzowski, JF, et al., 2000 J Neurosci 20(11): 3993–4001), with direct effects on glutamatergic AMPA receptor trafficking and actin cytoskeleton stability (Chowdhury, S, et al., 2006 Neuron 52(3):45–59; Messaoudi, E, et al., 2007 J Neurosci 27(39):10445–55). My work has shown that hippocampal Arc acts to constrain extinction of contextual fear memory in rats, concomitantly arguing against a role for Arc as a molecular correlate of memory reconsolidation (Trent, S. et al., 2015 Nat Commun 6:7897). This was demonstrated through the acute knockdown of Arc using antisense oligodeoxynucleotides at recall, which did not lead to the permanent disruption of memory and could be rescued under reminder conditions. Ongoing work within our laboratory now expands the known molecular milieu within which Arc regulates extinction, placing AMPARs as an upstream mediator of extinction-relevant Arc levels and actin cytoskeleton rearrangements as a downstream Arc effector pathway (Trent et al., 2016, submitted). Meanwhile, recent genomic studies performed in patients with psychiatric disorders have revealed the role of ARC-mediated pathways in conferring increased susceptibility to schizophrenia (Hall, J, et al., 2015 Biol Psychiatry 77(1):52–8). Moreover, the dysregulation of extinction in humans is often observed in a variety of disorders (Myers, KM, 2011 Neuropsychopharm 36:274–293) including schizophrenia (Hall, J, et al., 2009 Trends Neurosci 32:359–365; Holt, DJ, et al., 2009 Biol Psychiatry 65(6):455–63), suggesting the ARC-mediated control of extinction may be highly pertinent to the cognitive deficits and underlying pathology of schizophrenia. To study this further, I am utilising a novel rat and mouse model of reduced Cyfip1 dosage, a penetrant risk gene associated with schizophrenia and autism spectrum disorder (Fromer, M, et al., 2014 Nature 506(7487): 179–184), with a critical role in regulating Arc protein translation (Napoli, I, et al., 2008 Cell 134(6):1042–54). Current studies are determining whether Arc dysregulation underlies Cyfip1-relevant molecular and behavioural phenotypes, including extinction, and ultimately, whether the manipulation of Arc levels can provide a route towards phenotypic rescue.

PW2 OMEGA-3 POLYUNSATURATED FATTY ACIDS (N3 PUFAS), INFLAMMATION, DEPRESSION, AND ADHD Chang JP, G33.72, The Maurice Wohl, Clinical Neuroscience Institute, Cutcombe Road, London, SE5 9RT [email protected] Depressive disorders with predominantly somatic symptoms are the most common form of depression, on the other side, patients suffering with chronic medical disorders often present comorbid depression. The mechanisms behind the development of depression remain still largely unclear, however the phospholipid hypothesis of depression is proving an enlightening and promising path to increase our understanding of depression. In particular, omega-3 polyunsaturated fatty acids (n3 PUFAs), a type of essential fatty acids, has been suggested to provide the possible link between depression and somatic presentations in medical disorders via modulation of inflammatory pathways. A126 ABSTRACTS

In this presentation I will first describe two studies to demonstrate the important role of n3 PUFAs, including docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), in patients with depression, especially in the context of increased inflammation, such as in those with comorbid cardiovascular disorders and with interferon alpha induced depression. In the second part of the talk, I will describe a recent study of n3 intake/deficiency and cognitive function in children with attention deficit hyperactivity disorder (ADHD). Cellular (cytokine-related) -mediated immune mechanisms have been suggested to contribute to the etiology of ADHD and I will present some preliminary data to illustrate the potential role of n3 PUFAs in ADHD. In short, these studies have provided evidence of the role of n3 PUFAs in depression and in ADHD. Future studies will need to investigate the efficacy of n3 PUFAs as a potential treatment option for these patients.

PW3 BRIDGING THE GAP: THE UTILITY OF TRANSLATION AND BACK-TRANSLATION APPROACHES IN UNDERSTANDING NEUROBIOLOGICAL MECHANISMS Tunbridge E, Neurosciences Building, University Department of Psychiatry, Warneford Hospital, Oxford, OX3 7JX [email protected] The last 3 years have witnessed a revolution in our understanding of the detailed genetic basis of psychiatric disorders. Large scale studies have identified hundreds of genomic loci that are robustly associated with mental health conditions. However, understanding the mechanisms linking these markers with disease will require a herculean effort. Studies addressing this question will necessarily be hypothesis-driven and will require a multidisciplinary approach, as well as likely requiring translation and back-translation between animal models and human subjects. This presentation will consider some of the promises and complexities of this approach, using examples from my research to date. Specifically, I will discuss how translation and back-translation between rodents and humans has clarified the role of catechol-O-methyltransferase (COMT) on dopamine and cognitive function. I will show how rodent models have been useful in clarifying the mechanisms underlying observed relationships between COMT genotype and human cognitive function. I will then demonstrate how predictions from these rodent studies were followed up in humans. In turn, these human studies revealed a (perhaps surprising) effect of COMT on reward processing, which we have begun to investigate in an animal model of altered COMT function. These findings provide examples of the utility of both translational (animal to human), and back- translational (human to animal) approaches in understanding the neurobiological effects of COMT. I will conclude by considering to what extent this approach might be applied to loci emerging from large-scale genetics studies of mental health conditions, as well as the potential complexities that are likely to be encountered along the way.

PW4 WHAT DO TREATMENT TARGETS IN DEPRESSION AND ANXIETY LOOK LIKE? Browning M, Warneford Hospital, Oxford OX3 7JX [email protected] The targets of treatments for depression and anxiety—the specific processes that the treatments are designed to engage and alter—vary enormously depending on the form of the treatment. For example, SSRI antidepressants target serotonin reuptake from the synapse whereas cognitive behavioural therapy targets habits in how patients think. Although these targets are clearly very different in nature they actually describe a common system, the individual patient interacting with their environment, at different levels of analysis. This suggests that to truly understand how a treatment works one must look beyond the traditional target of that treatment, and assess how it influences the patient at all levels. I will briefly describe two studies which have taken this approach to a) address questions of treatment mechanism in anxiety and depression and b) attempt to translate the mechanistic findings into clinically useful tools. From the mechanistic perspective I will describe a study which demonstrates how a computational approach to modelling behaviour can be used to define novel treatment targets yb linking ABSTRACTS A127 symptoms to cognitive processes and the underlying neurochemistry. From the applied perspective I will describe an ongoing multi-site randomised controlled trial which is testing the clinical impact of using computer based measures of cognition to guide antidepressant use in primary care. Overall these studies indicate that it is possible, and perhaps useful, to extend how we think about treatment targets in psychiatry to incorporate molecular, cognitive and social levels of analysis. A128 PRESENTING AUTHORS

Presenting Author Page Presenting Author Page Adams RA A112 Dutta A A80 Al-Ali A A38 Edye ME A106 Alexander L A72 Freeman TP A53 Alford C A119 Galea LAM A11 Antonesei A A79 García-Bea A A41, A42 Baldwin DS A121 Gaspar P A122 Bannerman DM A3 Godlewska BR A83 Barlow RL A116 Goh J-Y A36 Barnes TRE A29 Gold JM A13 Bartlett JB A55 Gordon RP A105 Bergmann O A11 Grabski MG A58 Biernacki KC A60 Grafton-Clarke D A82 Bijleveld BA A83 Hales CA A66 Bloomfield MAP A31 Harrison PJ A8 Bolea Alamanac BM A104 Harte MK A1 Browning M A126 Hastings C A27 Burgess EE A118 Hawkins PCT A30 Butler K A102 Hayward A A115 Cadinu D A39 Heal DJ A19, A44, A45 Calevro AC A42 Heaney LM A35 Carhart-Harris RL A87 Herane Vives A A89, A90 Carlisi CO A110 Higgs S A18 Carlyle M A54 Hindocha C A63 Carr SM A46 Horst NK A114 Cassaday HJ A43 Hou R A102 Chang JP A108, A125 Hurlemann RH A6 Charpentier CJ A101 Husain M A14 Cheng X A96 Hussein AN A46 Chesters RA A48 Jackson MG A98 Clifton NE A37 Jauhar S A21 Cotel MC A37 Jones IR A6 Craddock N A7 Juruena MF A91, A92, A103 Crockett MJ A4 Kaisari P A109 Dakwar E A10 Kinch AJ A97 D’Ambrosio E A20 Kohli S A35 Das RK A100 Kotoula V A111 David J A65 Koulouris CR A64 Davies SJC A105 Kowalczyk OS A52 Dazzan P A16 Krupitsky E A9 Dean ZPC A75 Kulkarni RS A45 Demetriou L A73 Langfield TIR A57 den Ouden HEM A4 Lawn W A53 Deslandes PN A28 Lawrence NS A117 Di Capua G A68 Lawrie SM A8 Di Simplicio M A85 Leganes M A51 Doherty H A67 Little HJ A61 Dourish CT A2 Lovick TA A5 Dunphy-Doherty F A69 Lucassen PJ A12 PRESENTING AUTHORS A129

Presenting Author Page Presenting Author Page Lyons T A86 Shaban NDC A30 Marques TR A21 Sharpley AL A119 Marsh BR A57 Singh JB A9 Martens MAG A112 Smee JA A78 Marwood L A94 Smith SL A44 Masaki CO A75 Sohail A A59 McAllister-Williams RH A120 Solowij N A61 McAndrew A A51 Stavropoulou Deli A A77 McCabe C A15 Stice E A19 McClung CA A7 Stone JM A23 McDonnell CW A71 Stuart SA A123 Merritt K A27 Sykes LH A40 Mkrtchian A A100 Tait DS A124 Modebadze T A123 Talbot PS A1 Mokrysz C A62 Taylor MJ A88, A94 Moran PM A3 Thaweethee B A70 Morgan CJ A10 Thomson DM A122 Murphy SE A74 Timmermann CB A49 Nettis MA A26 Tiwari N A93 Newcomer J A25 Trent S A39, A125 Nord CL A72 Tunbridge E A126 O’Callaghan C A108 Turton RT A116 O’Driscoll D A64 Upton N A76 Olney L A90 Vaghi MM A81 Openshaw RL A113 Valentino RJ A5 Padovan CM A44, A70 Valton V A56 Panchal P A95 van Kesteren CFMG A17 Papciak JK A68 Vernon AC A16 Pariante C A12 Voon V A4 Paton C A96, A110 Wall MB A50 Pillinger TE A25 Walsh AEL A78 Podda GP A32 Walsh KH A55 Potkin S A24 Waltmann M A76 Preller KH A49 Wang L A89 Prinssen E A15 Warnock A A88 Reis Marques T A23 Wise T A81 Reveles Jensen K A47 Worrell C A22 Reynolds GP A34 Ziauddeen H A18 Robinson ESJ A13 Robinson OJ A99 Rogdaki M A32 Roseman L A86 Rzepa E A84 Sadler C A59 Sartor ML A98 Scott L A33 Seroiska R A107 Sethna V A104