Clinical Research

14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 103

Clinical research

Neurofeedback and networks of depression David E. J. Linden, Dr med, Dr phil, DPhil (Oxon)

Self-regulation through neurofeedback: technique and rationale

Since its invention 20 years ago, functional magnetic resonance imaging (fMRI) has become a central technique of cognitive and clinical neuroscience. The particular strengths of this noninvasive technique are its spatial resolution, fidelity, and ability to reach deep subcortical structures. Its whole-brain coverage enables the Recent advances in imaging technology and in the under- mapping of functionally connected networks and the standing of neural circuits relevant to emotion, motivation, extraction of information from distributed activation pat- and depression have boosted interest and experimental terns. These features make fMRI particularly suitable for work in neuromodulation for affective disorders. Real-time applications to mental disorders, where the pathology is functional magnetic resonance imaging (fMRI) can be used generally assumed to reside in faulty network activity, to train patients in the self-regulation of these circuits, and rather than focal lesions, and where deep structures play thus complement existing neurofeedback technologies a major role. The fMRI technique is particularly powerful based on electroencephalography (EEG). EEG neurofeed- in mapping correlates of mental states, another very back for depression has mainly been based on models of attractive feature for psychiatry, which deals predomi- altered hemispheric asymmetry. fMRI-based neurofeed- nantly with altered states of thought, emotion, and behav- back (fMRI-NF) can utilize functional localizer scans that ior. For example, fMRI scans acquired from patients with allow the dynamic adjustment of the target areas or net- chronic schizophrenia during the experience of auditory works for self-regulation training to individual patterns of verbal hallucinations have revealed activation in the audi- emotion processing. An initial application of fMRI-NF in tory cortex, very similar to that during stimulation with depression has produced promising clinical results, and fur- actual sounds.1 Beyond their major contribution to the ther clinical trials are under way. Challenges lie in the understanding of the brain correlates of psychopathology, design of appropriate control conditions for rigorous clin- fMRI studies have also informed our understanding of ical trials, and in the transfer of neurofeedback protocols from the laboratory to mobile devices to enhance the sus- Author affiliations: MRC Centre for Neuropsychiatric Genetics and Genomics, tainability of any clinical benefits. Neuroscience and Mental Health Research Institute, National Centre for Mental Health, Cardiff University, Cardiff, UK © 2014, AICH – Servier Research Group Dialogues Clin Neurosci. 2014;16:103-112. Address for correspondence: Institute of Psychological Medicine and Clinical Keywords: emotion; frontal lobe; functional magnetic resonance imaging; limbic Neurosciences, School of Medicine, Cardiff University, Hadyn Ellis Building, system; mood disorder; self-regulation Maindy Road, Cardiff, CF24 4HQ, UK (e-mail: [email protected])

Clinical research

Selected abbreviations and acronyms oxygenation level-dependent (BOLD) signal (Figure 1). DBS deep brain stimulation Thus, fMRI-NF can presently only be conducted while EEG electroencephalography participants are in a magnetic resonance system. The sig- EEG-NF electroencephalography–based neurofeedback nal can be based on the average time course of an indi- fMRI functional magnetic resonance imaging vidual area (such as the left primary motor cortex or the fMRI-NF functional magnetic resonance imaging–based right amygdala) or even on the time course of a single neurofeedback voxel anywhere in the brain (although this would make MEG magnetoencephalography it rather susceptible to noise). However, it can also be based on results of more complex computations, such as the effects of risk genes on cognitive and affective net- the activation difference or correlation between two works.2 These important research contributions have led areas, or the output of a multivariate pattern classifica- to strategies for the development of fMRI paradigms for tion algorithm. Unlike electrophysiological neurofeed- diagnostic, prognostic, or therapeutic use in mental disor- back techniques, such as EEG (electroencephalogra- ders, and are reviewed in the September 2013 issue of phy), the fMRI technique cannot provide truly Dialogues in Clinical Neuroscience (http://www.dialogues- “real-time” feedback because of the “hemodynamic” cns.org/wp-content/themes/dcnsv2/publication.php? delay of ≈5 seconds between the actual neural activity volume=15&issue=3). and the vascular response that creates the fMRI signal. Whereas concerns about power and reliability3 have However, this delay does not pose an obstacle to neuro- dampened hopes for imminent diagnostic uses of func- feedback training when participants are informed of it.4 tional imaging, there has recently been a surge of inter- Compared with other neurofeedback techniques est in a potential therapeutic application of fMRI-based (EEG or magnetoencephalography, MEG) and to non- neurofeedback (fMRI-NF). Imaging-based neurofeed- invasive physical stimulation techniques (transcranial back follows similar principles as other neuro- or direct current stimulation and transcranial magnetic biofeedback approaches. During neurofeedback train- stimulaiton), fMRI-NF has the advantage of higher ing, participants receive feedback on their brain activity localization accuracy and better access to deep brain in real time and are instructed to change this activation. structures. EEG-based neurofeedback (EEG-NF) has In the case of fMRI-NF, the feedback signal is computed the advantage of being more widely available and from a real-time analysis of the time course of the blood including ambulatory settings. It is a popular procedure, especially in child and adolescent mental health settings Real-time fMRI neurofeedback setup in application to attention deficit/ hyperactivity disorder (ADHD),5,6 although a recent meta-analysis has raised doubts about the specificity of the effects in ADHD.7 Feedback Several studies that have also been conducted with screen MRI EEG-NF in depression will be reviewed below. scanner Compared with deep brain stimulation (DBS),8,9 fMRI-NF has the advantage of noninvasiveness and spatial flexibility. However, it is too early to make any direct comparisons of the clinical effects of these two techniques in psychiatry, which have so far been used for very different patient populations due to the restriction of DBS to severe and treatment-refractory cases. Neurofeedback also differs from all external stimulation techniques in that it enables the patients themselves to control their brain activity and thus to contribute to their Analysis PC Stimulation PC experience of self-efficacy, which may be an important therapeutic factor.10 This aspect will be discussed in more Figure 1. Basic diagram of a real-time functional magnetic resonance imaging brain-computer interface for neurofeedback. detail below, in the section of links between neurofeed- Figure courtesy of Isabelle Habes back and social learning theory.

104 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 105

Neurofeedback and networks of depression - Linden Dialogues in Clinical Neuroscience - Vol 16 . No. 1 . 2014

There are, in principle, at least two ways in which self- the context of the patient’s primary pathology. For exam- regulation of brain activity through neurofeedback may ple, it may be beneficial to suppress canonical thought be beneficial for depression and other mental disorders. processes such as self-comparison in the context of a Self-regulation training might address a primary abnor- depressive disposition. The great progress in the under- mal process, such as hyper- or hypoactivation of specific standing of the circuits of cognitive, affective, and social brain areas or networks. For this approach, it would be information processing made through the last two necessary to identify such abnormal activation patterns decades of functional imaging can thus inform the design in individual patients beforehand. Although research of imaging-based clinical neurofeedback protocols, even with the fMRI technique (and metabolic imaging with in the absence of primarily abnormal imaging signals. positron emission tomography, PET) has yielded several potential disease-relevant targets for depression, notably Applications in depression imbalances between prefrontal and limbic areas,11-14 none of these have been validated as biomarkers for use in Symptoms of depression can be broadly grouped into individual patients. Similarly, although intriguing results the four domains of emotion regulation, cognition, moti- have been obtained with EEG mapping techniques in vation, and homoeostasis (Table I).19 Although like all relation to hemispheric asymmetries in depression (see categorization of psychological phenomena, this classi- EEG section below), these have not attained individual fication is somewhat artificial (and some symptoms map biomarker status either. At the present time, there is onto more than one category), it can help the search for insufficient evidence to identify any reliably abnormal, the biological mechanisms of depression.20 Furthermore local, or distributed brain activation patterns in individ- if the neural systems underlying some of these functional ual patients with depression that could be targeted with clusters prove to be modifiable (by pharmacological, neurofeedback (or indeed, any other neuromodulation psychological, or physical intervention) they may technique, including DBS). However, neuromodulation can also act in a differ- ICD-10 DSM-IV Domain ent way, by activating or suppressing circuits that are not Depressed mood ER primarily abnormal, but whose modulation may never- Loss of interest and enjoyment M, ER theless produce clinical benefits. I have argued for the Increased fatiguability M, H consideration of such a “functional systems” approach Reduced concentration and attention C, M recently elsewhere.15 Most biological treatments in psy- Ideas of guilt and unworthiness C chiatry probably follow this path already. For example, Ideas or acts of self-harm or suicide, thoughts of death C, ER monoamine reuptake inhibitors benefit many patients Sleep disturbance H with depression by increasing serotonergic and/or nora- Disturbed appetite/weight change H drenergic neurotransmission, but probably not by cor- Pessimistic view of the future C, ER recting an underlying monoaminergic deficit, for which Reduced self-esteem and self-confidence C, ER little evidence has been found.16,17 Similarly, current Early-morning awakening H lesion surgery and DBS approaches for depression or Mood worse in the morning H, ER obsessive-compulsive disorder (OCD), whose targets all Psychomotor retardation or agitation M, H converge onto pathways from brain stem and basal gan- Weight loss H glia to prefrontal cortex,18 work through—hitherto Loss of libido H poorly understood—effects on the function of these Table I. Symptoms of depression. Five symptoms are required over a 2- pathways in motivation and emotion regulation, but not week period for an episode of major depression (DSM-IV). ICD- necessarily because there are documented primary 10 defines depressive episodes by a combination of the most typ- abnormalities in these pathways. Regarding neurofeed- ical (printed in bold face) and other symptoms.19 The number of back, this implies that clinical benefits may be obtained symptoms determines the severity of the episode: 2 typical and from self-regulation training that activates compensatory 2 other: mild; 2 typical and 3 or 4 other: moderate; 3 typical and 4 or more other: severe. The column on the right indicates the circuits for particular cognitive processes (eg, emotion broad domains into which the symptoms can be tentatively clas- regulation) or inhibits circuits that, although normal sified: ER, emotion regulation; C, cognition; M, motivation; H, when viewed in isolation, contribute to dysfunction in homoeostasis.

105 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 106

Clinical research

become viable targets for new antidepressant therapies. Even if this combination of advanced functional (and New therapeutic strategies for depression are sorely structural) neuroimaging and a symptom cluster–based needed. Depression is expected to assume the first place approach to depression does not produce clear, individ- in the WHO’s global disease burden statistic by 2020. It ually targetable state markers, the knowledge of the func- affects up to 15% of the population of industrialised tional systems involved can still inform new treatment countries. The number of prescriptions of antidepres- approaches, notably in neuromodulation. I have argued15 sants for England alone was almost 50 million in 2011, at that the biological correlates of the mechanisms that help a cost of £270 million, but the overall health care and to overcome a mental illness, such as emotion regulation socioeconomic costs of depression are much higher, at or fear extinction,25 may be more consistent than those of about £11 billion per year in the UK according to a the original illness. Thus, if we can apply functional imag- recent House of Commons report.21 The impact of ing to reveal the neural correlates of successful treat- depression on health and wellbeing is not confined to ment26-29 (see also the article by Beauregard in this issue, the patients themselves, but frequently extends to their p xx), we can subsequently apply neuromodulation tech- human networks, negatively affecting social, familial, and niques to target these neural networks directly. This was occupational relationships.22 Depression is presently the rationale behind the target selection for the first DBS managed with psychological, pharmacological, or phys- protocol to depression, which targeted the subgenual cin- ical interventions or their combination. However, all pre- gulate cortex based on the observation that activation in sent treatment options have limitations, such as medica- this area was reduced after successful pharmacotherapy tion side effects, nonresponse (including a high of depression11,30 (see also the article by Holtzheimer in proportion of treatment-refractory patients who do not this issue, p xx). This idea of mimicking the neural corre- respond to any therapy),23 and frequent relapse. Even lates of successful treatments through direct brain inter- patients who have responded to antidepressant treat- vention can now be implemented even more flexibly ment are often reluctant to take medication in the long through self-regulation training with neurofeedback, term and thus experience an increased relapse risk.24 which can even track moving targets (unlike psychiatric Together these complex challenges underscore the need surgery, which is normally confined by a specific lesion or for better, and more effective, treatment and relapse pre- stimulation site), because functional localizers can be vention options for depression, and for solutions that are adjusted flexibly over treatment sessions. to be designed through interaction between researchers, clinicians, and the patients themselves. EEG neurofeedback in depression A functional imaging approach (in the broad sense, incorporating both fMRI and electrophysiological tech- EEG-NF studies of depression were originally based on niques) to elucidating the circuits underlying the symp- Davidson’s approach/withdrawal model of emotion,31 tom complexes of depression, but also of those involved which posited that appetitive and aversive emotional in their remediation, can be useful in this new therapeu- behaviors are subserved by the left and right frontal cor- tic endeavor in several respects. Firstly, it may allow tex respectively,32 and that hypoactivity of left frontal researchers to identify correlates of individual symptoms areas would be associated with depression.33,34 Because or symptom groups, for example, altered activation of alpha activity of the EEG is commonly linked with lower frontostriatal circuits during period of apathy and fatigue. metabolic activation, this relative left hypoactivity would If these imaging-based state markers can be shown to be be associated with relatively higher right than left frontal reliable and diagnostic, they may become new targets for alpha power. The logical consequence in neurofeedback self-regulation training through neurofeedback (or other terms would be to train patients to decrease left-hemi- neuromodulatory interventions). With further refinement spheric alpha activity, increase right-hemispheric alpha of functional imaging methods and higher signal-to-noise activity, or shift an asymmetry index toward the right in ratio obtained through higher field strengths, there may order to rebalance activation levels in favor of the left even be scope for a detailed functional mapping of brain hemisphere. This asymmetry model received initial sup- stem nuclei that may reveal information about the under- port from the stroke literature because depression lying chemical imbalances, thus possibly giving rise to seemed to occur more frequently after damage to the left new pharmacological strategies. than the right hemisphere. However, current neuropsy-

106 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 107

Neurofeedback and networks of depression - Linden Dialogues in Clinical Neuroscience - Vol 16 . No. 1 . 2014

chiatric evidence suggests that there is no such preferred pling) and has a much lower temporal resolution than association between depression and left-hemispheric the electrophysiological techniques (in the second range damage.35 The EEG literature has also been inconsistent compared with the millisecond precision of EEG and in that not all authors found higher left-hemispheric MEG), its spatial resolution and access to deeper brain alpha activity in patients with depression compared with structure make it an attractive tool for network mapping healthy controls,36 although a recent meta-analysis sup- in psychiatric disorders and neurofeedback.42 Our ported the asymmetry model based on resting EEG research group has designed an fMRI-NF protocol for data.37 The considerable interindividual variability of patients with depression (Figure 2). EEG asymmetry limits its usefulness as a neurofeedback Rather than using anatomically fixed target regions target.38 The main asymmetry-based EEG-NF protocol (as conventionally used in psychiatric surgery), the has used an asymmetry index of alpha power as feedback fMRI-NF approach gave us the opportunity to identify signal and trained patients to increase the right-to-left the relevant target areas in each training session using a ratio, essentially rebalancing a putative hypoactivation of functional localizer. Localizer scans with emotionally the left hemisphere. This asymmetry index is computed charged pictures can identify areas involved in the pro- as A=100×(R-L)/R+L), where R and L are the square cessing of positive or negative affective stimuli, and we root of power of alpha activity (obtained by Fast Fourier initially showed that healthy participants can attain con- Transformation) measured at a right and left frontal elec- trol over the activation levels in these areas.44,45 The “pos- trode respectively.39 itive emotion” areas were then used as the target for Compared with earlier research, which did not incor- fMRI-NF in a pilot study with patients with mild-to- porate control groups, a recent placebo-controlled ran- moderate levels of depression.46 domized (but not blinded) study has implemented sev- We tested eight patients, all with a longstanding his- eral design improvements.40 This study (again with the tory of depression. They were informed that the areas alpha asymmetry training protocol) included 24 patients they trained to upregulate had been associated with pos- with depression who were assigned to a 5-week EEG- itive emotional pictures, but no specific strategy was sug- NF or a psychotherapy control group. Patients in the gested to them. Most patients started their attempts to active group showed an improvement of over seven upregulate the target areas that, although varied in local- points on the 17-item Hamilton Depression Rating Scale (HDRS; from 11.33 to 4.08). Conversely, the psychotherapy group only showed minimal improvement (12.36 to 11.08). However, the scope of this study is lim- ited by the relatively low depression levels before treatment (a mean HDRS score of 12 indicates mild depression and is below the conventional cutoffs for treatment trials). Furthermore, the choice of control group does not exclude the impact of nonspecific effects of neurofeedback training, for example, the gaming component, which may make the training more interesting and engaging than conventional psychotherapy.

fMRI neurofeedback in depression 20 seconds 20 seconds 20 seconds ... Rest Imagine Rest

One of the limitations of EEG-NF is its low spatial pre- Figure 2. Display screen of visual neurofeedback with an outline of the cision, which is owed to the effects of volume conduc- protocol. The patients trained to increase activity in functionally tance and the attenuation of electrical signals on their localized areas during 20-second periods, alternating with 20 second periods of rest. Overall, they did this for 20 minutes each way from the source to the scalp, and the ill-posed in four weekly sessions. nature of the source localization problem.41 Although the Adapted from ref 43: Subramanian L, Hindle JV, Johnston S, et al. Real- time functional magnetic resonance imaging neurofeedback for treat- fMRI technique provides only indirect measures of ment of Parkinson's disease. J Neurosci. 2011;31:16309-16317. neural activity (obtained through neurovascular cou- Copyright © Society for Neuroscience 2011

107 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 108

Clinical research

ization, mostly included areas in the ventral prefrontal University (clinicaltrials.gov: NCT01544205) pits upreg- cortex and limbic system, by imagining the pictures, ulation of emotion networks against upregulation of a which included serene landscapes and uplifting sporting higher visual area, a rather conservative active control scenes. However, most of them reported at debriefing condition that also involves mental imagery and the that it worked better to try and evoke positive images rewarding experience of brain self-control. This trial will that related to themselves, for example, memories of also provide some initial information about any sus- happy events. In a psychotherapeutic sense, the neuro- tained benefits of fMRI-NF by including a follow-up feedback training may have helped them engage with assessment 1 month after the completion of the 2-month positive aspects of their own lives. The clinical effects of intervention. So far no published information is avail- the pilot study were also promising.46 The patients in the able about any longer-term benefits of fMRI-NF in neurofeedback group improved by about 30% on their depression or in any other mental or neurological dis- symptom score over the 1-month trial (about four points order. on the 17-item HDRS), whereas a control group, which performed emotional imagery for the same duration Neurofeedback and social learning theory outside the scanner, did not improve at all. The next step in the development of fMRI-NF into a Isolating the generic effects of the experience of self- potential therapeutic tool will be the investigation of its control from any region- or network-specific effects, is short- and long-term benefits and mechanisms in rigor- particularly relevant because neurofeedback may have ous trials. Essentially the same standards apply as those nonspecific positive effects on self-efficacy. If this was required before the introduction of a new drug. We can shown to be the main factor in any treatment effects, this be relatively certain that neurofeedback has no major information would considerably influence the develop- direct side effects,47 but cannot rule out that some ment of neurofeedback protocols. According to patients may experience parts of the procedure as stress- Bandura, “ People have to live with a psychic environ- ful. Furthermore, researchers in the field will have to ment that is largely of their own making. Many human show that the clinical benefits are not merely placebo distresses result from failures to control disturbing, rumi- effects induced by patients’ expectations, but genuinely native thoughts. Control of one’s thought processes is superior to other interventions. The design of appropri- therefore a key factor in self-regulation of emotional ate control conditions for clinical trials is a challenge. states.”48 The current standards of randomized controlled trials If neurofeedback is to benefit patients by helping were developed with drug studies in mind, where the them attain control of their own thought processes and aim is to distinguish the chemical effects of a drug from consequently their emotional states, this will probably the associated expectations. One of the principles of require fine tuning of self-regulation protocols to the these trials is that they are conducted in a “double-blind” appropriate neural networks. However, social learning fashion. Yet when treatments require the active collab- theory also posits that depression can be caused by a oration of the patients, which is the case in neurofeed- general low sense of agency and loss of experience of back (and also in all forms of psychotherapy), these control of the environment. For these patients the patients cannot be completely “blind.” Furthermore, the “imposed environment,” in Bandura’s terms, would take experience of gaining control over the brain, the precedence over the “constructed environment.” increased “self-efficacy” and heightened awareness of Successful control over one’s own brain activity (and in one’s own mental states may all be nonspecific compo- this scenario the exact region would probably matter nents of neurofeedback that contribute to improvement less) could then give patients a sense of agency and par- across disorders. Although we can control them with ticularly the experience that their own brain activity is sophisticated experimental designs, this may miss the constructed (rather than merely imposed). point, as these psychological mechanisms may actually Neuromodulation might receive a further, presently be valuable drivers of change for the patients, rather speculative, interesting inspiration from social learning than mere components of a placebo effect. An ongoing theory. Already in 1999 Bandura pointed out that, randomized controlled trial of neurofeedback for “Electronic technologies greatly extend human capabil- depression conducted by the author’s group at Cardiff ities to test the likely outcomes of given decisions and

108 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 109

Neurofeedback and networks of depression - Linden Dialogues in Clinical Neuroscience - Vol 16 . No. 1 . 2014

courses of action through the use of computerized enact- The importance and pervasiveness of such virtual ments in simulated realities without having to carry out simulations has increased even further in the past 14 the activities.”48 years and started to enter the field of mental health, for

Neurofeedback in depression: attractive neural targets

DLPFC

VS VLPFC

Cognitive control/ Reward circuitry emotion regulation circuitry (ventral striatum) (lateral PFC)

SubgACC

Amy

Cortical limbic circuitry Subcortical limbic circuitry (subgenual ACC) (amygdala)

Figure 3. Cognitive-affective brain systems that could become targets for neuromodulation in depression. DLPFC, dorsolateral prefrontal cortex; VLPFC, ventrolateral prefrontal cortex; ACC, anterior cingulate cortex; Amy, amygdala Adapted from ref 38: Esmail S, Linden D. Cogn Sci. 2011;6. Copyright © Nova Science Publishers, Inc.

109 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 110

Clinical research

example in the treatment of post-traumatic stress disor- iments that areas in the midbrain, striatum, and frontal der.49 Neurofeedback may now provide an avenue cortex, linked anatomically through the medial forebrain toward simulation, not just of environments and mental bundle and chemically through the neurotransmitter processes, but of the relevant brain processes themselves dopamine, support the ability to experience and learn (although the term “simulation” here is an incomplete from rewards. These “reward circuits” would therefore analogy, because even with neurofeedback the neural be another potentially suitable target for fMRI-NF, as changes would be real rather than virtual). One poten- they are for DBS. tial application might be the simulation of the effects of Another area for development in clinical research DBS, which could guide the later placement of perma- into fMRI-NF is the identification of suitable patient nent neuromodulation devices based on the behavioral, populations and predictive markers. For example the cognitive, or clinical effects of transient brain activation cognitive and motivational factors that underlie suc- changes during neurofeedback. cessful neurofeedback training are largely unknown. One option would be to include metacognitive scales Future developments such as the Thought Control Questionnaire (TCQ),53 the Thought Control Ability Questionnaire (TCAQ),54 and In addition to the rigorous testing of existing fMRI-NF the behavioral inhibition system and behavioral activa- protocols it will also be attractive to develop new pro- tion system (BIS/BAS) scale55 in order to enable predic- tocols based on different ways of extracting information tions of feasibility of neurofeedback and clinical effects. from brain activation data, eg, multivoxel pattern analy- Another recommendation would be to assess the short- sis,50-52 or on different brain networks. The choice of term changes associated with individual neurofeedback potential target areas can be informed both by the expe- sessions on patients’ mood and perceived self-regulation rience of other neuromodulation techniques, particularly ability in order to evaluate whether these immediate lesion surgery and DBS, and by the symptom clusters effects are associated with the longer-term clinical discussed above. The fMRI-NF approach to depression response. At the moment it is envisaged that neurofeed- has so far focused on emotion regulation, and thus the back, like DBS, will be a procedure that is added to overlap between the cognitive and affective domains. existing treatments, rather than one that replaces it. With The target areas for this approach are mainly in the these caveats the prospect of neurofeedback as a treat- frontal lobe (Figure 3).38 Another approach starts from ment for depression sound far more prosaic, but still the the observation that many patients with depression are potential is considerable. ❏ impaired in their ability to react to rewards or generally to have positive experiences (lack of enjoyment: “anhe- Acknowledgments: Supported by the Medical Research Council (MRC Developmental Clinical Studies grant G 1100629). Figures were kindly donia”). It has been well established through functional provided by Isabelle Habes and Dr Leena Subramanian, and expert imaging in humans and a long tradition of animal exper- graphic support from Lorraine Woods is gratefully acknowledged.

REFERENCES 7. Sonuga-Barke EJ, Brandeis D, Cortese S, et al. Nonpharmacological interventions for ADHD: systematic review and meta-analyses of random- 1. Dierks T, Linden DE, Jandl M, et al. Activation of Heschl's gyrus during ized controlled trials of dietary and psychological treatments. Am J Psychiatry. auditory hallucinations. Neuron. 1999;22:615-621. 2013;170:275-289. 2. Linden DE. The challenges and promise of neuroimaging in psychiatry. 8. Lozano AM, Lipsman N. Probing and regulating dysfunctional circuits Neuron. 2012;73:8-22. using deep brain stimulation. Neuron. 2013;77:406-424. 3. Uttal WR. Reliability in Cognitive Neuroscience: a Meta-Meta Analysis. 9. Goodman WK, Alterman RL. Deep brain stimulation for intractable psy- Cambridge, MA: MIT Press; 2013. chiatric disorders. Annu Rev Med. 2012;63:511-524. 4. Weiskopf N, Mathiak K, Bock S, et al. Principles of a brain-computer 10. Bandura A. Self-efficacy: the Exercise of Control. New York, NY: WH interface (BCI) based on real-time functional magnetic resonance imaging Freeman; 1997. (fMRI). IEEE Trans Biomed Eng. 2004;51:966-970. 11. Mayberg HS, Liotti M, Brannan SK, et al. Reciprocal limbic-cortical func- 5. Gevensleben H, Holl B, Albrecht B, et al. Neurofeedback training in chil- tion and negative mood: converging PET findings in depression and normal dren with ADHD: 6-month follow-up of a randomised controlled trial. Eur sadness. Am J Psychiatry. 1999;156:675-682. Child Adolesc Psychiatry. 2010;19:715-724. 12. Phillips M, Ladouceur C, Drevets W. A neural model of voluntary and 6. Gevensleben H, Rothenberger A, Moll GH, Heinrich H. Neurofeedback automatic emotion regulation: implications for understanding the patho- in children with ADHD: validation and challenges. Exp Rev Neurother. physiology and neurodevelopment of bipolar disorder. Mol Psychiatry. 2012;12:447-460. 2008;13:829,833-857.

110 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 111

Neurofeedback and networks of depression - Linden Dialogues in Clinical Neuroscience - Vol 16 . No. 1 . 2014

Retroalimentación neural y redes de la Neurofeedback et réseaux de la dépression depresión

Los recientes avances en la imaginología y en la com- Des avancées récentes dans la technologie de l’ima- prensión de los circuitos neurales relacionados con la gerie et dans la compréhension des circuits neuro- emoción, la motivación y la depresión han impulsado naux liés à l’émotion, à la motivation et à la dépres- el interés y el trabajo experimental en la neuromo- sion ont stimulé l’intérêt et le travail expérimental dulación en los trastornos afectivos. La resonancia en neuromodulation dans les troubles affectifs. magnética funcional en tiempo real (RNMf) se puede L’imagerie par résonance magnétique fonctionnelle emplear para entrenar a pacientes en la autorregu- en temps réel (IRMf) est utilisée pour entraîner les lación de estos circuitos y así complementar las tec- patients à l’autorégulation de ces circuits et vient nologías existentes de retroalimentación neural donc compléter les techniques de neurofeedback basadas en la electroencefalografía (EEG). La retroa- basées sur l’électroencéphalographie (EEG) princi- limentación neural EEG para la depresión se ha palement fondée, dans le cadre de la dépression, basado principalmente en modelos de alteración en sur des modèles de changements d’asymétrie des la asimetría hemisférica. La retroalimentación neural hémisphères. Le neurofeedback basé sur l’IRMf, basada en la RNMf (RN-RNMf) puede emplear explo- (IRMf-NF), utilise des images de localisation fonc- raciones de localizaciones funcionales que permiten tionnelle qui permettent l’ajustement dynamique el ajuste dinámico de las áreas o redes blanco para el des zones ou réseaux cibles pour l’entraînement par entrenamiento en autorregulación para patrones autorégulation des processus émotionnels indivi- individuales del procesamiento de las emociones. Una duels. Les premiers résultats de l’IRMf-NF dans la aplicación inicial de la RN-RNMf en la depresión ha dépression sont prometteurs et d’autres études cli- producido resultados clínicos promisorios y también niques sont en cours. Les difficultés résident dans la ensayos clínicos que están en desarrollo. Los desafíos création de bonnes conditions de contrôle pour des están en el diseño de condiciones control apropiadas études cliniques rigoureuses et dans le transfert des para ensayos clínicos rigurosos y en la transferencia protocoles de neurofeedback du laboratoire à des de protocolos de retroalimentación neural desde el dispositifs mobiles pour améliorer le maintien des laboratorio a los dispositivos móviles para reforzar la bénéfices cliniques. sostenibilidad de cualquier beneficio clínico.

13. Beck A. The evolution of the cognitive model of depression and its neu- 24. Kim KH, Lee SM, Paik JW, Kim NS. The effects of continuous antide- robiological correlates. Am J Psychiatry. 2008;165:969-977. pressant treatment during the first 6 months on relapse or recurrence of 14. Clark D, Beck A. Cognitive theory and therapy of anxiety and depres- depression. J Affect Disord. 2011;132:121-129. sion: convergence with neurobiological findings. Trends Cogn Sci. 25. Pine DS, Helfinstein SM, Bar-Haim Y, Nelson E, Fox NA. Challenges in 2010;14:418-424. developing novel treatments for childhood disorders: lessons from research 15. Linden D. Biological psychiatry: time for new paradigms. Br J Psychiatry. on anxiety. Neuropsychopharmacology. 2009;34:213-228. 2013;202:166-167. 26. Beutel ME, Stern E, Silbersweig DA. The emerging dialogue between 16. Healy D. The structure of psychopharmacological revolutions. Psychiatr psychoanalysis and neuroscience: neuroimaging perspectives. J Am Dev. 1987;5:349-376. Psychoanal Assoc. 2003;51:773-801. 17. Hyman SE. Revolution stalled. Sci Transl Med. 2012;4:155cm11. 27. Linden DE. How psychotherapy changes the brain—the contribution of 18. Schoene-Bake JC, Parpaley Y, Weber B, Panksepp J, Hurwitz TA, Coenen functional neuroimaging. Mol Psychiatry. 2006;11:528-538. VA. Tractographic analysis of historical lesion surgery for depression. 28. Linden DE. Brain imaging and psychotherapy: methodological consid- Neuropsychopharmacology. 2010;35:2553-2563. erations and practical implications. Eur Arch Psychiatry Clin Neurosci. 2008;258 19. Linden D. The Biology of Psychological Disorders. Basingstoke, UK: Palgrave (suppl 5):71-75. Macmillan; 2011. 29. DeRubeis R, Siegle G, Hollon S. Cognitive therapy versus medication for 20. Mayberg HS. Depression: a neuropsychiatric perspective. In: Panksepp depression: treatment outcomes and neural mechanisms. Nat Rev Neurosci. J, ed. Biological Psychiatry. Hoboken, NJ: Wiley-Liss; 2004:197-228. 2008;9:788-796. 21. Economics A-PPGoW. Cost of Depression in England, 2010. All-party 30. Mayberg H, Lozano A, Voon V, et al. Deep brain stimulation for treat- parliamentary group on wellbeing economics web site. http://parliamen- ment-resistant depression. Neuron. 2005;45:651-660. tarywellbeinggroup.org.uk/reports/. Accessed December 13, 2013. 31. Davidson RJ, Ekman P, Saron CD, Senulis JA, Friesen WV. Approach- 22. Henderson M, Harvey SB, Overland S, Mykletun A, Hotopf M. Work and withdrawal and cerebral asymmetry: emotional expression and brain phys- common psychiatric disorders. J R Soc Med. 2011;104:198-207. iology. I. J Pers Soc Psychol. 1990;58:330-341. 23. Rush A, Trivedi M, Wisniewski S, et al. Acute and longer-term outcomes 32. Harmon-Jones E, Gable PA, Peterson CK. The role of asymmetric frontal in depressed outpatients requiring one or several treatment steps: a STAR*D cortical activity in emotion-related phenomena: a review and update. Biol report. Am J Psychiatry. 2006;163:1905-1917. Psychol. 2010;84:451-462.

111 14_CH_8004_BA_INTERIEUR.qxd:DCNS#55 3/03/14 18:05 Page 112

Clinical research

33. Henriques JB, Davidson RJ. Regional brain electrical asymmetries dis- 44. Johnston SJ, Boehm SG, Healy D, Goebel R, Linden DE. Neurofeedback: criminate between previously depressed and healthy control subjects. J a promising tool for the self-regulation of emotion networks. Neuroimage. Abnorm Psychol. 1990;99:22-31. 2010;49:1066-1072. 34. Henriques JB, Davidson RJ. Left frontal hypoactivation in depression. J 45. Johnston S, Linden DE, Healy D, Goebel R, Habes I, Boehm SG. Abnorm Psychol. 1991;100:535-545. Upregulation of emotion areas through neurofeedback with a focus on pos- 35. Bhogal SK, Teasell R, Foley N, Speechley M. Lesion location and post- itive mood. Cogn Affect Behav Neurosci. 2011;11:44-51. stroke depression: systematic review of the methodological limitations in 46. Linden DE, Habes I, Johnston SJ, et al. Real-time self-regulation of emo- the literature. Stroke. 2004;35:794-802. tion networks in patients with depression. PLoS One. 2012;7:e38115. 36. Reid SA, Duke LM, Allen JJ. Resting frontal electroencephalographic 47. Hawkinson JE, Ross AJ, Parthasarathy S, et al. Quantification of adverse asymmetry in depression: inconsistencies suggest the need to identify medi- events associated with functional MRI scanning and with real-time fMRI- ating factors. Psychophysiology. 1998;35:389-404. based training. Int J Behav Med. 2012;19:372-381. 37. Thibodeau R, Jorgensen RS, Kim S. Depression, anxiety, and resting 48. Bandura A. Social cognitive theory: an agentic perspective. Asian J Soc frontal EEG asymmetry: a meta-analytic review. J Abnorm Psychol. Psychol. 1999;2:21-41. 2006;115:715-729. 49. Gonçalves R, Pedrozo AL, Coutinho ES, Figueira I, Ventura P. Efficacy of 38. Esmail S, Linden D. Emotion regulation networks and neurofeedback virtual reality exposure therapy in the treatment of PTSD: a systematic in depression. Cogn Sci. 2011;6(2). review. PLoS One. 2012;7:e48469. 39. Rosenfeld JP, Baehr E, Baehr R, Gotlib IH, Ranganath C. Preliminary evi- 50. Sitaram R, Lee S, Ruiz S, Rana M, Veit R, Birbaumer N. Real-time sup- dence that daily changes in frontal alpha asymmetry correlate with port vector classification and feedback of multiple emotional brain states. changes in affect in therapy sessions. Int J Psychophysiol. 1996;23:137- Neuroimage. 2011;56:753-765. 141. 51. LaConte SM. Decoding fMRI brain states in real-time. Neuroimage. 40. Choi SW, Chi SE, Chung SY, Kim JW, Ahn CY, Kim HT. Is alpha wave neu- 2011;56:440-454. rofeedback effective with randomized clinical trials in depression? A pilot 52. Habes I, Krall S, Johnston S, et al. Pattern classification of valence in study. Neuropsychobiology. 2011;63:43-51. depression. Neuroimage Clin. 2013;2:675-683. 41. Bledowski C, Linden DE, Wibral M. Combining electrophysiology and 53. Wells A, Davies M. The Thought Control Questionnaire: a measure of functional imaging - different methods for different questions. Trends Cogn individual differences in the control of unwanted thoughts. Behav Res Ther. Sci. 2007;11:500-502. 1994;32:871-878. 42. Sulzer J, Haller S, Scharnowski F, et al. Real-time fMRI neurofeedback: 54. Luciano JV, Algarabel S, Tomas JM, Martinez JL. Development and validation progress and challenges. Neuroimage. 2013;76:396-399. of the thought control ability questionnaire. Pers Individ Dif. 2005;38:997-1008. 43. Subramanian L, Hindle JV, Johnston S, et al. Real-time functional mag- 55. Carver CS, White TL. Behavioral inhibition, behavioral activation, and netic resonance imaging neurofeedback for treatment of Parkinson's dis- affective responses to impending reward and punishment: the BIS/BAS ease. J Neurosci. 2011;31:16309-16317. scales. J Pers Soc Psychol. 1994;67:319-333.

112