Sleep and Psychopathology in the general population and bipolar disorder

Susanne Verkooijen Sleep and Psychopathology in the general population and bipolar disorder

Slaap en Psychopathologie in de algemene bevolking en bipolaire stoornis

(met een samenvatting in het Nederlands)

Proefschrift

Ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag The studies described in this thesis were performed at the Brain Center Rudolf van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit Magnus, Department of Psychiatry, University Medical Center Utrecht and van het college voor promoties in het openbaar te verdedigen op donderdag 21 were funded by the National Institute of Mental Health (NIMH), Grant december 2017 des ochtends te 10.30 uur number: R01 MH090553 to prof. dr. Ophoff.

Copyright © 2017 Susanne Verkooijen door All rights reserved to the author. No part of this publication may be reproduced in any form by any electronic or mechanical means without prior permission of Susanne Verkooijen the author.

ISBN: 978-94-028-0844-5 geboren op 4 september 1987 te Middelburg Printed by: Ipskamp Printing, proefschriften.net Layout and design by: Lara Leijtens, persoonlijkproefschrift.nl Cover design by: Rolf Winia Promotoren: Prof. dr. R.S. Kahn Voor Trijneke Prof. dr. R.A. Ophoff

Copromotor: Dr. M.P.M. Boks Table of Contents

Chapter 1 General introduction 8

Chapter 2 The Dutch Bipolar Cohort, a large study of bipolar disorder I and related phenotypes in 2551 participants: study protocol 22

Chapter 3 An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls 62

Chapter 4 The association of sleep and physical activity with integrity of white matter microstructure in bipolar disorder patients and healthy controls 90

Chapter 5 Sleep disturbances, psychosocial difficulties and health risk behaviour in 16,781 Dutch adolescents 124

Chapter 6 Summary and general discussion 144

Chapter 7 Nederlandse samenvatting 156

List of Publications 172

Curriculum Vitae 174

Dankwoord 176 Chapter 1

General introduction “O sleep, O gentle sleep, Sleep has inspired artists, philosophers and scientists since the dawn of time. Nature’s soft nurse, how have I frightened thee, From dream analyses by the ancient Egyptians, to Aristotle’s theory on sleep, to That thou no more wilt weigh my eyelids down modern polysomnography, sleep has been a continuous subject for debate with And steep my sense in forgetfulness?” both great advocates of the state of sleep and just as many individuals who detest the fact that roughly a third of our lifetime is spent asleep. William Shakespeare (1564-1616)

Although humans have been fascinated by sleep for centuries, to this day the Chapter 1 subject remains surrounded by a cloud of mystery. Researchers have tried to unravel the big question of why we sleep and have come up with a multitude “Sleep is the golden chain that ties health and our bodies together.” of hypotheses, focusing on energy conservation, metabolism, clearance of waste products in the brain, immune restorations and cognitive restructuring Thomas Dekker (1572-1632) which are, most likely, not mutually exclusive (Bryant, Trinder, & Curtis, 2004; Diekelmann & Born, 2010; Kang et al., 2009; Morselli, Guyon, & Spiegel, 2012; Tononi & Cirelli, 2006). What we do know is that sleep is of crucial importance for normal functioning. It is not until one experiences a few “Sleep is an enormous waste of time, a heritance of our cave days.” nights of insomnia, that the true virtue of a good night sleep becomes painfully apparent. Studies showed that acute sleep deprivation leads to impairments in Thomas Edison (1847-1931) postural control, disruptions in a wide range of cognitive processes, metabolic alterations, hormonal changes, impairments in the immune system, changes in brain activity and deterioration in emotional well-being (for a review, see Orzeł- Gryglewska, 2010). Therefore, the increasing prevalence of sleep disturbances “Sleep is for wimps.” in our modern 24-h society is a major public health concern (Ford et al., 2014; Ford, Cunningham, Giles, & Croft, 2015). Margaret Thatcher (1925-2013) The relation between sleep disturbances and health is the reason we set out to investigate the association between sleep and psychopathology. We explore to what extent both factors are interrelated and whether these associations are present in a psychiatric sample and in the general population. In this thesis we focus on two subgroups: bipolar disorder patients and adolescents. As we will

General introduction describe later in this introduction, these populations are characterized by mood

10 11 fluctuations and, as increasing evidence suggests, are vulnerable for disturbances projections on the pineal gland resulting in the secretion of melatonin (Lack & in sleep-wake patterns. In this introductory chapter we will briefly describe the Wright, 2007). Both the homeostatic sleep drive (process S) and the circadian proposed underlying processes that control sleep and propose a model of how regulation of sleep and arousal (process C) interact in such a way that, ideally, sleep and psychopathology may interact. We then describe our sample of bipolar sleep onset follows both a full day of wakefulness and the environmental onset disorder patients and adolescents from the general population in more detail and of night time (Figure 1). explain why focusing on their sleep pattern is of clinical relevance. Chapter 1 Sleep regulation As mentioned previously, there is to this day still no general consensus on the question why we sleep. However, the question how sleep is regulated has become increasingly clear in the past decades. The underlying neurobiological processes of sleep were first described by Borbely (1982). It is thought that the daily shift from wakefulness to sleep is regulated by the interaction of two separate processes, one involving circadian rhythmicity (process C) and one describing Figure 1 – Interaction between homeostatic sleep drive and circadian rhythm of arousal resulting sleep homeostasis (process S). Process S is best described as an hour glass, and in sleep propensity when the distance between S and C is largest (Daan, Beersma, & Borbély, serves as the homeostatic sleep drive, tracking the hours of wakefulness through 1984) increasing levels of adenosine. This results in sleepiness after roughly 16 hours of wakefulness (Strecker et al., 2000). Process C aligns human physiological Interaction between sleep and psychopathology processes to the rotation of earth. The suprachiasmatic nucleus (SCN) in the The high comorbidity between sleep disturbances and psychiatric illnesses hypothalamus, also referred to as the Master Clock, orchestrates endogenous has resulted in a widely held assumption that sleep disturbances are merely a circadian rhythms of approximately 24 hours and 10 minutes (Czeisler et al., symptom of psychiatric disorders (Krystal, 2012). However, more recent studies 1999). To correct for this slight deviation, the SCN is synchronized by so- have focused on several other pathways in which sleep and psychopathology called zeitgebers, meaning environmental cues that help entrain the endogenous are interrelated. Harvey et al. (2011) suggested a theoretical framework for the circadian rhythm to 24 hours. One of the most prominent zeitgebers is light. etiological link between circadian rhythms, sleep disturbances and emotional Photic information, mostly from the blue-green spectrum, is transferred from deregulation that could ultimately lead to psychiatric disorders (Figure 2). retinal ganglion cells to the SCN to provide input regarding the external day Their first assumption (shown in box 2) is a bidirectional relation between sleep and night cycle (Reppert & Weaver, 2002). Non-photic cues, such as social disturbances and emotional deregulation. For one, a wide range of studies interactions, exercise, food consumption and arousal compose a second category conclude that sleep deprivation increases negative mood, such as depression and of zeitgebers that entrain the SCN (Grandin, Alloy, & Abramson, 2006). The anxiety (Orzeł-Gryglewska, 2010). A possible underlying mechanism for this

General introduction SCN regulates daily arousal levels and promotes sleep propensity through its relation is the increased amygdala activity and loss of functional connectivity

12 13 with the medial-prefrontal cortical activity, suggesting impaired regulation of Sleep in bipolar disorder emotional responses following sleep loss (Yoo, Gujar, Hu, Jolesz, & Walker, Bipolar disorder is a psychiatric disorder that has consistently been linked to 2007). Second, high levels of emotional arousal can lead to problems initiating sleep disturbances; during manic episodes 69-99% of patients exhibit a reduced and maintaining sleep, creating a vicious cycle between sleep disturbances need for sleep and during depression both insomnia and hypersomnia are and emotional deregulation (Baglioni, Spiegelhalder, Lombardo, & Riemann, common symptoms (Harvey, 2008). Moreover, challenges to zeitgebers such as 2010). A second assumption (shown in box 1) is that in several psychiatric shift work and jet lags may precipitate mood episodes and sleep disturbances

disorders, this bidirectional relationship between sleep and emotion is are now considered as one of the most prominent prodromal symptoms Chapter 1 affected by disturbances in the underlying neurobiological sleep-wake and (Jackson, Cavanagh, & Scott, 2003). These sleep disturbances have not only circadian processes. It has been suggested that the experience of life events been associated with mood episodes, some studies suggest that they also persist and consequential stress can disrupt social zeitgebers (i.e. daily routines, social during phases of euthymia (Geoffroy et al., 2015; Ng et al., 2014). In addition, interactions) that subsequently disturb circadian rhythmicity resulting in the several lines of evidence suggest that bipolar disorder is also characterized by experience of mood disorders (Grandin et al., 2006). Further circumstantial altered circadian rhythms. Indeed, phase advances, phase delays and general evidence for the involvement of circadian biology in psychopathology comes phase instabilities have been reported for several measures of rhythmicity, from genetic studies that link circadian genes, known to regulate and generate such as body temperature, nocturnal cortisol levels and peak melatonin time circadian rhythms, to several psychiatric disorders (Harvey et al., 2011). in bipolar patients (Cervantes, Gelber, Ng Ying Kin, Nair, & Schwartz, 2001; Examples of circadian genes that have been associated with mood disorders are Nurnberger et al., 2000; Wehr, Goodwin, Wirz-Justice, Breitmaier, & Craig, Timeless, Clock, BMal1, PERIOD2, NPAS2. Furthermore, a recent study by 1982). Hammerschlag et al. (2017) found strong positive genetic correlations between insomnia complaints and anxiety, depressive symptoms, neuroticism and major This raises the question whether sleep disturbances are a state marker of mood depressive disorder. episodes, or rather a trait of bipolar disorder. If bipolar disorder is characterized by disturbances in circadian biology (box 1), then these disturbances could subsequently result in overt problems in both sleep disturbances and emotion regulation (box 2). As a consequence, bipolar patients would show increased sleep disturbances, even during phases of euthymia. Moreover, non-affected siblings who are genetically at risk for developing traits of bipolar disorder, could have a higher incidence of sleep disturbances compared to the general population. If sleep disturbances in bipolar patients are indeed a stable trait, it would be conceivable that they furthermore relate to an unfavourable clinical Figure 2 – Proposed association between sleep and psychopathology. Figure derived from presentation of the disorder, with psychotic symptoms, earlier age at onset and

General introduction Harvey et. al (2011) more mood recurrences.

14 15 Lastly, converging evidence suggests that sleep affects integrity of white matter Given this increased vulnerability for disturbances in circadian biology (box 1) (Bellesi et al., 2013; Cirelli, Faraguna, & Tononi, 2006) and that bipolar during adolescence, we sought to explore the bidirectional relation between sleep disorder is characterized by widespread alterations in integrity of white matter disturbances and emotional deregulation (box 2) by focussing on the association (O’Donoghue, Holleran, Cannon, & McDonald, 2017). Disruptions in white between sleep disturbances and the presence of psychosocial difficulties and matter microstructure have been suggested to possibly underlie the association health risk behaviour. between sleep disturbances and emotional dysregulation (Benedetti et al., 2017).

If so, disturbances in sleep pattern could relate to alterations in white matter Outline of this thesis Chapter 1 microstructure. As of yet, this association has scarcely been investigated in a In this thesis we investigate the association between sleep disturbances and bipolar sample. psychopathology, focusing on both bipolar disorder patients and adolescents. Chapter 2 starts with an overview of the bipolar phenotype and shortly Sleep in adolescents addresses previous findings with regard to circadian rhythmicity in this disorder. Adolescents are the second population of interest vulnerable to sleep Following that, in chapter 3 we address the question whether bipolar patients disturbances, with one in four adolescents acquiring less than 6 hours of sleep, show persisting sleep disturbances during non-clinical phases of the disorder, or and prevalence rates of insomnia between 4-24% (Hysing, Pallesen, Stormark, whether sleep disturbances mostly relate to (subclinical) mood symptomology. Lundervold, & Sivertsen, 2013; Johnson, Roth, Schultz, & Breslau, 2006; These results are then compared to previous findings in a replication of a Ohayon, 2000; Roberts, Roberts, & Duong, 2009). This subgroup of the general recent meta-analysis. Furthermore, we examine if non-affected siblings show population is characterized by a phase delay in the sleep-wake pattern resulting a more disturbed sleep pattern compared to controls. Lastly, we investigate in a delay in bed times. This shift in circadian rhythmicity is not necessarily in this chapter whether the sleep pattern of bipolar patients relates to clinical problematic in itself, but since schools generally do not delay their hours in characteristics of the disorder. Following that, chapter 4 focusses on the relation correspondence with the circadian rhythm of adolescents, the total sleep time is between sleep-activity patterns and integrity of white matter in bipolar patients. negatively affected in many adolescents. The majority of adolescents subsequently More specifically, we investigate if there is an association between sleep-activity increase the total sleep time on weekends to compensate for missed hours of patterns in general and whether we find differences between bipolar patients sleep during week days, which further disrupts their sleep-wake rhythmicity and controls regarding this association. Lastly, in chapter 5 we cross over to on the following week days (Dahl & Lewin, 2002). These developmental the adolescent population to investigate the prevalence of sleep disturbances in characteristics of the adolescent sleep-wake schedule coincide with several this age group. We then address the question to what extent adolescent sleep unfavourable challenges to the circadian rhythms, such as late-night exposure to disturbances relate to psychosocial difficulties and health risk behaviour, taking social zeitgebers, increased blue light from screens, substance use and preferred gender differences and the mediating role of emotional problems into account. autonomy regarding bed times (Carskadon, 2011). General introduction

16 17 Ford, E. S., Wheaton, A. G., Cunningham, T. J., Giles, W. H., Chapman, D. P., & Croft, J. References B. (2014). Trends in outpatient visits for insomnia, sleep apnea, and prescriptions for sleep medications among US adults: findings from the national ambulatory medical care survey 1999-2010. Sleep, 37(8), 1283–1293. Baglioni, C., Spiegelhalder, K., Lombardo, C., & Riemann, D. (2010). Sleep and emotions: A focus on insomnia. Sleep Medicine Reviews, 14(4), 227–238. Geoffroy, P. A., Scott, J., Boudebesse, C., Lajnef, M., Henry, C., Leboyer, M., … Etain, B. (2015). Sleep in patients with remitted bipolar disorders: a meta-analysis of actigraphy Bellesi, M., Pfister-Genskow, M., Maret, S., Keles, S., Tononi, G., & Cirelli, C. (2013). Effects studies. Acta Psychiatrica Scandinavica, 131(2), 89–99. of sleep and wake on oligodendrocytes and their precursors. The Journal of Neuroscience, 33(36), 14288–300. Grandin, L. D., Alloy, L. B., & Abramson, L. Y. (2006). The social zeitgeber theory, circadian

rhythms, and mood disorders: Review and evaluation. Clinical Psychology Review, 26(6), Chapter 1 Benedetti, F., Melloni, E. M. T., Dallaspezia, S., Bollettini, I., Locatelli, C., Poletti, S., 679–694. & Colombo, C. (2017). Night sleep influences white matter microstructure in bipolar depression. Journal of Affective Disorders, 218, 380–387. Hammerschlag, A. R., Stringer, S., Leeuw, C. A. De, Sniekers, S., Taskesen, E., Watanabe, K., … Someren, E. J. W. Van. (2017). Genome-wide association analysis of insomnia Borbely, A. A. (1982). A two process model of sleep regulation. Human Neurobiology, 1(3), 195- complaints identifies risk genes and genetic overlap with psychiatric and metabolic traits, 204. Nature Genetics.

Bryant, P. A., Trinder, J., & Curtis, N. (2004). Sick and tired: does sleep have a vital role in the Harvey, A. G. (2008). Sleep and circadian rhythms in bipolar disorder: Seeking synchrony, immune system? Nature Reviews Immunology, 4(6), 457–467. harmony, and regulation. American Journal of Psychiatry, 165(7), 820–829.

Carskadon, M. A. (2011). Sleep in adolescents: The perfect storm.Pediatric Clinics of North Harvey, A. G., Murray, G., Chandler, R. A., & Soehner, A. (2011). Sleep disturbance as America, 58(3), 637–647. transdiagnostic: consideration of neurobiological mechanisms. Clinical Psychology Review, 31(2), 225–35. Cervantes, P., Gelber, S., Ng Ying Kin, F. N. K., Nair, V. N. P., & Schwartz, G. (2001). Circadian secretion of cortisol in bipolar disorder. Journal of Psychiatry and Neuroscience, Hysing, M., Pallesen, S., Stormark, K. M., Lundervold, A. J., & Sivertsen, B. (2013). Sleep 26(5), 411–416. patterns and insomnia among adolescents: A population-based study. Journal of Sleep Research, 22(5), 549–556. Cirelli, C., Faraguna, U., & Tononi, G. (2006). Changes in brain gene expression after long- term sleep deprivation. Journal of Neurochemistry 98(5), 1632–1645. Jackson, A., Cavanagh, J., & Scott, J. (2003). A systematic review of manic and depressive prodromes. Journal of Affective Disorders, 74(3), 209–217. Czeisler, C. A., Duffy, J. F., Shanahan, T. L., Brown, E. N., Mitchell, J. F., Rimmer, D. W., … Kronauer, R. E. (1999). Stability , precision , and near – 24-hour period of the human Johnson, E. O., Roth, T., Schultz, L., & Breslau, N. (2006). Epidemiology of DSM-IV circadian pacemaker. Science, 284(5423), 2177–2182. insomnia in adolescence: Lifetime prevalence, chronicity, and an emergent gender difference. Pediatrics, 117(2), e247-56. Daan, S., Beersma, D. G., & Borbély, A. A. (1984). Timing of human sleep: recovery process gated by a circadian pacemaker. The American Journal of Physiology, 246(2 Pt 2), R161–R183. Kang, J.-E., Lim, M. M., Bateman, R. J., Lee, J. J., Smyth, L. P., Cirrito, J. R., … Holtzman, D. M. (2009). Amyloid-b dynamics are regulated by orexin and the sleep-wake cycle. Sleep, Dahl, R. E., & Lewin, D. S. (2002). Pathways to adolescent health: Sleep regulation and 326(5955), 1005–1007. behavior. Journal of Adolescent Health, 31(6 SUPPL.), 175–184. Krystal, A. D. (2012). Psychiatric Disorders and Sleep. Neurologic Clinics, 30(4), 1389–1413. Diekelmann, S., & Born, J. (2010). The memory function of sleep. Nature Reviews Neuroscience, 11(2), 114–126. Lack, L. C., & Wright, H. R. (2007). Chronobiology of sleep in humans. Cellular and Molecular Life Sciences, 64(10), 1205–15. Ford, E. S., Cunningham, T. J., Giles, W. H., & Croft, J. B. (2015). Trends in insomnia and excessive daytime sleepiness among US adults from 2002 to 2012. Sleep Medicine, 16(3), General introduction Morselli, L. L., Guyon, A., & Spiegel, K. (2012). Sleep and metabolic function. Pflugers Archiv 372–378. European Journal of Physiology, 463(1), 139–160.

18 19 Ng, T. H., Chung, K. F., Ho, F. Y. Y., Yeung, W. F., Yung, K. P., & Lam, T. H. (2014). Sleep- wake disturbance in interepisode bipolar disorder and high-risk individuals: A systematic review and meta-analysis. Sleep Medicine Reviews, 20, 46–58.

Nurnberger, J. I., Adkins, S., Lahiri, D. K., Mayeda, A., Hu, K., Lewy, A., … Davis-Singh, D. (2000). Melatonin suppression by light in euthymic bipolar and unipolar patients. Archives of General Psychiatry, 57, 572–579.

O’Donoghue, S., Holleran, L., Cannon, D. M., & McDonald, C. (2017). Anatomical dysconnectivity in bipolar disorder compared with schizophrenia: A selective review of

structural network analyses using diffusion MRI. Journal of Affective Disorders, 209, 217–228. Chapter 1

Ohayon, M. M. (2000). Prevalence of hallucinations and their pathological associations in the general population. Psychiatry Research, 97(2–3), 153–164.

Orzeł-Gryglewska, J. (2010). Consequences of sleep deprivation. International Journal of Occupational Medicine and Environmental Health, 23(1), 95–114.

Reppert, S. M., & Weaver, D. R. (2002). Coordination of circadian timing in mammals. Nature, 418(6901), 935–941.

Roberts, R. E., Roberts, C. R., & Duong, H. T. (2009). Sleepless in adolescence: Prospective data on sleep deprivation, health and functioning. Journal of Adolescence, 32(5), 1045–1057.

Strecker, R. E., Morairty, S., Thakkar, M. M., Porkka-Heiskanen, T., Basheer, R., Dauphin, L. J., … McCarley, R. W. (2000). Adenosinergic modulation of basal forebrain and preoptic/ anterior hypothalamic neuronal activity in the control of behavioral state. Behavioural Brain Research, 115(2), 183–204.

Tononi, G., & Cirelli, C. (2006). Sleep function and synaptic homeostasis. Sleep Medicine Reviews, 10(1), 49–62.

Wehr, T., Goodwin, F., Wirz-Justice, A., Breitmaier, J., & Craig, C. (1982). 48-hour sleep- wake cycles in manic-depressive illness: naturalistic observations and sleep deprivation experiments. Archives of General Psychiatry, 39(5), 559–565.

Yoo, S. S., Gujar, N., Hu, P., Jolesz, F. a., & Walker, M. P. (2007). The human emotional brain without sleep - a prefrontal amygdala disconnect. Current Biology, 17(20), 877–878. General introduction

20 21 Chapter 2

The Dutch Bipolar Cohort, a large study of bipolar disorder I and related phenotypes in 2551 participants: study protocol

Sanne Verkooijen, Annabel Vreeker, Lucija Abramovic, Annet H. van Bergen, Diane Ramakers, Anil Ori, Timothy Wu, Loes Olde Loohuis, Manon H.J. Hillegers, Annet T. Spijker, Erik Hoencamp, Eline J. Regeer, Rixt F. Riemersma-Van der Lek, Robert A. Schoevers, Stefan E. Knapen, Anja W.M.M. Stevens, Peter F.J. Schulte, Ronald Vonk, Rocco Hoekstra, Nico J.M. van Beveren, Ralph Kupka, René S. Kahn , Marco P. M. Boks, Roel A. Ophoff Abstract bipolar disorder in combination with extensive genetic information provides a great opportunity to study a wide range of biological, genomic and epidemiologic Background The Dutch Bipolar Cohort (DBC) is a large, homogeneous risk factors of bipolar disorder. sample of bipolar I patients, their first degree relatives and healthy controls from the Netherlands. The study is primarily designed as a deep-phenotype characterization of bipolar I patients and their first-degree relatives and includes a wide range of (neuro)psychological and biological measures. The current paper presents the recruitment strategies, assessment methods and cohort characteristics. Chapter 2

Method A cross-sectional cohort of bipolar disorder I patients, their first degree relatives and controls was recruited in the Netherlands between June 2011 and April 2015. Individual assessment lasted several hours per subject and included collection of biomaterials (blood samples, hair sample and skin biopsies). All participants were extensively phenotyped using a clinical interview, cognitive tests and psychological questionnaires. Subgroups participated in 3T MRI scannings and actigraphy measurements.

Results A total of 1695 bipolar I patients, 590 relatives and 266 controls were included. Patients had a mean age of 49.8 years (sd = 12.3), relatives were on average 56.0 years old (sd = 15.2), and the controls were 47.5 years old (sd = 15.4). Of the patients, 603 (43.2%) were male, compared to 215 (36.5%) in relatives and 131 (49.2%) in controls. Blood samples were obtained in all participants. Hairs were collected in 1318 participants and skin biopsies were acquired in 187 participants. A subgroup of 437 participants underwent 3T MRI-scanning. Actigraphy measurements were obtained in 261 participants.

Discussion The DBC study is one of the largest cohorts and stands out for its comprehensive phenotyping and inclusion of a large number of family

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol members. The detailed information regarding symptomatology and course of

24 25 Introduction Family and twin studies first described the aggregation of bipolar disorder within families and point to a genetic contribution to bipolar disorder risk Bipolar disorder is a chronic psychiatric disorder characterized by episodes (Barnett & Smoller, 2009). According to family studies, a first-degree relative of severe fluctuations in mood, ranging from depression to (hypo)mania. of a bipolar disorder patient has an increased lifetime prevalence rate of 5-13 Prevalence rates are estimated between 1-2% of the general adult population and %, compared to the general population rate of 1-2% (Craddock & Sklar, 2013; onset is generally in early adulthood (Belmaker, 2004; Müller-Oerlinghausen, Mesman, Nolen, Reichart, Wals, & Hillegers, 2013; Taylor, Faraone, & Berghöfer, & Bauer, 2002). Although most patients experience periods in which Tsuang, 2002). Twin studies further showed that this increased risk can not mood symptoms reach subclinical levels or completely subside, many patients only be contributed to a shared environment: concordance rates of monozygotic never regain their normal level of daily functioning and require lifelong health twins are between 40 – 80% compared to a concordance rate between 1-8 %

care (Sajatovic, 2005). Mortality rates of bipolar patients are estimated to be two in dizygotic twins. Heritability estimates (the proportion of variance in risk Chapter 2 to three times higher than that of the general population (Müller-Oerlinghausen that is explained by genetic variation) range from 79-93% (Barnett & Smoller, et al., 2002). According to the Global Burden of Disease 2000 study, bipolar 2009). Given that most of the genetic variance in liability to mania is specific disorder is in the top ten causes of disability worldwide, and has around the to the manic syndrome, bipolar I disorder is thought to be the most heritable of same percentage of Years Lived with Disability as chronic obstructive lung all bipolar subtypes (McGuffin et al., 2003). Genome wide association studies conditions (Lopez & Murray, 1998). (GWAS) of bipolar disorder have identified single-nucleotide polymorphisms (SNPs) at a small number of loci that confer risk to bipolar disorder. These risk Bipolar disorder is divided into four subtypes. Bipolar disorder I classification alleles have small effect sizes, confirming the idea that bipolar disorder is a requires at least one full blown manic episode which is often, but not necessarily, polygenic trait (Kerner, 2014). However, as of yet, only a small percentage of followed or preceded by depressive episodes. Psychotic experiences such as variance can be explained by the identified alleles, resulting in a large part of the hallucinations and delusions frequently occur during episodes. Bipolar disorder heritability still unaccounted for (Gershon, Alliey-Rodriguez, & Liu, 2011). A II is characterized by less impairing hypomanic episodes along with depressive useful tool to analyse genetic loading in disorders that include a large number of episodes. A third subtype, cyclothymia, is diagnosed by subclinical depressive causal genetic variants, is the use of polygenic risk scores. To calculate individual symptoms not meeting the criteria for a depressive episode, alternating with polygenic risk scores in a sample, first a GWAS is performed in a ‘discovery hypomanic episodes. Lastly, bipolar disorder ‘not otherwise specified’ (NOS) is sample’, from which SNPs are selected (e.g. exceeding a certain threshold). used for patients with subclinical manic and depressive symptoms, that do not Individuals from a new ‘target sample’ receive a score that is composed of the meet the full diagnostic criteria for bipolar disorder or depression (American weighted sum of these risk alleles (Dudbridge, 2013). This way, loci with no Psychiatric Association, 2013). Of all subtypes, bipolar disorder I is considered separate effect on disease risk may together significantly predict disease status to be the most clinically distinct. (Wray, Goddard, & Visscher, 2007). Furthermore, polygenic risk score analysis allows for genotype-phenotype associations, as the polygenic risk scores can be

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol linked to any phenotype outcome measure. Polygenic risk score analysis was first

26 27 performed in a large schizophrenia sample and provided support for a polygenic of deep phenotyping involves the elaborate collection of biological, clinical basis (Purcell et al., 2009). However, polygenic analyses based on scores from a and (neuro)psychological information which can subsequently be associated bipolar discovery sample have not yet been performed. with genetic variation. The emphasis on collecting precise and comprehensive phenotypic data is becoming more widespread in regular medicine and One of the challenges for psychiatric genetic research is the heterogeneity of neuropsychiatric research (Robinson, 2012). One of such efforts is the Research disorders. As a result of psychiatry’s diagnostic systems, disorders can consist of Domain Criteria (RDoC) matrx initiated by the National Institute of Mental multiple combinations of symptoms. Although efforts have been made to improve Health (NIMH). This matrix provides a new framework for future research in homogeneity by introducing subtypes of disorders, variation in symptoms and neuropsychiatry and focusses entirely on the collection of data on a symptom course of illness still results in large within-group differences (Wardenaar & de level (Insel et al., 2010). Combining data from GWAS analyses, whole genome

Jonge, 2013). Another challenge of the psychiatric phenotypes is the similarity sequencing and polygenic risk scores with phenotype data, allows for a more Chapter 2 between clinical diagnoses. Several key symptoms are not unique to bipolar direct analysis of underlying mechanisms. The focus of the Dutch Bipolar disorder but are shared with other mood and/or psychotic disorders, such as Cohort (DBC) study is therefore to recruit a large homogeneous cohort of major depressive disorder and schizophrenia (Maier, Zobel, & Wagner, 2006; bipolar I patients, first degree relatives and controls for a collection of genome- Thomas, 2004). This symptomatic overlap between disorders has further been wide genetic, (neuro)psychological and biological measures. corroborated by genetic similarities between bipolar disorder and several other psychiatric disorders. Increased rates of unipolar depression and schizophrenia Phenotypes of bipolar disorder have been found in first-degree relatives of bipolar I patients (Barnett & Smoller, Neurocognitive functioning 2009; Mesman et al., 2013; Thomas, 2004). In 2013, additional evidence for this Several studies have demonstrated that during phases of acute illness, bipolar overlap came from the largest psychiatric genome-wide analysis to date by the patients show cognitive deficits on multiple domains, such as attention, memory, Psychiatric Genome Consortium (PGC). SNPs at several loci, most of whom executive functioning and psychomotor speed (Bearden, Hoffman, & Cannon, were previously associated to just one disorder, were associated to five major 2001). Also, general intellectual functioning (IQ ) has been found to be lower in psychiatric disorders (i.e. bipolar disorder, schizophrenia, major depressive manic and depressed patients compared to controls (Martínez-Arán et al., 2000; disorder, attention-deficit, hyperactivity disorder and autism-spectrum disorder) Quraishi & Frangou, 2002). Whether these cognitive deficits remain as stable (Smoller et al., 2013). The challenges of heterogeneity within the patient over time as seen in schizophrenia patients is less evident. Several cognitive sample and the overlapping psychiatric definitions between disorders may domains appear to be impaired during phases of euthymia. A meta-analysis by have hampered the identification of susceptibility genes underlying psychiatric Bora et al. (Bora, Yucel, & Pantelis, 2009) reported underperformance on tasks disorders, among which bipolar disorder (Gottesman & Gould, 2003). that measured executive functioning and verbal memory in euthymic patients. These findings could not be fully attributed to medication effects. Studies on To overcome these problems, the concept of deep phenotyping can help IQ in asymptomatic patients are inconclusive, with some studies reporting

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol strengthen the genotype-phenotype association (Delude, 2015). The method deficits (Eric, Halari, Cheng, Leung, & Young, 2013; McIntosh, Harrison,

28 29 Forrester, Lawrie, & Johnstone, 2005; Toulopoulou, Quraishi, McDonald, & 1999), although smaller amygdala volumes have also been reported (DelBello, Murray, 2006), while others report IQ levels comparable to controls (Glahn, Zimmerman, Mills, Getz, & Strakowski, 2004; Hibar et al., 2016). An Bearden, Bowden, & Soares, 2006; Robinson et al., 2006). These contradictory important limitation of most structural neuroimaging studies are the relatively findings may be the result of methodological differences in IQ measurements small sample sizes. Furthermore, heterogeneous samples comprised of several and the inclusion of different subtypes of bipolar patients. Possible cognitive bipolar subtypes could have contributed to the conflicting results, given that impairments have also been studied in first-degree relatives and twins of bipolar these subtypes may differ regarding their neuroanatomical characteristics. patients. Results indicate that these genetically susceptible individuals show Another important issue in neuroimaging studies of bipolar disorder is the similar impairments on executive functioning and memory, although effect use of psychotropic medication. Several commonly prescribes drugs, such as sizes are generally smaller compared to the results from the bipolar samples lithium and antipsychotic medication are hypothesized to affect brain volumes

(Arts, Jabben, Krabbendam, & van Os, 2008; Bora et al., 2009; Christensen, (Hafeman, Chang, Garrett, Sanders, & Phillips, 2012; Ho, Andreasen, Ziebell, Chapter 2 Kyvik, & Kessing, 2006; Vreeker et al., 2016). Considering the significant effect Pierson, & Magnotta, 2011; Moncrieff & Leo, 2010; Navari & Dazzan, 2009; of cognition on functioning after episodes, it is of paramount importance to Scheepers et al., 2001). However, delineating the specific effects per medication understand whether cognitive dysfunction reflects a stable trait of the disorder or type from the direct illness effects has proven to be a challenging task. As of only a transient symptom solely occurring during mood episodes. We therefore yet, no single centre study has used a bipolar sample of more than a hundred measure both premorbid and current cognitive functioning to study whether participants with strict diagnostic inclusion criteria. Therefore, high resolution euthymic bipolar patients show underperformance compared to controls and 3T-scans are acquired in the largest sample of bipolar I patients to date, while whether relatives show similar deficiencies indicative of familial vulnerability. simultaneously addressing the issue of medication use.

Neuroanatomy Circadian Rhythms Accumulating evidence suggests that bipolar disorder is accompanied by Several lines of evidence show altered circadian rhythmicity in bipolar disorder brain abnormalities, including white matter hyperintensities and ventricular (Hasler, Drevets, Gould, Gottesman, & Manji, 2006). Phase advances in body enlargements (Hallahan et al., 2011; Kempton, Geddes, Ettinger, Williams, & temperature, nocturnal cortisol levels and peak melatonin time have previously Grasby, 2008; Lyoo, Lee, Jung, Noam, & Renshaw, 2002; Pillai et al., 2002). been identified in bipolar patients, supporting the idea that the endogenous 24- White matter hyperintensities may contribute to the pathophysiology of bipolar hour cycle is altered in bipolar patients (Milhiet, Etain, Boudebesse, & Bellivier, disorder by interrupting connecting structures involved in mood regulation 2011; Nurnberger et al., 2000). Moreover, sleep-wake disturbances were found (Mahon, Burdick, & Szeszko, 2010). The anomalies have also been found to to be one of the most evident prodromal symptoms and have also been reported be prevalent in white matter and subcortical grey nuclei of unaffected relatives to persist during phases of euthymia (Bauer et al., 2006; Harvey, Schmidt, (Gulseren, Gurcan, Gulseren, Gelal, & Erol, 2006; Tighe et al., 2012). In Scarnà, Semler, & Goodwin, 2005; Jackson, Cavanagh, & Scott, 2003). It has addition, several studies have observed enlarged amygdala structures in bipolar been hypothesized that one of the therapeutic effects of lithium is lengthening

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol patients (Altshuler et al., 2000; Brambilla et al., 2003; Strakowski et al., the circadian phase (Johnsson, Pflug, Engelmann, & Klemke, 1979), and other

30 31 treatment strategies that act on circadian rhythmicity, such as sleep deprivation However, given the small sample size, this promising study requires replication and social rhythm therapy, have proven to be successful treatments of mood in a larger set. With the inclusion of unaffected relatives, the possibility of these episodes (Gonzalez, 2014). altered expression patterns as an endophenotype of bipolar disorder can be further investigated. The main circadian pacemaker consists of transcriptional-translational feedback loops that cause rhythmic expressions of clock genes, ultimately leading to Metabolic functioning the temporal organization of peripheral rhythms of the body (Takahashi et Another promising endophenotype is metabolic functioning as increasing al., 2008). Mice carrying a mutation in the CLOCK-gene display mania-like evidence suggests that it is affected in bipolar patients (McIntyre et al., 2010; behaviour, including hyperactivity, decreased sleep and enhanced reward Sicras, Rejas, Navarro, Serrat, & Blanca, 2008). Bipolar disorder has consistently

sensitivity. This behaviour seems reversible after chronic administration of been linked with overweight and (central) obesity, in both medicated as well Chapter 2 lithium (Monteleone & Maj, 2008). Polymorphisms in clock genes have also as drug naïve patients (Maina, Salvi, Vitalucci, D’Ambrosio, & Bogetto, been identified in bipolar patients and were associated with specific disease 2008). Hypertension and dyslipidaemia (e.g. increased cholesterol levels and characteristics such as age at onset (Benedetti et al., 2007, 2008; Mansour et al., hypertriglyceridemia) has also been found to be more prevalent in bipolar 2009; Nievergelt et al., 2006; Shi et al., 2008). These results came from studies patients compared to the general population (Fagiolini, Frank, Scott, Turkin, using a candidate-gene approach. However, GWAS-studies have so far failed to & Kupfer, 2005; Goldstein, Fagiolini, Houck, & Kupfer, 2009; Gurpegui replicate the association between clock genes and bipolar disorder (McCarthy, et al., 2012; Johannessen, Strudsholm, Foldager, & Munk-Jørgensen, 2006; Nievergelt, Kelsoe, & Welsh, 2012). The question whether this is due to the McElroy & Keck, 2012). According to the current working definition by the statistical power of previous GWAS-studies, or to a lack of genetic association International Diabetes Federation, the overlapping metabolic syndrome consists between the circadian timing system and bipolar disorder, remains unanswered. of i) central obesity, ii) hypertriglyceridemia, iii) elevated blood pressure, iv) low high-density-lipoproteïne (HDL) levels and v) elevated fasting glucose (Alberti, Besides studying circadian rhythm as an endophenotype using genome wide Zimmet, & Shaw, 2005). A recent meta-analysis found the odds for metabolic association, genes identified in the GWAS can be studied as an in an in vitro syndrome to be almost twice as high for bipolar patients compared to age- and model for altered circadian rhythm in bipolar disorder. Several studies have gender-matched control subjects (Vancampfort et al., 2013). Moreover, one shown that circadian oscillators in fibroblasts are similar to those operative in study has reported an increased prevalence of bipolar disorder in first-degree the brain as they originate from the same embryonic layer as neurons. Circadian relatives of morbidly obese patients, suggesting shared genetic risk factors expression patterns of clock genes in fibroblasts can therefore serve as a valid (Black, Goldstein, Mason, Bell, & Blum, 1992). However, whether there is a in vitro model for molecular oscillators in the brain (Pagani et al., 2010). So clear genetic link between bipolar disorder and metabolic syndrome has yet to far, only one study examined cultured fibroblasts of bipolar patients and found be elucidated. Hence, in all participants overt physical fitness phenotypes, such rhythmic expression patterns in several clock genes, CLOCK, BMAL1, PER3 as height, weight and blood pressure are measured, and biochemical markers of

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol and TIMELESS (Yang, Van Dongen, Wang, Berrettini, & Bućan, 2009). metabolism, such as triglyceride and HDL levels, are determined from plasma.

32 33 This detailed picture of metabolic functioning combined with GWAS-data can healthy controls from the Netherlands. The primary objective is to collect a help illuminate whether altered metabolic functioning in bipolar disorder indeed deep phenotype characterization of the cohort. This includes an extensive reflects a genetic component. analysis of clinical, circadian, neuroanatomical, endocrine, and metabolic characteristics of the participants, which are hypothesized as possible affected Regulation of Stress Hormones phenotypes of bipolar I disorder. Moreover, genome-wide genotype data In several mood disorders, among which bipolar disorder, the Hypothalamic and genome-wide expression data are collected to identify genetic variation Pituitary Adrenocortical (HPA) axis function is altered (de Kloet, Joëls, & involved in susceptibility of bipolar I disorder. Genetic data from this study will Holsboer, 2005). In bipolar patients, the adaptational autoregulatory response complement the already existing repository for genome-wide association studies of the HPA axis is changed (Daban, Vieta, Mackin, & Young, 2005), and these from the National Institute of Mental Health (NIMH) and can be added to

responses have been related to genetic variation in the stress response genes the Psychiatric Genetics Consortium (PGC), allowing for mega-analyses of Chapter 2 (DeRijk & de Kloet, 2008; Spijker et al., 2011; Spijker et al., 2009). Changes genetic data. Both the NIMH repository and the PGC are aimed at collectively in HPA functioning have not only been found in patients diagnosed with a studying the genetic aetiology of psychiatric disorders, with the idea that mood disorder, but also, to a lesser extent, in non-affected relatives (Langan & independent studies are generally underpowered to find replicable results, but McDonald, 2009; Watson, Gallagher, Ritchie, Ferrier, & Young, 2004). In the that combining and sharing of data will yield more robust findings. Ultimately, present study, a novel parameter is used to measure cortisol and other endocrine characteristics from deep phenotype analyses can be studied for their association parameters, which involves high pressure liquid chromatography (HPLC) in with previously identified loci to put the earlier GWAS findings in a broader scalp hair (Kirschbaum, Tietze, Skoluda, & Dettenborn, 2009; Manenschijn, perspective. The combination of genetic and phenotype data can facilitate the Koper, Lamberts, & Van Rossum, 2011; Sauvé, Koren, Walsh, Tokmakejian, interpretation of genetic discoveries to unravel the biological underpinnings of & Van Uum, 2007; Van Uum et al., 2008). This provides the opportunity to bipolar disorder. measure long-term cortisol levels (reflecting mean levels of the past months) in a non-invasive way. Also, contrary to other cortisol assessments (e.g. in saliva or Methods urine), HPLC in scalp hair does not reflect a state-measurement complicated by circadian fluctuations or external circumstances, such as stress or infection Study design (Manenschijn et al., 2011). Simultaneously analysing long-term cortisol action DBC is a case-control study, carried out by the University Medical Centre in relation to genetic background enables studying the complete picture of Utrecht (UMCU), in collaboration with the University of California Los cortisol exposure at the tissue level in bipolar disorder. Angeles (UCLA) and funded by the NIMH. The team of DBC-investigators consisted of scientific leaders from the following Dutch health care institutes: Objective Altrecht (Utrecht), PSYQ Haaglanden (Den Haag), Parnassia BAVO The aim of the Dutch Bipolar Cohort (DBC) study is to recruit a large, Group (Den Haag), GGZ Ingeest (Amsterdam), Dimence (Almelo), Mental

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol homogeneous cohort of bipolar I patients, their first-degree relatives and Health Services North-Holland-North (Heerhugowaard), Reinier van Arkel

34 35 (‘s Hertogenbosch), Erasmus University Medical Centre (Rotterdam), the between excluded and included controls was the higher rate of anxiety disorders University Medical Centre Groningen (UMCG) and UMCU (Utrecht). in the excluded group (χ2 [1, 293] = 9.27, p < 0.01). Inclusions took place between June 2011 and April 2015. The study protocol was

approved and monitored by the Ethical Review Board of the UMCU and by the Table 1 - In- and exclusion criteria

local review board of the UMCG. Patients Relatives Controls Dutch ancestry Dutch ancestry Dutch ancestry Inclusion and Exclusion Premorbid IQ > 80 Premorbid IQ > 80 Premorbid IQ > 80 > 18 years old > 18 years old > 18 years old Table 1 shows the in- and exclusion criteria. Dutch ancestry was defined as Good command of Dutch language Good command of Dutch language Good command Dutch language having at least three grandparents of Dutch descent. A diagnosis of bipolar BP I diagnosis 1st degree relative with BP I diagnosis No psychotic or BP diagnosis

disorder was not an exclusion nor an inclusion criteria for relatives. However, No compulsory treatment under Relatives may be diagnosed with or No 1st or 2nd degree relative Chapter 2 governmental mental health act without BP themselves with psychotic or BP diagnosis controls with a life-time diagnosis of any type of bipolar disorder or psychotic disorder were excluded. Any other type of psychiatric disorder was not

an exclusion criteria in order to avoid ‘super-normal controls’, that are not Table 2 - Excluded participants after assessment

comparable with the general population. Patients Relatives Controls No BP I diagnosis N=171 Proband excluded N=22 Bipolar or psychotic diagnosis N=9 A total of 231 participants were excluded after the original assessment (see table Major somatic illness N=3 Not related to proband N=2 1st / 2nd degree relative with disorder N=17 Withdrew consent N=1 Unreliable assessment N=1 Foreign (grand)parents N=1 2). Of the patients that were excluded due to aberrant diagnosis, 86 patients were diagnosed with bipolar disorder type II, 11 with bipolar disorder NOS, 25 with depressive disorder and 23 with schizoaffective disorder. Comparisons Recruitment between in- and excluded patients and demographic differences were analysed All participants were recruited in the Netherlands. The Netherlands has the using Analysis of variance (ANOVA) with post-hoc Bonferroni corrections for unique quality for genetic research that although the geographical area is all continuous measures and chi-square analyses were used for all categorical small, the population density is relatively high (Boomsma et al., 2014). Also, measures. Compared to the included patients, excluded patients were older (F immigration has been low over the past two centuries, all contributing to [5,2472] = 27.50, p < 0.001) and had lower levels of education (χ2 [5, 1536] = a relatively homogeneous population (Francioli et al., 2014). An additional 19.10, p < 0.01). Also, not surprisingly given the high rate of exclusion due to a advantage is that medical care is generally accessible and the infrastructure is non-bipolar diagnosis, these patients were more often not using lithium (χ2 [1 advanced, making it possible to include patients from all regions of the country. , 991] = 271.10, p < 0.001). Excluded patients were more frequently recruited Moreover, previous psychiatric population studies in the Netherlands have also through their pharmacy (χ2 [7, 1558] = 44.81, p < 0.001). Included and excluded resulted in large cohorts showing a high willingness to participate in scientific relatives showed no differences in characteristics and the only difference research (Korver et al., 2012; Penninx et al., 2008). The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol

36 37 With a life time prevalence of 1.3% in the general population between the may be used for future scientific research. Also, a cohort of 500 genotyped age of 18-64 and a 12-month prevalence of 1.1%, it is estimated that the total and phenotyped healthy control subjects was already available from a previous number of Bipolar I patients in the Netherlands lies around 160,000. In order GWAS on schizophrenia (Stefansson et al., 2009). to reach these patients, the UMCU collaborated closely with a wide range of clinical centres, health care professionals and the Dutch Foundation for See figure 2 for a chart flow from the total number of participants approached Bipolar Disorders. Also, the Dutch bipolar disorder advocacy group, repeatedly for participation to the final number of participants included. informed their members about the possibility of participating in the study. Clinical patients were approached by their treating physician, but informed consent was obtained by the study team, independent from treatment and

fully according Good Clinical Practice and Helsinki guidelines. An important Chapter 2 strategy to contact bipolar patients was through pharmacy network UPPER, in which pharmacy practitioners collaborate with scientific researchers. Through this network, a total of 320 pharmacies from various regions of the Netherlands invited their patients who had received a lithium prescription in the previous 5 years to contact the study team. Since lithium is primarily used in the treatment of bipolar disorder, lithium use is an excellent method to screen for bipolar patients. On average, 20% of the notified patients volunteered to participate in the study. In addition, patients and healthy controls were recruited through advertisements in local and national media and the Internet. Patients and controls who had previously participated in psychiatric studies at the UMCU Figure 1 - Chart pies of the various recruitment strategies were invited to participate again. Lastly, patients were asked to contact their first-degree relatives to ask their willingness to participate in the study. Figure Procedure 1 shows the relative contribution of the various recruitment strategies. After The assessments were administered by a group of carefully trained staff members eligible participants had been informed of the study procedure, they were and consisted of psychology and medical students, PhD-students, research screened for the in- and exclusion criteria by telephone. A few criteria, such as assistants and residents in psychiatry under the supervision of experienced the level of premorbid IQ , could only be objectively assessed as part of the study psychiatrists. Training consisted of a one-day training workshop to practice procedure. all measurements of the study protocol. Furthermore, new interviewers were supervised during at least 3 assessments by experienced supervisors, Additional participants came from previous studies that had included bipolar I before independently interviewing participants. During every assessment, all

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol patients. These participants had given signed consent that their data and samples interviewers had to report to one of the supervisors for consent diagnosis.

38 39 The assessments took place at the UMCU or UMCG. To reduce the travelling interview, assessment of physical measures and neuropsychological testing. time for participants, the study team also visited the following health care Furthermore, all participants were asked to complete an online questionnaire institutes for regional inclusions: PsyQ (Den Haag), Reinier van Arkel (Den containing additional psychological assessments. Bosch and Vught), Dimence (Deventer), Antes Mental Health Care (Rotterdam and Poortugaal), St. Annaziekenhuis (Geldrop). In 32 cases, in-home After completion of the first assessment, a subsample of participants was assessments were provided. The session for patients lasted on average three to reproached for sMRI scanning and actigraphy measurements. four hours and the session for relatives and controls lasted around two hours. To minimize the effect of fatigue, the session of the patients included one break of Measurements thirty minutes and all participants were allowed short breaks as required. Before Phenotype assessment

proceeding with any of the assessments, participants gave a written informed For a complete overview of questionnaires used during the phenotype Chapter 2 consent. The session was roughly divided into four parts: blood sampling, clinical assessment, see table 3. The patient assessment started with an inquiry about the current level of depressive and manic symptoms, using the IDS-SR (Rush et al., Approached for participation 2000) and ASRM (Altman et al., 1997). These self-report questionnaires were Patients Relatives Controls Total used to quickly give an indication of the presence of a mood episode in the past

N = 3364 N = 884 N = 409 N = 4647 week. If this was the case, the session was terminated if the interviewer deemed it to be necessary for the wellbeing of the patient. The subsequent clinical interview consisted of administration of the SCID (First et al., 1997), in which Clinical Assessment the diagnosis of bipolar disorder I was verified. This semi-structured interview Patients Relatives Controls Total was also used to assess the presence of rapid cycling, seasonality and comorbid N = 1575 N = 614 N = 293 N = 2482 anxiety disorders. Psychotic symptoms were evaluated with both the SCID and the psychosis section of the CASH (Andreasen et al., 1992). The QBP Assessments remaining after exclusion Samples from previous studies (Leverich et al., 2001; Suppes et al., 2001) was administered to evaluate bipolar

Patients Relatives Controls Total Patients Relatives Controls Total illness characteristics, and the CIDI (Robins et al., 1988) sections B, J and L,

N = 1401 N = 585 N = 265 N = 2251 N = 299 N = 1 N = 0 N = 300 were administered to assess lifetime and current substance and alcohol abuse. A comprehensive list of current and lifetime medication was used, along with the question whether the medication history could be requested from patients’ Total included in the study pharmacy. Also, a questionnaire regarding satisfaction with lithium use was Patients Relatives Controls Total administered. Lastly, the FIGS (Maxwell, 1992) contained questions regarding N = 1700 N = 586 N = 265 N = 2551 the presence of psychiatric disorders in first or second degree relatives. The The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol

Figure 2- Recruitment flow of participants included in the DBC study

40 41 Table 3 - Collected clinical, neuropsychological and biological measurements Table 3 - Continued

Phenotype Assessment Patients Relatives Controls Phenotype Assessment Patients Relatives Controls

Migraine screener Lumina(van Oosterhout et al., 2011) 1327 573 260 Clinical Assessment Inventory of Depressive Symptoms-Self Report (IDS-SR)(Rush, 1351 - - Lithium side-effect list 1003 412 222 Carmody, & Reimitz, 2000) Paternal and maternal age 1311 565 259 Altman Self-Rating Mania Scale (ASRM)(Altman, Hedeker, Peterson, 1353 - - Social economic status(Currie et al., 2008) 1325 573 260 & Davis, 1997) Medical consumption 1298 564 257 Structural Clinical Interview for DSM IV (SCID)(First, Spitzer, 1376 - - Gibbon, & Williams, 1997) Endophenotype Assessment Patients Relatives Controls Comprehensive Assessment of Symptoms and History (CASH) 1364 - - (Andreasen, Flaum, & Arndt, 1992) Biomaterials Questionnaire for Bipolar Illness (QBP)(Leverich et al., 2001; Suppes et 1362 - - Paxgene blood sample 1395 589 266 al., 2001) EDTA blood sample 1394 589 266 Chapter 2 Composite International Diagnostic Interview (CIDI)(Robins et al., 1037 - - Serum blood sample 808 315 141 1988) Cell lines blood sample 1279 - - List of current and life-time medication 1343 - - 50 mg of scalp hairs 811 353 154 Lithium satisfaction questionnaire 1334 - - Family Interview for Genetic Studies (FIGS)(Maxwell, 1992) 1303 548 248 Physical assessment

Mini-International Neuropsychiatric Interview (M.I.N.I)(Sheehan et - 588 265 Height 1085 434 226 al., 1998) Weight 1079 434 224 Mental Health Screening form (MHSF)(Caroll & McGinley, 2001) - 587 266 Blood pressure 964 422 224

Online Assessment Neuropsychological assessment Community Assessment of Psychic Experiences (CAPE)(Stefanis et al., 1307 564 257 2002) Wais subtask ‘Information’(Wechsler, 1997) 1120 580 257 Symptoms checklist-90 (SCL-90)(Derogatis & Cleary, 1977) 1318 565 259 Wais subtaks ‘Block design’(Wechsler, 1997) 1113 582 256 State-trait anxiety Inventory Medical Consumption (STAI)(Spielberger, 1299 564 257 Wais subtask ‘Arithmetic’(Wechsler, 1997) 1119 582 257 Gorsuch, & Lushene, 1970) Wais subtask ‘Digit Symbol Coding’(Wechsler, 1997) 1118 580 255 Schizotypical Personality Questionnaire (SPQ)(Raine, 1991) 1311 566 259 National Adult Reading Test (NART)(Bright, Jaldow, & Kopelman, 1395 588 266 Everyday Problem Checklist (EPCL)(Vingerhoets & Tilburg, 1994) 1324 571 260 2002) Childhood Trauma Questionnaire (CTQ)(Bernstein, Ahluvalia, Pogge, 1324 571 260 Circadian rhythm assessment & Handelsman, 1997) Life Stressor Checklist-Revised (LSC-R) (Wolfe, Kimerling, Brown, 1325 572 260 14 days actigraphy measurements 107 74 80 Chrestman, & Levin, 1996) Fibroblasts 87 45 56 Neuroticism-Extraversion-Openness Five Factor Inventory (NEO-FFI) 1303 564 257 (McCrae & Costa, 2004) Neuroimaging (structural 3T-MRI) Peters et al. Delusion Inventory (PDI)(Peters, Joseph, & Garety, 1999) 84 24 - High resolution T1-weighted scan 266 - 171 Beliefs about Medicines Questionnaire (BMQ)(Horne, Weinman, & 1165 66 4 T2 weighted scan 266 - 171 Hankins, 1999) Diffusion Tensor Imaging (DTI) scans 266 - 171 Medical Symptoms Questionnaire 1298 564 257 Magnetization Transfer Imaging (MTR) scan 266 - 171 Cannabis Use Inventory (CUI)(Schubart et al., 2011) 1327 573 260 PRESTO and EPI Resting State fMRI scans 266 - 171 The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol Fagerström Test for Nicotine Dependence(Fagerstroem, 1978) 1327 573 260

42 43 clinical assessment of relatives and controls differed from the patient assessment, Neuropsychological assessment in that the presence of axis-I disorders was checked using the M.I.N.I. (Sheehan A short form of the Wechsler Adult Intelligence Scale-III (WAIS-III) was et al., 1998) Also, the FIGS was completed, which was of particular importance used to estimate current Intelligence Quotient (IQ ). A previous study found for the controls, given the fact that any bipolar or psychotic diagnosis in the first the combination of ‘Information’, ‘Block Design’, ‘Digit Symbol Coding’ and or second degree was an exclusion criteria. ‘Arithmetic’ to account for the greatest amount of variance in full-scale IQ in schizophrenia patients and controls (respectively R2=0.92 and R2=0.91) (Blyler, Biomaterials Gold, Iannone, & Buchanan, 2000). The National Adult Reading test is a short Blood samples were obtained from all participants through venepuncture. tool that was used to estimate premorbid IQ level. It has also proven to be a From patients 73,5 ml was collected and from relatives and controls 46,5 ml. strong predictor of intelligence in the general population (Bright et al., 2002).

15 ml of the total blood sample was used for DNA and RNA extraction. From Chapter 2 April 10 2014 onward, peripheral blood mononuclear cell (PBMC) counts were Actigraphy assessment determined, since this is a potential confounder in RNA expression analysis. Behavioural circadian rhythmicity and sleep-wake measurements were acquired Viable PBMC was separated from whole blood using ficoll gradients and with an Actiwatch (the Actiwatch 2, Philips Respironics). The Actiwatch has stored with DMSO at the UMCU Biobank. For genome-wide genotyping the a solid-state piezo-electric accelerometer and a lithium rechargeable battery. It Illumina 660W quad Infinium array was used with 660,000 haplotype tagging records wrist movements and the sum of wrist movements is scored in epochs of SNPs and for copy-number variation detection. RNA was extracted from whole 1 minute. Participants wore the Actiwatch for a period of 14 consecutive days on blood PaxGene tubes. In addition to genome-wide genotyping, genome-wide their non-dominant wrist while also keeping a diary of their sleeping behaviour. gene RNA sequencing data was obtained in a subset. The healthy controls for In order to obtain 116 circadian rhythm and sleep phenotypes from the the analysis were used from both the current study and other gene expression actigraphic data, several algorithms in R statistical package were used according studies. to the procedure developed by Pagani et. al (2016). The acquired phenotypes can be subdivided in the following categories: rest period, sleep period, non- The fibroblast cell lines were obtained by a skin biopsy at a 3mm punch (Stiefel). parametric analysis (i.e. fragmentation and consolidation of activity) and the Subsequently, the cells were cultured at 37°C in humidified atmosphere fitting of a cosinor curve, a Hill-transformed curve and a Fourier-fitting curve. containing 5% CO2 in Dulbecco’s modified Eagle’s medium, supplemented with 15% fetal bovine serum (FBS) and 1% penicillin/streptomycin/amphotericin B. Neuroimaging The cells were seeded to 6-well plates with each well containing an individual Structural Magnetic Resonance Images were acquired on a 3 Tesla Philips cell line. Achieva scanner (Philips Healthcare, Best, the Netherlands), equipped with an 8-channel SENSE-headcoil. Morphological characteristics (global and focal brain volume, cortical thickness, volume and surface area) were studied, along

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol with white matter abnormalities (coherence, organisation and density of white

44 45 matter tracts), and connectivity networks. The total protocol lasted 43 minutes. comorbid substance use disorder, compared with relatives (1.5%, χ2 [1, 1136] = Processing was done on the neuroimaging computer network of the UMCU - 93.96, p < 0.001) and controls (1.1%, χ2 [1, 816] = 48.82, p < 0.001). Again, no Brain Centre Rudolf Magnus, Utrecht, the Netherlands. differences were found between relatives and controls (χ2 [1, 850] = 0.22, p < 0.64). Results Table 4 - Demographical and clinical characteristics of the DBC sample

Demographic and clinical characteristics Patients Relatives Controls A total of 1695 bipolar patients, 590 relatives and 266 controls participated Age M (sd) * 49.8 (12.3) 56.0 (15.2) 47.5 (15.4) in the DBC study. From 299 patients and 1 relative only blood samples Gender male n (%) * 603 (43.2%) 215 (36.5%) 131 (49.2%) Level of education n (%) * and clinical diagnosis were available from participation in earlier studies, Chapter 2 1. Low education 136 (9.8%) 58 (10.0%) 30 (10.1%) resulting in a total of 1396 patients, 589 relatives and 266 controls of whom 2. Intermediate secondary education 148 (10.7%) 60 (10.6%) 20 (7.8%) the comprehensive assessment was completed. Table 4 shows the demographic 3. Intermediate professional education 260 (18.8%) 99 (17.5%) 37 (14.5%) 4. High prepatory vocational / pre-university 203 (14.6%) 58 (10.2%) 43 (16.8%) and clinical characteristics of the sample. Age differed significantly between 5. Bachelor degree 384 (27.7%) 190 (33.6%) 69 (27.0%) patients, relatives and controls (F [2, 2248] = 55.50, p < 0.001), with significant 6. Master or PhD degree 255 (18.4%) 101 (17.8%) 57 (22.3%) differences between all groups (p < 0.01). Also, the ratio male to female differed Alcohol disorder n (%) 237 (33.6%) 49 (8.3%) 21 (7.9%) between groups (χ2 [1, 2251] = 13.82, p = 0.001). Within the group of relatives Substance use disorder n (%) 103 (18.7%) 9 (1.5%) 3 (1.1%) Anxiety disorder n (%) * 358 (25.6%) 66 (11.2%) 16 (6.0%) 36.5 % was male, which was significantly lower than the percentages of males Age at onset Mean (sd) 30.9 (10.6) - - χ2 in patients (43.2%, [1, 1985] = 7.66, p < 0.01). Groups differed on educational # of depressive episodes Mean (sd) 7.0 (12.0) - - 2 performance (χ [1, 2208] = 19.69, p < 0.05), with a post-hoc significant # of manic episodes Mean(sd) 5.18 (8.4) - - difference between relatives and controls (χ2 [1, 822] = 13.14, p < 0.05). As Lithium use n (%) 853 (65.4%) - - * Significant between-group difference (p<0.05) expected, anxiety disorders were significantly more frequent in patients (25.6%) compared to relatives (11.2%, χ2 [1, 1985] = 51.41, p < 0.001) and controls (6.0%, χ2 [1, 1662] = 49.37, p < 0.001). The difference in anxiety disorders MRI scans between relatives and controls was also significant (χ2 [1, 855] = 5.69, p < 0.05). Structural MRI scans were obtained from 266 patients and 171 controls. Additionally, 33.6 % of patients had a comorbid alcohol use disorder, compared Patients did not differ from controls on age (M age patients = 48.3, controls = to 8.3% of relatives (χ2 [1, 1293] = 118.97, p < 0.001) and 7.9% of controls (χ2 [1, 45.5, t[310.25] = -1.97, p = 0.05), gender (% male in patients = 48.9%, in controls 971] = 65.49 , p < 0.001). There was no difference between relatives and controls = 50.3%, χ2 [1, 437] = 0.08, p = 0.77) and handedness distribution (% right on percentage of alcohol use disorders (χ2 [1, 854)] = 0.05, p = 0.83). A similar handed patients = 86.8%, controls = 85.3%, χ2 [1, 435] = 0.78, p = 0.68). Patients pattern was observed regarding substance use disorder (χ2 [1, 1401] = 132.55, and controls participating in the MRI scans did not differ on age (Patients:F

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol p < 0.001). A significant higher number of patients (18.7%), suffered from a [1, 1660] = 3.32, p = 0.07; Controls: F [1, 435] = 1.82, p = 0.18) and gender

46 47 (Patients: χ2 [1, 1662] = 2.92, p = 0.92; Controls: χ2 [1, 437] = 0.05, p = 0.85), (BP I, BP II, BP NOS and the bipolar subtype of schizoaffective disorder) from the overall sample. (Mühleisen et al., 2014). Additional phenotype information of these cohorts is limited. The risk of focusing solely on a broad and heterogeneous definition of Actigraphy bipolar disorder is that the association with the identified loci becomes hard to Actigraphy was measured in 107 patients, 74 siblings and 80 controls. Siblings interpret. Also, a heterogeneous definition of bipolar status limits the statistical and controls differed on age (M age patients = 50.3, siblings = 54.7, controls = power and underestimates the genotype-phenotype association (Tracy, 2008). 46.8, F[2, 258] = 6.67, p < 0.01), but not on gender (% male in patients = 43.9%, Studying the link between overt phenotypes and genotype data, (e.g. GWAS siblings = 39.2%, controls = 48.8%, χ2 [1, 261] = 1.43, p = 0.49). Depressive results, polygenic risk scores, whole genome sequencing) gives the opportunity symptoms were significantly higher in patients compared to controls (mean to analyse the underlying pathophysiology of bipolar disorder in more detail,

IDS-score in patients = 15.2, siblings = 6.9, controls = 5.8, F [2, 257] 35.72, p < without the noise that results from including the broad spectrum of bipolar Chapter 2 0.001 ). Compared to the overall sample, patients, siblings and controls from the illness. The NIMH initiated RDoC framework is in line with this idea; rather actigraphy study did not differ on age F( [1, 1501] = 0.21, p = 0.64 & F [1, 661] = than using the current diagnostic definitions, RDoC intends to focus on 0.56, p = 0.46 & F [1, 344] = 0.14, p = 0.71) and gender (χ2 [1, 1503] = 0.02, p = symptom level, ultimately leading to a revised classification system based on 0.92 & χ2 [1, 663] = 0.20, p = 0.70 & χ2 [1, 346] = 0.01, p = 1.00). objectively measureable biosignatures. While the DBC study still worked from the classic DSM-IV framework, our deep-phenotype data may be combined Discussion with similar deep-phenotype data from different psychiatric populations. Using our data for cross-disorder studies that measure on symptom level can foster The current paper presents the methods and sample characteristics of the DBC research, such as RDoC, aimed at unraveling the etiology of psychopathology. study, investigating susceptibility to bipolar disorder type I. With 1695 bipolar I patients, 590 relatives and 266 controls, this project has collected one of the The current study also came with some limitations. For one, patients, relatives largest cohorts in its field (Mühleisen et al., 2014; Sklar et al., 2011). Selection and controls differed on age and gender. This is mainly due to the fact that based on Dutch origin has resulted in genetically homogeneous participants and matching was not possible for the group of relatives. We found that the parents with one executive research centre, we have ensured uniform data collection. and female relatives of the included patients were most inclined to participate. The extensive deep phenotype characterization makes this a unique cohort Given that we were aiming for as many relatives as possible, selection based on for many future studies investigating psychological, cognitive and biological age and gender was not possible. In further analysis we should take these group mechanisms that may underlie bipolar disorder. By simultaneously collecting differences into account. Also, with a mean age of little under 50 years old and a a wide variety of phenotypes, while also providing genome-wide genetic data, mean age at onset of 30 years old, our patient sample has been ill for a relatively the current study aimed to bridge the gap between genotype and phenotype long time. Future analysis of biological parameters, could consider including analyses. The latest combined GWAS in bipolar disorder detected 56 SNPs duration of medication use into the analyses, to account for the possible

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol in 5 chromosomal regions but included a broad definition of bipolar disorder confounding medication effects. Second, the assessment protocol for patients

48 49 lasted in total three to four hours and contained delicate questions regarding Reference list their psychiatric background. This was carefully explained to the eligible

participants during the first screening, resulting in a drop-out of several patients Alberti, K. G. M. M., Zimmet, P., & Shaw, J. (2005). The metabolic syndrome - A new who did not feel emotionally up to the challenge. This selection may have led worldwide definition.Lancet, 366(9491), 1059–1062. to a sample of bipolar patients with a higher level of functioning compared Altman, E. G., Hedeker, D., Peterson, J. L., & Davis, J. M. (1997). The altman self-rating mania scale. Biological Psychiatry, 42(10), 948–955. to the average bipolar patient. Since we had no clinical data on patients who dropped out after the initial screening, we could not evaluate this potential Altshuler, L. L., Bartzokis, G., Grieder, T., Curran, J., Jimenez, T., Leight, K., … Mintz, J. (2000). An MRI study of temporal lobe structures in men with bipolar disorder or selection. Third, the majority of participants was recruited via their pharmacies schizophrenia. Biological Psychiatry, 48(2), 147–162. based on a previous lithium prescription. One could argue that this has led to a American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: biased selection of lithium using bipolar patients. However, given that lithium is DSM-5. Washington DC. Chapter 2 recommended by international guidelines as one of the first-line pharmacological Andreasen, N. C., Flaum, M., & Arndt, S. (1992). The Comprehensive Assessment of Symptoms treatments for both acute mania and as maintenance therapy, the majority of the and History (CASH). An instrument for assessing diagnosis and psychopathology. Archives general bipolar population has a history of lithium use (Yatham et al., 2013). of General Psychiatry, 49(8), 615–623.

With 65,4% of patients currently using lithium, these patients do not appear Arts, B., Jabben, N., Krabbendam, L., & van Os, J. (2008). Meta-analyses of cognitive overrepresented in our sample. functioning in euthymic bipolar patients and their first-degree relatives.Psychological Medicine, 38(6), 771–785.

Overall, the DBC study is one of the largest cohorts to investigate the underlying Barnett, J. H., & Smoller, J. W. (2009). The genetics of bipolar disorder.Neuroscience, 164(1), 331–343. mechanisms of bipolar disorder. With a clear need for larger genotyped samples, Bauer, M., Grof, P., Rasgon, N., Bschor, T., Glenn, T., & Whybrow, P. C. (2006). Temporal this study will complement international efforts that are currently being made to relation between sleep and mood in patients with bipolar disorder. Bipolar Disorders, 8(2), delineate genetic susceptibility factors. A more comprehensive understanding of 160–167.

biological, genomic and epidemiological risk factors will not only help us better Bearden, C. E., Hoffman, K. M., & Cannon, T. D. (2001). The neuropsychology and understand bipolar disorder, but may ultimately lead to novel prevention and neuroanatomy of bipolar affective disorder: a critical review. Bipolar Disorders, 3(3), 106–150.

treatment strategies. Belmaker, R. H. (2004). Bipolar disorder. The New England Journal of Medicine, 351, 476–486.

Benedetti, F., Dallaspezia, S., Colombo, C., Pirovano, A., Marino, E., & Smeraldi, E. (2008). A length polymorphism in the circadian clock gene Per3 influences age at onset of bipolar disorder. Neuroscience Letters, 445(2), 184–187.

Benedetti, F., Dallaspezia, S., Fulgosi, M. C., Lorenzi, C., Serretti, A., Barbini, B., … Smeraldi, E. (2007). Actimetric evidence that CLOCK 3111 T/C SNP influences sleep and activity patterns in patients affected by bipolar depression. American Journal of Medical Genetics, Part B: Neuropsychiatric Genetics, 144(5), 631–635. The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol

50 51 Bernstein, D. P., Ahluvalia, T., Pogge, D., & Handelsman, L. (1997). Validity of the Childhood DelBello, M. P., Zimmerman, M. E., Mills, N. P., Getz, G. E., & Strakowski, S. M. (2004). Trauma Questionnaire in an adolescent psychiatric population. Journal of the American Magnetic resonance imaging analysis of amygdala and other subcortical brain regions in Academy of Child and Adolescent Psychiatry, 36(3), 340–348. adolescents with bipolar disorder. Bipolar Disorders, 6(1), 43–52.

Black, D. W., Goldstein, R. B., Mason, E. E., Bell, S. E., & Blum, N. (1992). Depression Delude, C. M. (2015). Deep phenotyping: The details of disease. Nature, 527, S14–S15. and other mental-disorders in the relatives of morbidly obese patients. Journal of Affective Disorders, 25(2), 91–96. DeRijk, R. H., & de Kloet, E. R. (2008). Corticosteroid receptor polymorphisms: Determinants of vulnerability and resilience. European Journal of Pharmacology, 583(2–3), 303–311. Blyler, C. R., Gold, J. M., Iannone, V. N., & Buchanan, R. W. (2000). Short form of the WAIS- III for use with patients with schizophrenia. Schizophrenia Research, 46(2–3), 209–215. Derogatis, L. R., & Cleary, P. A. (1977). Confirmation of the dimensional structure of the scl‐90: A study in construct validation. Journal of Clinical Psychology, 33(4), 981–989. Boomsma, D. I., Wijmenga, C., Slagboom, E. P., Swertz, M. A., Karssen, L. C., Abdellaoui, A., … van Duijn, C. M. (2014). The Genome of the Netherlands: design, and project goals. Dudbridge, F. (2013). Power and predictive accuracy of polygenic risk scores. PLoS Genetics, European Journal of Human Genetics, 22(2), 221–7. 9(3), e1003348

Bora, E., Yucel, M., & Pantelis, C. (2009). Cognitive endophenotypes of bipolar disorder: a Eric, Y. W., Halari, R., Cheng, K. M., Leung, S. K., & Young, A. H. (2013). Cognitive Chapter 2 meta-analysis of neuropsychological deficits in euthymic patients and their first-degree performance is impaired in euthymic Chinese patients with Bipolar 1 Disorder. Journal of relatives. Journal of Affective Disorders, 113(1–2), 1–20. Affective Disorders, 151(1), 156–163.

Brambilla, P., Harenski, K., Nicoletti, M., Sassi, R. B., Mallinger, A. G., Frank, E., … Soares, Fagerstroem, K. O. (1978). Measuring degree of physical dependence to tobacco smoking with J. C. (2003). MRI investigation of temporal lobe structures in bipolar patients. Journal of reference to individualization of treatment. Addictive Behaviors, 3(3–4), 235–241. Psychiatric Research, 37(4), 287–295. Fagiolini, A., Frank, E., Scott, J. A., Turkin, S., & Kupfer, D. J. (2005). Metabolic syndrome Bright, P., Jaldow, E., & Kopelman, M. D. (2002). The National Adult Reading Test as a in bipolar disorder: findings from the Bipolar Disorder Center for Pennsylvanians. Bipolar measure of premorbid intelligence: a comparison with estimates derived from demographic Disorders, 7(5), 424–30. variables. Journal of the International Neuropsychological Society, 8(6), 847–854. First, M. B. et, Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1997). Structured Clinical Caroll, J. F. X., & McGinley, J. J. (2001). A screening form for identifying mental health Interview for DSM-IV Axis I Disorders (SCID-I). New York: Biometric Research problems in alcohol/other drug dependent persons. Alcoholism Treatment Quarterly, 19(4), Department. 33–47. Francioli, L. C., Menelaou, A., Pulit, S. L., van Dijk, F., Palamara, P. F., Elbers, C. C., … Christensen, M. V., Kyvik, K. O., & Kessing, L. V. (2006). Cognitive function in unaffected Wijmenga, C. (2014). Whole-genome sequence variation, population structure and twins discordant for affective disorder.Psychological Medicine, 36(8), 1119–1129. demographic history of the Dutch population. Nature Genetics, 46(8), 818–25.

Craddock, N., & Sklar, P. (2013). Genetics of bipolar disorder. Lancet, 381(9878), 1654–62. Gershon, E. S., Alliey-Rodriguez, N., & Liu, C. (2011). After GWAS: searching for genetic risk for schizophrenia and bipolar disorder. The American Journal of Psychiatry, 168(3), 253– Currie, C., Molcho, M., Boyce, W., Holstein, B., Torsheim, T., & Richter, M. (2008). 256. Researching health inequalities in adolescents: The development of the Health Behaviour in School-Aged Children (HBSC) Family Affluence Scale.Social Science and Medicine, 66(6), Glahn, D. C., Bearden, C. E., Bowden, C. L., & Soares, J. C. (2006). Reduced educational 1429–1436. attainment in bipolar disorder. Journal of Affective Disorders, 92(2–3), 309–312.

Daban, C., Vieta, E., Mackin, P., & Young, A. H. (2005). Hypothalamic-pituitary-adrenal axis Goldstein, B. I., Fagiolini, A., Houck, P., & Kupfer, D. J. (2009). Cardiovascular disease and and bipolar disorder. The Psychiatric Clinics of North America, 28(2), 469–480. hypertension among adults with bipolar I disorder in the United States. Bipolar Disorders, 11(6), 657–62. de Kloet, E. R., Joëls, M., & Holsboer, F. (2005). Stress and the brain: from adaptation to Gonzalez, R. (2014). The relationship between bipolar disorder and biological rhythms. Journal The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol disease. Nature Reviews Neuroscience, 6(6), 463–475. of Clinical Psychiatry, 75(4), 323–331.

52 53 Gottesman, I. I., & Gould, T. D. (2003). Reviews and overviews the endophenotype concept in Johannessen, L., Strudsholm, U., Foldager, L., & Munk-Jørgensen, P. (2006). Increased risk psychiatry: Etymology and strategic intentions. American Journal of Psychiatry 160, 636–645. of hypertension in patients with bipolar disorder and patients with anxiety compared to background population and patients with schizophrenia. Journal of Affective Disorders, 95(1– Gulseren, S., Gurcan, M., Gulseren, L., Gelal, F., & Erol, A. (2006). T2 hyperintensities in 3), 13–7. bipolar patients and their healthy siblings. Archives of Medical Research, 37(1), 79–85. Johnsson, A., Pflug, B., Engelmann, W., & Klemke, W. (1979). Effect of lithium carbonate Gurpegui, M., Martínez-Ortega, J. M., Gutiérrez-Rojas, L., Rivero, J., Rojas, C., & Jurado, D. on circadian periodicity in humans. Pharmakopsychiatrie, Neuro-Psychopharmakologie, 12(6), (2012). Overweight and obesity in patients with bipolar disorder or schizophrenia compared 423–425. with a non-psychiatric sample. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 37(1), 169–75. Kempton, M. J., Geddes, J. R., Ettinger, U., Williams, S. C. R., & Grasby, P. M. (2008). Meta- analysis, database, and meta-regression of 98 structural imaging studies in bipolar disorder. Hafeman, D. M., Chang, K. D., Garrett, A. S., Sanders, E. M., & Phillips, M. L. (2012). Archives of General Psychiatry, 65(9), 1017–1032. Effects of medication on neuroimaging findings in bipolar disorder: An updated review. Bipolar Disorders, 14(4), 375–410. KenBis. Retrieved March 29, 2016, from http://www.kenniscentrumbipolairestoornissen.nl

Hallahan, B., Newell, J., Soares, J. C., Brambilla, P., Strakowski, S. M., Fleck, D. E., … Kerner, B. (2014). Genetics of bipolar disorder. Application of Clinical Genetics, 7, 33–42. Chapter 2 McDonald, C. (2011). Structural magnetic resonance imaging in bipolar disorder: An international collaborative mega-analysis of individual adult patient data. Biological Kirschbaum, C., Tietze, A., Skoluda, N., & Dettenborn, L. (2009). Hair as a retrospective Psychiatry, 69(4), 326–335. calendar of cortisol production-Increased cortisol incorporation into hair in the third trimester of pregnancy. Psychoneuroendocrinology, 34(1), 32–37. Harvey, A. G., Schmidt, D. A., Scarnà, A., Semler, C. N., & Goodwin, G. M. (2005). Sleep- related functioning in euthymic patients with bipolar disorder, patients with insomnia, and Korver, N., Quee, P. J., Boos, H. B. M., Simons, C. J. P., De Haan, L., & GROUP Investigators. subjects without sleep problems. American Journal of Psychiatry, 162, 50–57. (2012). Genetic Risk and Outcome of Psychosis (GROUP), a multi site longitudinal cohort study focused on gene–environment interaction: objectives, sample characteristics, Hasler, G., Drevets, W. C., Gould, T. D., Gottesman, I. I., & Manji, H. K. (2006). Toward recruitment and assessment methods. International Journal of Methods in Psychiatric Research, constructing an endophenotype strategy for Bipolar Disorders. Biological Psychiatry, 60(2), 21(3), 205–221. 93–105. Langan, C., & McDonald, C. (2009). Neurobiological trait abnormalities in bipolar disorder. Hibar, D. P., Westlye, L. T., van Erp, T. G. M., Rasmussen, J., Leonardo, C. D., Faskowitz, Molecular Psychiatry, 14(9), 833–846. J., … Andreassen, O. a. (2016). Subcortical volumetric abnormalities in bipolar disorder. Molecular Psychiatry, 1–7. Leverich, G. S., Nolen, W. A., Rush, A. J., McElroy, S. L., Keck, P. E., Denicoff, K. D., … Post, R. M. (2001). The Stanley Foundation Bipolar Treatment Outcome Network: I. Ho, B.-C., Andreasen, N. C., Ziebell, S., Pierson, R., & Magnotta, V. (2011). Long- Longitudinal methodology. Journal of Affective Disorders, 67, 33–44. term antipsychotic treatment and brain volumes: a longitudinal study of first-episode schizophrenia. Archives of General Psychiatry, 68(2), 128–137. Lopez, A. D., & Murray, C. C. J. . (1998). The global burden of disease , 1990 – 2020. Nature Medicine, 4(11), 1241–1243. Horne, R., Weinman, J., & Hankins, M. (1999). The beliefs about medicines questionnaire: The development and evaluation of a new method for assessing the cognitive representation of Lyoo, I. K., Lee, H. K., Jung, J. H., Noam, G. G., & Renshaw, P. F. (2002). White matter medication. Psychology & Health, 14(1), 1–24. hyperintensities on magnetic resonance imaging of the brain in children with psychiatric disorders. Comprehensive Psychiatry, 43, 361–368. Insel, T., Cuthbert, B., Garvey, M., Heinssen, R., Pine, D. S., Quinn, K., … Wang, P. (2010). Research Domain Criteria (RDoC): Toward a new classification framework for research on Mahon, K., Burdick, K. E., & Szeszko, P. R. (2010). A role for white matter abnormalities in mental disorders. American Journal of Psychiatry, 167(7), 748–751. the pathophysiology of bipolar disorder. Neuroscience and Biobehavioral Reviews, 34(4), 533– 554. Jackson, A., Cavanagh, J., & Scott, J. (2003). A systematic review of manic and depressive The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol prodromes. Journal of Affective Disorders, 74(3), 209–217. Maier, W., Zobel, A., & Wagner, M. (2006). Schizophrenia and bipolar disorder: differences and overlaps. Current Opinion in Psychiatry, 19(2), 165–170.

54 55 Maina, G., Salvi, V., Vitalucci, A., D’Ambrosio, V., & Bogetto, F. (2008). Prevalence and Moncrieff, J., & Leo, J. (2010). A systematic review of the effects of antipsychotic drugs on brain correlates of overweight in drug-naïve patients with bipolar disorder. Journal of Affective volume. Psychological Medicine, 40(9), 1409–1422. Disorders, 110(1–2), 149–55. Monteleone, P., & Maj, M. (2008). The circadian basis of mood disorders: recent developments Manenschijn, L., Koper, J. W., Lamberts, S. W. J., & Van Rossum, E. F. C. (2011). Evaluation and treatment implications. European Neuropsychopharmacology, 18(10), 701–11. of a method to measure long term cortisol levels. Steroids, 76(10–11), 1032–1036. Mühleisen, T. W., Leber, M., Schulze, T. G., Strohmaier, J., Degenhardt, F., Treutlein, J., Mansour, H. A., Talkowski, M. E., Wood, J., Chowdari, K. V., McClain, L., Prasad, K., … Cichon, S. (2014). Genome-wide association study reveals two new risk loci for bipolar Nimgaonkar, V. L. (2009). Association study of 21 circadian genes with bipolar I disorder, disorder. Nature Communications, 5(3339), 1-8. schizoaffective disorder, and schizophrenia.Bipolar Disorders, 11(7), 701–710. Müller-Oerlinghausen, B., Berghöfer, A., & Bauer, M. (2002). Bipolar disorder. The Lancet, Martínez-Arán, A., Vieta, E., Colom, F., Reinares, M., Benabarre, A., Gastó, C., & Salamero, 359(9302), 241–247. M. (2000). Cognitive dysfunctions in bipolar disorder: Evidence of neuropsychological disturbances. Psychotherapy and Psychosomatics, 69(1), 2–18. Navari, S., & Dazzan, P. (2009). Do antipsychotic drugs affect brain structure? A systematic and critical review of MRI findings.Psychological Medicine, 39(11), 1763–1777. Maxwell, M. E. (1992). Family Interview for Genetic Studies (FIGS): a manual for FIGS. Chapter 2 Bethesda, MD: National Institute of Mental Health: Clinical Neurogenetics Branch. Nievergelt, C. M., Kripke, D. F., Barrett, T. B., Burg, E., Remick, R. A., Sadovnick, A. D., … Kelsoe, J. R. (2006). Suggestive evidence for association of the circadian genes PERIOD3 McCarthy, M. J., Nievergelt, C. M., Kelsoe, J. R., & Welsh, D. K. (2012). A survey of genomic and ARNTL with bipolar disorder. American Journal of Medical Genetics - Neuropsychiatric studies supports association of circadian clock genes with bipolar disorder spectrum illnesses Genetics, 141 B(3), 234–241. and lithium response. PLoS ONE, 7(2), e32091. Nurnberger, J. I., Adkins, S., Lahiri, D. K., Mayeda, A., Hu, K., Lewy, A., … Davis-Singh, D. McCrae, R. R., & Costa, P. T. (2004). A contemplated revision of the NEO Five-Factor (2000). Melatonin suppression by light in euthymic bipolar and unipolar patients. Archives of Inventory. Personality and Individual Differences, 36(3), 587–596. General Psychiatry, 57, 572–579.

McElroy, S. L., & Keck, P. E. (2012). Obesity in bipolar disorder: an overview. Current Pagani, L., Semenova, E. A., Moriggi, E., Revell, V. L., Hack, L. M., Lockley, S. W., … Psychiatry Reports, 14(6), 650–8. Brown, S. A. (2010). The physiological period length of the human circadian clock In Vivo is directly proportional to period in human fibroblasts.PLoS ONE, 5(10), e13376. McGuffin, P., Rijsdijk, F., Andrew, M., Sham, P., Katz, R., & Cardno, A. (2003). The heritability of bipolar affective disorder and the genetic relationship to unipolar depression. Pagani, L., St. Clair, P. a., Teshiba, T. M., Service, S. K., Fears, S. C., Araya, C., … Freimer, N. Archives of General Psychiatry, 60(5), 497–502. B. (2016). Genetic contributions to circadian activity rhythm and sleep pattern phenotypes in pedigrees segregating for severe bipolar disorder. Proceedings of the National Academy of McIntosh, A. M., Harrison, L. K., Forrester, K., Lawrie, S. M., & Johnstone, E. C. (2005). Sciences, 113(6), E754–E761. Neuropsychological impairments in people with schizophrenia or bipolar disorder and their unaffected relatives.British Journal of Psychiatry, 186, 378–385. Penninx, B. W. J. H., Beekman, A. T. F., Smit, J. H., Zitman, F. G., Nolen, W. A., Spinhoven, P., … van Dyck, R. (2008). The Netherlands Study of Depression and Anxiety (NESDA): McIntyre, R. S., Danilewitz, M., Liauw, S. S., Kemp, D. E., Nguyen, H. T. T., Kahn, L. S., … Rationale, objectives and methods. International Journal of Methods in Psychiatric Research, Taylor, V. H. (2010). Bipolar disorder and metabolic syndrome: An international perspective. 17(3), 121–140. Journal of Affective Disorders, 126(3), 366–387. Peters, E. R., Joseph, S. A., & Garety, P. A. (1999). Measurement of delusional ideation in the Mesman, E., Nolen, W. A., Reichart, C. G., Wals, M., & Hillegers, M. H. J. (2013). The dutch normal population: introducing the PDI (Peters et al. Delusions Inventory). Schizophrenia bipolar offspring study: 12-year follow-up.American Journal of Psychiatry, 170(5), 542–549. Bulletin, 25(3), 553–576.

Milhiet, V., Etain, B., Boudebesse, C., & Bellivier, F. (2011). Circadian biomarkers, circadian Pillai, J. J., Friedman, L., Stuve, T. A., Trinidad, S., Jesberger, J. A., Lewin, J. S., … Schulz, S. genes and bipolar disorders. Journal of Physiology Paris, 105, 183–189. C. (2002). Increased presence of white matter hyperintensities in adolescent patients with The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol bipolar disorder. Psychiatry Research, 114(1), 51–56.

56 57 Purcell, S. M., Wray, N. R., Stone, J. L., Visscher, P. M., O’Donovan, M. C., Sullivan, P. F., Shi, J., Wittke-Thompson, J. K., Badner, J. A., Hattori, E., Potash, J. B., Willour, V. L., … & Sklar, P. (2009). Common polygenic variation contributes to risk of schizophrenia and Liu, C. (2008). Clock genes may influence bipolar disorder susceptibility and dysfunctional bipolar disorder. Nature, 460(7256), 748–752. circadian rhythm. American Journal of Medical Genetics, Part B: Neuropsychiatric Genetics, 147(7), 1047–1055. Quraishi, S., & Frangou, S. (2002). Neuropsychology of bipolar disorder: A review. Journal of Affective Disorders, 72(3), 209–226. Sicras, A., Rejas, J., Navarro, R., Serrat, J., & Blanca, M. (2008). Metabolic syndrome in bipolar disorder: a cross-sectional assessment of a Health Management Organization database. Raine, A. (1991). The SPQ: a scale for the assessment of schizotypal personality based on DSM- Bipolar Disorders, 10(5), 607–16. III-R criteria. Schizophrenia Bulletin, 17(4), 555–564. Sklar, P., Ripke, S., Scott, L. J., Andreassen, O. A., Cichon, S., Craddock, N., … Purcell, S. Robins, L. N., Wing, J., Wittchen, H. U., Helzer, J. E., Babor, T. F., Burke, J., … Regier, D. A. M. (2011). Large-scale genome-wide association analysis of bipolar disorder identifies a new (1988). The Composite International Diagnostic Interview. Archives of General Psychiatry, 45, susceptibility locus near ODZ4. Nature Genetics, 43(10), 977–983. 1069–77. Smoller, J. W., Craddock, N., Kendler, K., Lee, P. H., Neale, B. M., Nurnberger, J. I., … Robinson, L. J., Thompson, J. M., Gallagher, P., Goswami, U., Young, A. H., Ferrier, I. N., & Sullivan, P. F. (2013). Identification of risk loci with shared effects on five major psychiatric Chapter 2 Moore, P. B. (2006). A meta-analysis of cognitive deficits in euthymic patients with bipolar disorders: a genome-wide analysis. Lancet, 381(9875), 1371–1379. disorder. Journal of Affective Disorders, 93(1–3), 105–15. Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the state-trait anxiety Robinson, P. N. (2012). Deep phenotyping for precision medicine. Human Mutation, 33(5), inventory, 1–23. 777–780. Spijker, A. T., Giltay, E. J., van Rossum, E. F. C., Manenschijn, L., DeRijk, R. H., Haffmans, Rush, A. J., Carmody, T., & Reimitz, P. (2000). The Inventory of Depressive Symptomatology J., … Hoencamp, E. (2011). Glucocorticoid and mineralocorticoid receptor polymorphisms (IDS): Clinician (IDS-C) and Self-Report (IDS-SR) ratings of depressive symptoms. and clinical characteristics in bipolar disorder patients. Psychoneuroendocrinology, 36(10), International Journal of Methods in Psychiatric Research, 9, 45–59. 1460–1469.

Sajatovic, M. (2005). Bipolar disorder: disease burden. The American Journal of Managed Care, Spijker, A. T., Van Rossum, E. F. C., Hoencamp, E., Derijk, R. H., Haffmans, J., Blom, M., 11(3 Suppl), S80–S84. … Zitman, F. G. (2009). Functional polymorphism of the glucocorticoid receptor gene Sauvé, B., Koren, G., Walsh, G., Tokmakejian, S., & Van Uum, S. H. M. (2007). Measurement associates with mania and hypomania in bipolar disorder. Bipolar Disorders, 11(1), 95–101. of cortisol in human hair as a biomarker of systemic exposure. Clinical and Investigative Stefanis, N. C., Hanssen, M., Smirnis, N. K., Avramopoulos, D. a, Evdokimidis, I. K., Stefanis, Medicine, 30(5), E183–E191. C. N., … Van Os, J. (2002). Evidence that three dimensions of psychosis have a distribution Scheepers, F. E., De Wied, C. C. G., Pol, H. E. H., Van De Flier, W., Van Der Linden, J. A., in the general population. Psychological Medicine, 32(2), 347–358. & Kahn, R. S. (2001). The effect of clozapine on caudate nucleus volume in schizophrenic Stefansson, H., Ophoff, R. A., Steinberg, S., Andreassen, O. A., Cichon, S., Rujescu, D., … patients previously treated with typical antipsychotics. Neuropsychopharmacology, 24(1), 47– Collier, D. A. (2009). Common variants conferring risk of schizophrenia. Nature, 460(7256), 54. 744–747.

Schubart, C. D., Sommer, I. E. C., van Gastel, W. A., Goetgebuer, R. L., Kahn, R. S., & Boks, Strakowski, S. M., DelBello, M. P., Sax, K. W., Zimmerman, M. E., Shear, P. K., Hawkins, J. M. P. M. (2011). Cannabis with high cannabidiol content is associated with fewer psychotic M., & Larson, E. R. (1999). Brain magnetic resonance imaging of structural abnormalities experiences. Schizophrenia Research, 130(1–3), 216–221. in bipolar disorder. Archives of General Psychiatry, 56(3), 254–260.

Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., … Suppes, T., Leverich, G. S., Keck, P. E., Nolen, W. A., Denicoff, K. D., Altshuler, L. L., Dunbar, G. C. (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.): The … Post, R. M. (2001). The Stanley Foundation Bipolar Treatment Outcome Network development and validation of a structured diagnostic psychiatric interview for DSM-IV - II. Demographics and illness characteristics of the first 261 patients. Journal of Affective and ICD-10. Journal of Clinical Psychiatry, 59, 22–33. Disorders, 67, 45–59. The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol

58 59 Takahashi, J. S., Takahashi, J. S., Hong, H.-K., Hong, H.-K., Ko, C. H., Ko, C. H., … Wardenaar, K. J., & de Jonge, P. (2013). Diagnostic heterogeneity in psychiatry: towards an McDearmon, E. L. (2008). The genetics of mammalian circadian order and disorder: empirical solution. BMC Medicine, 11(1), 201. implications for physiology and disease. Nature Reviews. Genetics, 9(10), 764–75. Watson, S., Gallagher, P., Ritchie, J. C., Ferrier, I. N., & Young, A. H. (2004). Hypothalamic- Taylor, L., Faraone, S. V, & Tsuang, M. T. (2002). Family, twin, and adoption studies of bipolar pituitary-adrenal axis function in patients with bipolar disorder. The British Journal of disease. Current Psychiatric Reports, 4, 130–133. Psychiatry, 184, 496–502.

Thomas, P. (2004). The many forms of bipolar disorder: A modern look at an old illness. Journal Wechsler, D. (1997). WAIS­III: Wechsler Adult Intelligence Scale (3rd ed.) Administration and scoring of Affective Disorders, 79 SUPPL., 3–8. manual. The Psychological Corporation, San Antonio, TX.

Tighe, S. K., Reading, S. A., Rivkin, P., Caffo, B., Schweizer, B., Pearlson, G., … Bassett, S. Wolfe, J., Kimerling, R., Brown, P., Chrestman, K., & Levin, K. (1996). Psychometric review S. (2012). Total white matter hyperintensity volume in bipolar disorder patients and their of the life stressor checklist-revised. In: Instrumentation in stress, trauma, and adaptation (pp. healthy relatives. Bipolar Disorders, 14(8), 888–893. 141–151). Lutherville, MD.

Toulopoulou, T., Quraishi, S., McDonald, C., & Murray, R. M. (2006). The Maudsley Family Wray, N., Goddard, M., & Visscher, P. (2007). Prediction of individual genetic risk to disease Study: premorbid and current general intellectual function levels in familial bipolar I from genome-wide association studies. Genome Research, 17, 1520–1528. Chapter 2 disorder and schizophrenia. Journal of Clinical and Experimental Neuropsychology, 28(2), 243– 259. Yang, S., Van Dongen, H. P. A., Wang, K., Berrettini, W., & Bućan, M. (2009). Assessment of circadian function in fibroblasts of patients with bipolar disorder. Molecular Psychiatry, 14(2), Tracy, R. P. (2008). “Deep phenotyping”: characterizing populations in the era of genomics and 143–155. systems biology. Current Opinion in Lipidology, 19(2), 151–7. Yatham, L. N., Kennedy, S. H., Parikh, S. V, Schaffer, A., Beaulieu, S., Alda, M., …Berk, UPPER. Retrieved March 29, 2016, from http://www.uu.nl/organisatie/departement- M. (2013). Canadian Network for Mood and Anxiety Treatments (CANMAT) and farmaceutische-wetenschappen/upper International Society for Bipolar Disorders (ISBD) collaborative update of CANMAT guidelines for the management of patients with bipolar disorder: update 2013. Bipolar van Oosterhout, W., Weller, C., Stam, A., Bakels, F., Stijnen, T., Ferrari, M., & Terwindt, G. Disorders, 15(1), 1–44. (2011). Validation of the web-based LUMINA questionnaire for recruiting large cohorts of migraineurs. Cephalalgia, 31(13), 1359–1367.

Van Uum, S. H. M., Sauvé, B., Fraser, L. A., Morley-Forster, P., Paul, T. L., & Koren, G. (2008). Elevated content of cortisol in hair of patients with severe chronic pain: a novel biomarker for stress. Stress, 11(6), 483–488.

Vancampfort, D., Vansteelandt, K., Correll, C. U., Mitchell, A. J., De Herdt, A., Sienaert, P., … De Hert, M. (2013). Metabolic syndrome and metabolic abnormalities in bipolar disorder: a meta-analysis of prevalence rates and moderators. The American Journal of Psychiatry, 170(3), 265–74.

Vingerhoets, A. J. J. M., & Tilburg, M. A. L. van. (1994). Alledaagse problemen lijst. Lisse: Swets en Zeitlinger.

VMDB. Retrieved March 29, 2016, from http://www.vmdb.nl/

Vreeker, A., Boks, M., Abramovic, L., Verkooijen, S., van Bergen, A., Hillegers, M., … Investigators, G. (2016). High educational performance is a distinctive feature of bipolar

The DutchBipolar Cohort, a large studybipolarof disorder I and relatedphenotypes participants: in 2551 study protocol disorder; a study on cognition in bipolar disorder or schizophrenia patients, relatives and controls. Psychological Medicine, 46, 807–818.

60 61 Chapter 3

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls

Sanne Verkooijen, Annet H. van Bergen, Stefan E. Knapen, Annabel Vreeker, Lucija Abramovic, Lucia Pagani, Yoon Jung, Rixt Riemersma- van der Lek, Robert A. Schoevers, Joseph S. Takahashi, René S. Kahn, Marco P. M. Boks, Roel A. Ophoff

Journal of Affective Disorders (2017), 208, 248-254 Abstract disturbances are a reflection of current mood state, but are unrelated to the course of the disorder. Objectives Disturbances in sleep and waking patterns are highly prevalent during mood episodes in bipolar disorder. The question remains whether these disturbances persist during phases of euthymia and what the role is of genetic susceptibility without the presence of bipolar disorder. The current study investigates objective sleep measures in a large sample of bipolar I patients, non- affected siblings and controls.

Methods A total of 107 bipolar disorder I patients, 74 non-affected siblings, and 80 controls were included. Sleep was measured with actigraphy over the course of 14 days. Seven sleep parameters were analysed for group

differences and their relationship with age at onset, number of episodes and Chapter 3 psychotic symptoms using linear mixed model analysis to account for family dependencies.

Results Patients had a longer sleep duration and later time of sleep offset compared to the non-affected siblings but these differences were entirely attributable to differences in mood symptoms. We found no difference between patients and controls or siblings and controls when the analyses were restricted to euthymic patients. None of the bipolar illness characteristics were associated with sleep.

Limitations Medication use was not taken into account which may have influenced our findings and controls were younger compared to non-affected siblings.

Conclusions In the largest study to date, our findings suggest that recovered bipolar I patients do not experience persisting sleep disturbances and that sleep An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls

64 65 Introduction allow for measurements in a natural environment over several weeks. Validation showed better performance compared to observational measurements, sleep Bipolar disorder is a chronic psychiatric disorder characterized by severe logs and diaries (Ancoli-Israel et al., 2003). When assessed in a bipolar fluctuations in mood, which affects 1-2% of the general population (Belmaker, sample, actigraphy showed high correlations with polysomnography and high 2004). According to the DSM-V criteria, one prominent manifestation of to moderate correlations with subjective measures of sleep (Boudebesse et al., a mood episode is a shift in sleep-wake behaviour (American Psychiatric 2014). Nevertheless, studies using actigraphy in the analysis of sleep in bipolar Association, 2013). The majority of patients experience insomnia or hypersomnia disorder have come up with conflicting results. While some studies found during depression and a reduced need for sleep during mania (Harvey, 2008). differences between bipolar patients and controls on several sleep parameters These disturbances are thought to be a hallmark of a current mood episode, (e.g. sleep duration, sleep onset latency, sleep efficiency) (Pierre Alexis Geoffroy and often precede mood episodes, suggesting utility as a marker of prodromal et al., 2014; Harvey et al., 2005; Millar, Espie, & Scott, 2004; Salvatore et al., symptoms (Jackson, Cavanagh, & Scott, 2003). 2008), other studies failed to replicate these findings (Jones, Hare, & Evershed, 2005; Kaplan, Talbot, Gruber, & Harvey, 2012; St-Amand, Provencher,

Studies focusing on bipolar patients have tried to delineate whether these sleep Bélanger, & Morin, 2013). These contradictory results can potentially be Chapter 3 disturbances can be considered as more than a state marker of the disorder and explained by differences in methodology. Measurement periods varied represent a general deregulation of the endogenous circadian cycle independent between 2 and 54 nights and different diagnostic criteria may have resulted in from current episodes. In accordance with that idea, phase advances, phase heterogeneous patient samples. Moreover, sample sizes were generally small (<36 delays and general phase instabilities have been reported for several measures cases and controls), raising the possibility that some studies were underpowered. of rhythmicity, such as body temperature, nocturnal cortisol levels and peak Recently, two meta-analyses concluded that bipolar patients differed from melatonin time in euthymic bipolar patients (Milhiet, Etain, Boudebesse, & controls on measures of sleep duration, sleep onset latency and wake after sleep Bellivier, 2011; Nurnberger et al., 2000). Sleep (the most evident behavioural onset (Geoffroy et al., 2015b; Ng et al., 2014). Sleep efficiency of bipolar patients reflection of circadian rhythms) has also been reported to be disturbed during was significantly lower in only one of the two meta-analyses . According to the non-clinical phases of the disorder. According to Harvey et al. (2005), 70% authors, the number of actigraphy studies in bipolar disorder is limited and lags of euthymic patients reported clinically significant sleep problems, with 55% behind similar research in depression and ADHD. Furthermore, Geoffroy et al. of patients also meeting the criteria for insomnia. Several other studies using pointed to heterogeneity in methodologies and found that age matching, level of self-report measures identified worse sleep quality and more disturbed sleep- depressive symptoms and actigraphy device potentially influence the actigraphy timing preferences in non-clinical bipolar patients compared to controls (Cretu, analyses and should be taken into account in future research (Geoffroy et al., Culver, Goffin, Shah, & Ketter, 2016; Rocha, Neves, & Corrêa, 2013; Seleem 2015a, 2015b). Moreover, associations between objective sleep parameters and et al., 2015). Actigraphy has proven to be an indispensable tool in the objective bipolar illness characteristics have so far not been studied. If sleep disturbances assessment of sleep-wake parameters (Sadeh, Hauri, Kripke, & Lavie, 1995). indeed reflect a continuous aberration of the circadian rhythm, it is conceivable

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls Actiwatches are wrist-worn devices that continuously record movement and that these disturbances correlate with an unfavourable course of the disorder.

66 67 Although the presence of sleep disturbances in bipolar patients has gained Methods increasing attention, the question whether disturbances in sleep patterns can be considered as a heritable trait of bipolar disorder, has scarcely been investigated. Sample As of yet, only two studies objectively studied the difference in sleep-wake The current study is a follow up of the Dutch Bipolar Cohort (DBC) study, which behaviour between bipolar patients and their relatives. Jones et al. (2006) is a collaboration between the University Medical Centre Utrecht (UMCU), studied children of bipolar patients and concluded that sleep onset latency various health care institutes in the Netherlands and the University of California and sleep fragmentation were lower in children of bipolar parents compared to Los Angeles (UCLA). In short, the DBC study is designed to provide a deep- control children. However, when affected bipolar offspring was excluded the phenotype characterization of bipolar I patients and their first-degree relatives. effects were no longer significant. Pagani et al. (2015) analysed 26 pedigrees The cohort included 1700 bipolar I patients, 586 relatives and 265 controls. After ascertained for bipolar I disorder and showed that bipolar patients slept longer completion of the DBC protocol, a subgroup of patients, siblings and controls and woke up later compared to their non-affected relatives. The authors also were re-approached to participate in the actigraphy protocol. Both the DBC provided evidence that a number of sleep measures are heritable. Extending study and the current study were approved by the medical ethical committee

this pedigree study by also including independent control subjects gives the of the UMCU and were in accordance with the Helsinki Declaration. Written Chapter 3 opportunity to study the sleep-wake pattern in individuals who are genetically informed consent was obtained from all participants prior to participation. susceptible for the disorder, but lack the direct illness and its sequelae such as medication use. If first-degree relatives indeed show disturbed sleep patterns All participants had a minimum age of 18, at least three grandparents of similar to probands, it would support the hypothesis that the sleep-wake pattern Dutch descent. Exclusion criteria for all participants were self-reported is a trait of bipolar disorder. major somatic illness (e.g. sleep apnoea) and pregnancy. Inclusion criteria for patients was a bipolar I diagnosis, verified using the Structural Clinical The current study aims at extending previous research in a large, homogenous Interview for DSM-IV (SCID-I) (First, Spitzer, Gibbon, & Williams, 1997) sample of bipolar I patients using an elaborate collection of objective measures and no current admission for their bipolar illness. None of the patients reported of sleeping behaviour. First, the question will be addressed whether euthymic being in a current mood episode. However, 17 patients scored above the IDS- bipolar patients show differences in sleep pattern compared to controls and SR threshold for mild depressive symptoms. Additionally, 4 patients had an whether the non-affected siblings display similar patterns of sleeping behaviour. ASRM score indicative of a manic or hypomanic state. Siblings and control Sleep duration, timing of sleep onset, timing of sleep offset, sleep onset subjects with a diagnosis of bipolar disorder or other psychotic disorders were latency, sleep efficiency, wake after sleep onset (WASO) and sleep inertia will excluded, as were control subjects who had a first or second degree relative be measured objectively using actigraphy. Subsequently, the association with with such diagnoses. Both siblings and controls were assessed using the Mini- current mood symptom level and life-time illness characteristics (i.e. age at International Neuropsychiatric Interview (M.I.N.I) (Sheehan et al., 1998). In onset, number of mood episodes, presence of psychotic symptoms and history of total, 466 eligible candidates were approached for participation via telephone,

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls suicidal behaviour) will be analysed. post or e-mail. 106 subjects did not respond to our invitation, 57 subjects

68 69 refused to participate and 17 subjects did not show up for their appointment. social time, the Saturday prior to DST, the Sunday of DST and the following In 5 participants with an appointment the Actiwatch never started recording Monday were discarded. This way, any jet lag caused by the shift in time was and in 5 participants the data showed consecutive nights in which the minimum excluded from analyses. activity never reached 0, indicative of measurement errors. These 10 participants were excluded from further analyses. After excluding participants who did not In the next algorithm, the rest period was set for every 24-hr period. The meet the inclusion criteria (sleep apnoea n=3, other somatic illness n=1, non- starting point of rest was defined as the first moment after 6 PM when the compliance to the protocol due to a mood episode n=3, sibling not meeting participant started resting. A threshold was set (the median of overall activity, diagnostic criteria n=7, controls not meeting diagnostic criteria n=1) the sample not exceeding 50 counts per epoch), which was used to identify consecutive consisted of 107 patients, 74 siblings and 80 controls. epochs with low activity. The beginning of the rest interval was defined as the epoch that was preceded by consistent activity above the threshold, followed by Actigraphy recordings and sleep logs a minimum of 15 epochs of activity below the threshold, allowing 2 minutes Sleep-wake measurements were recorded with the Actiwatch 2 (Philips to be above threshold. The end of the rest interval was similarly defined as the

Respironics Inc, Murrysville, PA, USA). The Actiwatch has a solid state piezo- epoch that was followed by consistent activity above the threshold and preceded Chapter 3 electric accelerometer and a lithium rechargeable battery. It records wrist by a minimum of 15 epochs of activity below the threshold, also allowing 2 movements and the sum of wrist movements is scored in epochs of 1 minute. All minutes to be above threshold. The reported rest intervals from the sleep logs participants were instructed to wear the Actiwatch for a period of 14 consecutive were used to cross-check the identified intervals set by R. To prevent daytime days on their non-dominant wrist and only to remove it when exposed to water naps to get incorporated in the rest intervals, the script excluded intervals that for long periods of time (e.g. swimming). During the 14-day recording period, had a rest period beginning 120 minutes before 6 PM. Also, when participants participants kept a sleep diary in which bed times, nap times and off-wrist reported an off-wrist interval in their sleep logs, the recording was checked for periods were noted. All Actiwatches were subjected to two calibration protocols 5 consecutive epochs or more with an activity count of 0 in the reported interval prior to initial data collection and after every battery service (for details, see minus and plus 60 minutes. 24 hour-periods with more than 60 minutes defined Pagani et al., 2015). as off-wrist were discarded.

Actigraphy data analysis After the rest period was defined, the sleep algorithm was used to calculate sleep To calculate the 7 sleep phenotypes, a series of algorithms in R statistical duration, timing of sleep onset, timing of sleep offset, sleep onset latency, sleep package (R Development Core Team, 2014) were used according to the efficiency (i.e. minutes asleep divided by minutes in bed), wake after sleep onset procedure developed by Pagani et al. (2015). First, alterations had to be made (WASO) and sleep inertia (wake interval between sleep offset and time out of for the recording days in which daylight saving time (DST) either set the clock bed). This algorithm was based on the Respironics Software Actiware. back with one hour (fall) or set the clock forward with one hour (spring). After

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls adjusting the time of the subsequent recordings so that it would match the new

70 71 Disease characteristics (means and standard deviations of sleep duration, sleep onset, sleep onset Patients were assessed using several (semi)structured clinical interviews. latency, WASO, sleep efficiency, sleep offset, sleep inertia) two separate models The Questionnaire for Bipolar Disorder (Leverich et al., 2001; Suppes et al., were used. One for the comparison of patients and siblings versus controls and 2001) was used to inquire age at onset of the disorder (defined as age of first one for patients and controls versus siblings. A random factor for relatedness was medication). The presence of psychotic symptoms during mood episodes, used, along with age and gender as covariates. A second series of analyses tested the number of episodes and presence of rapid cycling in previous year were the possible confounding effect of depression and work by adding IDS-SR score evaluated with the SCID-I (First et al., 1997). The Comprehensive Assessment and the number of workdays as covariate to the previous described mixed-effects of Symptoms and History (CASH) was used to screen for history of attempted models. The association between the illness characteristics and actigraphy sleep suicide (Andreasen, Flaum, & Arndt, 1992) pattern was analysed using linear regression.

At the start of the actigraphy measurement period and after 7 days, depressive Means and standard deviations from a recent meta-analysis by Geoffroy et symptoms were assessed using the Inventory for Depressive Symptoms - Self al. (2015) were combined with the means and standard deviations from the

Report (IDS-SR) (Rush, Carmody, & Reimitz, 2000) and manic symptoms current study in order to assess the effect sizes of sleep duration, sleep onset Chapter 3 were assessed using the Altman Self-Rating Mania Scale (ASRM) (Altman, latency, WASO and sleep efficiency. Standardized mean differences (SMD) Hedeker, Peterson, & Davis, 1997). For symptom measurement, the average of were calculated with Comprehensive Meta-Analysis 3.3.070 (Pierce, 2008), both time points was used in the analyses. using random effects models. Only subjects from the current study with IDS-SR scores < 26 and ASRM scores < 6 were included in the meta-analysis. Statistical analysis All analyses were carried out using either the R statistical package or IBM SPSS Results Statistics 21.0. Assumptions were checked and in case of significant outliers, sensitivity analyses were performed to assess the impact of the outliers on the Participants overall model. Demographic data, group differences on continuous variables A total of 107 patients, 74 siblings and 80 controls were included with a were analysed using analysis of variance (ANOVA) with Bonferroni post hoc mean measurement period of 14.4 days. Patients, siblings and controls had an tests. Group differences on categorical variables were analysed using chi-square equal male-female ratio, but groups differed on mean age, with siblings being analyses. Possible measurement bias of the sleep measurements by distribution significantly older than controls F( [2, 258] = 6.66, p = 0.001). All groups of Actiwatch devices was assessed using multivariate analysis of variance reported similarly low levels of manic symptoms. Current depressive symptoms (MANOVA). were significantly higher in patients F( [2, 257] = 35.72, p < 0.001). The total number of workdays during the measurement period was lowest in patients, To analyse differences in actigraphy sleep pattern between patients, siblings and while siblings and controls worked an equal number of days (F [2, 258] = 4.05,

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls controls, mixed-effects model analyses were conducted. Per outcome measure p = 0.02). See table 1 for additional sample characteristics. Compared to the

72 73 overall DBC cohort, the sample of patients, siblings and controls participating Group differences in the actigraphy study did not differ on age F( [1, 1501] = 0.21, p = 0.64 & F Table 2 shows the sleep parameters of the three groups and table 3 the summary [1, 661] = 0.56, p = 0.46 & F [1, 344] = 0.14, p = 0.71) and gender (χ2 [1, 1503] = of mixed-effect models. We found no differences between patients and controls 0.02, p = 0.92 & χ2 [1, 663] = 0.20, p = 0.70 & χ2 [1, 346] = 0.01, p = 1.00). on any of the sleep parameters. Compared to the non-affected siblings, patients slept significantly longer and had a later timing of sleep offset. Adjusting for At time of inclusion, 3 patients did not use any form of psychotropic medication, depression symptom score by adding IDS-SR score as covariate, rendered the whereas the remaining patients used between 2 to 5 types of medication (median differences between patients and siblings non-significant at a Bonferroni-corrected = 2). 58 patients received lithium, 36 patients received other mood stabilizers, 22 significance level of 0.007. Adjusting for IDS-SR score in the remaining non- patients received antidepressants, 35 patients received antipsychotics (atypical = significant models (i.e. sleep onset, sleep onset latency, sleep efficiency, WASO and 29, typical = 6) and 29 patients received benzodiazepines. With the exception of sleep inertia) did not change the results. Adding number of workdays as additional 2 patients, all subjects used the same type of medication for at least 3 months. covariate in all previous models did not change any of the results. We subsequently None of the subjects reported doing shift work during the measurement period. analysed whether the groups differed on variability in sleep measures, but found

no significant differences between patients and controls, patients and siblings or Chapter 3 siblings and controls (see supplementary table 1). Since Actiwatch device was not Table 1 – Sample characteristics associated with the sleep measurements (V=0.73, F[192,1368]=0.99, p=0.54), we Patients Siblings Controls N = 107 N = 74 N = 80 decided not to add it to the analyses as covariate. Supplementary figure 1 shows Gender male n (%) 47 (43.9%) 29 (39.2%) 39 (48.8%) examples of actograms of a bipolar patient (1A), sibling (1B) and control (1C). Age M (sd) a 50.3 (11.6) 54.7 (12.1) 46.8 (16.3) Altman M (sd) 1.9 (1.9) 1.2 (1.4) 1.6 (2.2) Relation current depressive symptoms with sleep IDS M (sd) b 15.2 (11.1) 6.9 (6.5) 5.8 (4.9) The direct association between sleep duration and IDS score was tested in linear Waist circumference M (sd) 95.8 (13.9) 91.3 (12.3) 91.4 (12.8) regression analyses with age and gender as covariates. For patients and siblings, Days Work M (sd) c 4.1 (3.7) 5.0 (3.6) 5.6 (3.9) a significant association was found (see figure 1). To prevent circular reasoning, Age at onset 33.8 (13.3) - - Number of episodes 15.5 (24.5) - - the analyses were repeated after questions regarding sleep were eliminated from History of psychotic symptoms N (%) 73 (69.5%) - - the IDS-SR score. The association between depression score and sleep duration a Difference siblings and controls significant (p < 0.01) remained significant in both patients and sibling β( =0.28, t=3.01, p=0.003 and b Difference patients and siblings and patients and controls significant (p < 0.001) β=0.33, t=2.90, p=0.005). c Difference patients and controls significant (p < 0.05)

Linear regression analyses testing the direct relation between sleep offset and IDS score showed a significant association in patients and controls (see figure 2). An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls

74 75 Table 2 – Means (standard deviations) for sleep measures

Patients Siblings Controls Sleep duration 475.16 (82.68) 444.00 (55.47) 452.10 (44.53) Sleep onset -3.54 (76.22) -16.86 (54.45) 9.59 (74.13) Sleep onset latency 5.32 (5.50) 5.33 (7.91) 6.01 (13.00) Sleep efficiency 86.20 (6.32) 85.95 (6.42) 87.17 (4.41) WASO 54.72 (23.15) 50.95 (16.99) 49.60 (18.15) Sleep offset 471.63 (79.52) 427.14 (44.90) 461.69 (77.64) Sleep inertia 5.89 (5.78) 6.87 (7.63) 4.92 (5.36)

Table 3 – Analysis of differences in sleep parameters between groups in crude model vs model adjusted for current level of depression

Patients vs Siblings Patients vs Controls Siblings vs Controls Chapter 3 β p β p β p Figure 1 - Association between mean sleep duration and IDS score A Sleep duration 31.75 0.003* 22.20 0.03 9.55 0.37 Sleep onset 7.34 0.49 -8.02 0.44 15.36 0.17 Sleep onset latency 0.01 1.00 -0.70 0.61 0.70 0.64 Sleep efficiency 0.48 0.59 -1.18 0.18 1.66 0.08 WASO 0.56 0.83 5.03 0.11 -4.47 0.18 Sleep offset 40.30 < 0.001* 15.15 0.16 25.15 0.03 Sleep inertia -1.27 0.19 1.22 0.20 -2.49 0.02

B Sleep duration 10.90 0.28 0.63 0.95 10.27 0.31 Sleep onset 12.60 0.27 -2.18 0.85 14.77 0.19 Sleep onset latency 0.30 0.84 -0.37 0.81 0.67 0.66 Sleep efficiency 0.33 0.73 -1.35 0.17 1.68 0.08 WASO -2.02 0.48 2.13 0.52 -4.15 0.20 Sleep offset 24.47 0.03 -0.23 0.98 24.70 0.03 Sleep inertia -1.11 0.29 1.50 0.16 -2.60 0.01

Beta (β) and p-values of patients vs siblings, patients versus controls and siblings versus controls; Table A: Random factor: relatedness; Covariates: age and gender Table B: Random factor: relatedness; Covariates: age, gender and IDS-SR score * Significant at Bonferroni-corrected significance level of 0.007

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls Figure 2 – Association between mean sleep offset and IDS score

76 77 The association also remained significant after elimination of the sleep questions supplementary table 2 for a summary of the regression models. Analyses in the from the IDS score (β=0.21, t=2.25 p=0.03 and β=0.29, t=2.75, p=0.008). euthymic subsample resulted in a non-significant association between WASO and suicidal behaviour (β=0.26, p=0.05) and similar non-significant associations Differences in sleep parameters compared to euthymic patients only between sleep and the remaining illness characteristics. To account for the relatively large dispersion in current mood symptomatology, the analyses were repeated in participants scoring under the cut-off indicative of Table 4 – Association of sleep with illness characteristics mild depressive and manic symptoms (IDS-SR < 26 and ASRM < 6)(Altman, Age at onset Psychosis Total nr of episodes Suicidal behaviour Hedeker, Peterson, & Davis, 2001; Karsten, Hartman, Ormel, Nolen, & Penninx, β p β p β p β p 2010). This subsample consisted of 87 patients, 71 siblings and 75 controls. Sleep duration 0.12 0.31 -0.05 0.63 -0.04 0.71 0.10 0.39 Current depression symptom score was not significantly different between groups Sleep onset -0.15 0.19 0.02 0.85 -0.01 0.92 0.04 0.75 (IDS-SR mean ± sd ; patients: 11.8 ± 7.3, siblings: 6.2 ± 5.2 and controls: 5.7 Sleep onset latency 0.08 0.52 -0.09 0.39 -0.19 0.08 -0.07 0.57 Sleep efficiency 0.11 0.33 0.08 0.41 0.23 0.03 -0.26 0.03 ± 5.0). Also, no group differences on manic symptoms were found (ASRM mean WASO -0.09 0.44 -0.13 0.19 -0.22 0.05 0.38 0.001* ± sd ; patients: 1.6 ± 1.5, siblings: 1.1 ± 1.2 and controls: 1.2 ± 1.5), nor on the Chapter 3 Sleep offset -0.03 0.83 -0.03 0.75 -0.06 0.61 0.15 0.20 number of days work (patients 4.4 ± 3.8, siblings 5.2 ± 3.6, 5.6 ± 4.0). When Sleep inertia 0.05 0.66 0.01 0.90 -0.17 0.12 0.13 0.27

comparing the groups on the sleep parameters, the only significant difference we Beta (β) and p-values of association between illness characteristics and sleep pattern β found was between patients and siblings on timing of sleep offset ( =38.66, t=3.61 * Significant at Bonferroni-corrected significance level of 0.007 p=0.002). However, after adjusting for IDS-SR score, the group difference in the sleep offset model did not reach the required Bonferroni-corrected significance Meta-analysis level of 0.007 (β= 23.30, t=2.06, p = 0.05). Adjusting for IDS-SR score in the To facilitate the interpretation of our findings we pooled the data of all available 6 remaining models (i.e. sleep duration, sleep onset, sleep onset latency, sleep studies using actigraphy. We calculated the standardized mean differences efficiency, WASO and sleep inertia) did not change the results. Adding number (SDM) between patients and controls of sleep duration, sleep onset latency, of work days as covariate to all previous models did not change any of the results. WASO and sleep efficiency. The meta-analysis of sleep duration (10 studies; N= 289), showed a significant difference in SMD between bipolar patients and Life-time illness characteristics controls of 0.52 (SE=0.13, z=3.86, p<0.001). Sleep onset latency (8 studies; None of the actigraphy sleep pattern variables was significantly associated with N=239) was higher in bipolar patients compared to controls with an SMD of 0.31 age at onset of the bipolar illness. Also, a history of psychotic symptomatology (SE=0.09, z=3.38, p=0.001). Sleep efficiency was lower in patients compared to during episodes was not related with the sleep pattern variables. Total number controls (8 studies, N=239), with an SMD of -0.30 (SE=0.09, z=-3.17, p=0.002). of mood episodes was not related to the sleep pattern variables, as were Lastly, bipolar patients had a higher WASO compared to controls (8 studies; number of episodes by polarity and presence of rapid cycling. Only WASO N=239) with an SMD of 0.24 (SE=0.09, z=2.61, p=0.009). Current levels of

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls was significantly associated with history of suicidal behaviour. See table 4 and manic and depressive symptoms were not accounted for in this meta-analysis.

78 79 Discussion study selectively included bipolar patients with a comorbid insomnia diagnosis, these results may not reflect the sleep pattern of bipolar patients in general. Investigating a large bipolar I sample with a long actigraphy measurement period of 14 days, the current study aimed to expand previous findings on The direct association between measures of depression and sleep pattern in sleep disturbances in bipolar disorder. We found a longer sleep duration and patients suggests that sleep disturbances are a state-dependent phenomenon. later timing of sleep offset in patients compared to non-affected siblings, but Given that sleep disturbances are a core feature of depression (Harvey, 2008), it these findings were attributable to differences in current depressive symptoms. is not surprising that we found depressive symptoms to be associated with longer Euthymic bipolar patients and healthy controls did not differ on any of the sleep sleep duration and later sleep offset. These associations also held after removing parameter, nor was there a relation between sleep pattern and bipolar illness sleep symptoms items from depression assessment. The direct relationship characteristics after adjusting for depressive symptoms. between sleep and mood symptomatology is in keeping with previous actigraphy studies: Gershon et al. (2012) found coupling of sleep disturbances and negative When combining our data with results from the recent meta-analysis by affect in interepisodic patients and Bauer et al. (2006) found changes in sleep

Geoffroy et al. (2015) the overall evidence points in the direction of a lower and bed rest to precede changes in mood, with an average latency of one day. Chapter 3 sleep efficiency and higher sleep duration, WASO and sleep onset latency in Moreover, a meta-analysis of actigraphy in patients with depression concluded bipolar patients. This would indicate that bipolar disorder patients do show that these patients showed higher sleep efficiency and lower sleep duration after trait-like differences in sleep pattern compared to controls. Nevertheless, treatment of depressive symptoms (Burton et al., 2013). The coupling between the effect sizes from the meta-analysis are small, with the exception of sleep sleep duration and depressive symptoms appears stronger for bipolar patients duration that showed a medium effect size. Whether such longer sleep and their siblings as compared to controls, which is in accordance with the duration is indicative of a sleep disorder is questionable, considering that the longitudinal measures by Gershon et al. (2012). However, a similar association mean sleep duration never exceeded 9 hours, which is still within the range between sleep offset and depressive symptoms was absent in siblings, suggesting of recommended sleep for adults (Hirshkowitz et al., 2015). A previous meta- that mood and sleep are not indisputably related in siblings. analysis by Ng et al. (2014) found similar effect sizes for sleep duration, WASO and sleep onset latency, but no difference in sleep efficiency between bipolar By including non-affected siblings of bipolar patients, we addressed the question patients and controls. Also, bipolar patients did not show any deviations on the whether sleep disturbances are a heritable trait of bipolar disorder. Bipolar polysomnography measures, suggesting that sleep of remitted bipolar patients disorder has as an estimated heritability of roughly 80% and several disorder has similar restorative qualities as compared to controls. These results suggest characteristics are expressed at a higher rate in non-affected relatives than in the that any differences in sleep pattern between bipolar patients and controls, is general population (Barnett & Smoller, 2009). The extensive pedigree analysis of disputable clinical importance. Of note is that a retrospective study found by Pagani et al. (2015) revealed that several sleep phenotypes, among which persisting sleep disturbances across all phases of bipolar disorder, including the sleep duration and sleep offset, are significantly heritable. In agreement with our

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls inter-episode months (Kanady, Soehnera, & Harvey, 2015). However, as this observations, they also found that bipolar patients slept longer and woke up later

80 81 compared to non-affected relatives. However, we found no difference between be the selection of patients in our sample. While this is an inherent problem patients, non-affected siblings and controls, after adjustments for current mood of all cross-sectional studies, our sample consisted of patients who were fit symptom levels. These findings suggest that the sleep phenotypes, although being to participate in an elaborate assessment and were willing to participate in a heritable traits, are mood state-related rather than a trait of bipolar disorder. follow-up actigraphy study. Although we cannot rule out selection bias, there is also no reason to assume this, as the patients in the current actigraphy study Limitations did not differ from the overall sample of DBC-participants on demographic The current findings need to be interpreted in the context of several limitations. characteristics. Furthermore, a potential limitation is the difference in age First, although sleep in bipolar patients was not affected in our study, we cannot between siblings and controls. Age is known to influence sleep characteristics, conclude that general circadian rhythmicity is not at all altered in bipolar patients. with older subjects experiencing a lower sleep efficiency and a higher rate of Clinical guidelines on the maintenance treatment of bipolar disorder recommend insomnia (Ancoli-Israel, 2009). However, it is unlikely that this age difference restoring behavioural rhythmicity and sleep by, for example, prescribing sleep- has masked an effect in the sibling-control comparison, since controls were promoting medication and social rhythm therapy (Hirschfeld, 2005). Perhaps, our on average younger compared to the average age of siblings. We also need

sample of bipolar patients consisted of successfully treated bipolar patients who do to point out that other than sleep apnoea, we did not inquire other somatic Chapter 3 not show overt signs of circadian rhythm disturbances, while other manifestations or psychological conditions that may have resulted in distorted sleep in the of circadian rhythms (nocturnal body temperature, sleep architecture, hormone participants. Finally, by selectively including patients with bipolar disorder type secretion) may still show irregularities. Future studies could address this question I we studied sleep in a clinically homogeneous sample. We cannot, however, by combining both behavioural and biological measures of circadian rhythmicity. extend our findings to other types of mood disorders.

Second, the effect of pharmacological treatment is one of the hardest hurdles to Conclusion overcome when studying sleep in bipolar disorder. Bipolar patients, even after In a large homogeneous sample of bipolar I sample, we did not identify being functionally recovered, are advised to continue psychotropic medication persisting sleep disturbances in bipolar patients other than related to increased to prevent relapse (Hirschfeld, 2005). Possibly, bipolar patients did not differ depressive symptoms. The meta-analysis points in the direction of longer sleep from controls due to the sedative side-effects of antipsychotic medication and duration, WASO and sleep onset latency and lower sleep efficiency in bipolar benzodiazepines or the normalizing effect of lithium on the 24-hour endogenous patients. The majority of these differences in the meta-analysis were however rhythmicity (Klemfuss, 1992). Given that the majority of patients used more small, the overall number of participants limited and the question remains than one type of medication, delineating the specific effect per medication type whether they reflect clinically significant sleep disturbances. Furthermore, non- was not possible. Studying non-affected siblings could serve as a reliable way of affected siblings showed normal sleep-wake patterns. Overall in our data sleep overcoming the problem of medication. It is noteworthy that these genetically disturbances are a state marker of mood symptomatology which underscores the susceptible individuals were not using lithium or antipsychotic medication and need for further study of the persistence of disturbance in sleep-wake patterns in

An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls also did not show any signs of sleep disturbances. A third limitation could euthymic bipolar disorder patients and in patients before disease onset.

82 83 Geoffroy, P. A., Boudebesse, C., Bellivier, F., Lajnef, M., Henry, C., Leboyer, M., … Etain, B. References (2014). Sleep in remitted bipolar disorder: A naturalistic case-control study using actigraphy. Journal of Affective Disorders, 158, 1–7.

Altman, E. G., Hedeker, D., Peterson, J. L., & Davis, J. M. (1997). The altman self-rating Geoffroy, P. A., Scott, J., Boudebesse, C., Lajnef, M., Henry, C., Leboyer, M., … Etain, B. mania scale. Biological Psychiatry, 42(10), 948–955. (2015a). Reply: Sleep in Patients with Remitted Bipolar Disorders: Analyses stratified on actigraphy devices, age and gender. Acta Psychiatrica Scandinavica, 131(5), 400. Altman, E. G., Hedeker, D., Peterson, J. L., & Davis, J. M. (2001). A comparative evaluation of three self-rating scales for acute mania. Biological Psychiatry, 50, 468–471. Geoffroy, P. A., Scott, J., Boudebesse, C., Lajnef, M., Henry, C., Leboyer, M., … Etain, B. (2015b). Sleep in patients with remitted bipolar disorders: a meta-analysis of actigraphy American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: studies. Acta Psychiatrica Scandinavica, 131, 89–99. DSM-5. Washington DC. Gershon, A., Thompson, W. K., Eidelman, P., McGlinchey, E. L., Kaplan, K. A., & Harvey, A. Ancoli-Israel, S. (2009). Sleep and its disorders in aging populations. Sleep Medicine, 10 SUPPL G. (2012). Restless Pillow, Ruffled Mind: Sleep and Affect Coupling in Interepisode Bipolar 1, S7–S11. Disorder. Journal of Abnormal Psychology, 121, 863–873.

Ancoli-Israel, S., Cole, R., Alessi, C., Chambers, M., Moorcroft, W., & Pollak, C. P. (2003). Harvey, A. G. (2008). Sleep and circadian rhythms in bipolar disorder: Seeking synchrony, The role of actigraphy in the study of sleep and circadian rhythms. Sleep, 26, 342–392. harmony, and regulation. American Journal of Psychiatry, 165, 820–829.

Andreasen, N. C., Flaum, M., & Arndt, S. (1992). The Comprehensive Assessment of Symptoms Harvey, A. G., Schmidt, D. A., Scarnà, A., Semler, C. N., & Goodwin, G. M. (2005). Sleep- Chapter 3 and History (CASH). An instrument for assessing diagnosis and psychopathology. Archives related functioning in euthymic patients with bipolar disorder, patients with insomnia, and of General Psychiatry, 49(8), 615–623. subjects without sleep problems. American Journal of Psychiatry, 162, 50–57.

Barnett, J. H., & Smoller, J. W. (2009). The genetics of bipolar disorder. Neuroscience, 164, Hirschfeld, R. M. A. (2005). Guideline watch: Practice guideline for the treatment of patients with 331–343. bipolar disorder, 2nd edition. Arlington, VA: American Psychiatric Association.

Bauer, M., Grof, P., Rasgon, N., Bschor, T., Glenn, T., & Whybrow, P. C. (2006). Temporal Hirshkowitz, M., Whiton, K., Albert, S. M., Alessi, C., Bruni, O., DonCarlos, L., … Adams relation between sleep and mood in patients with bipolar disorder. Bipolar Disorders, 8(2), Hillard, P. J. (2015). National Sleep Foundation’s sleep time duration recommendations: 160–167. methodology and results summary. Sleep Health, 1, 40–43.

Belmaker, R. H. (2004). Bipolar disorder. The New England Journal of Medicine, 351, 476–486. Jackson, A., Cavanagh, J., & Scott, J. (2003). A systematic review of manic and depressive prodromes. Journal of Affective Disorders, 74, 209–217. Boudebesse, C., Geoffroy, P. A., Bellivier, F., Henry, C., Folkard, S., Leboyer, M., & Etain, B. (2014). Correlations between objective and subjective sleep and circadian markers in Jones, S. H., Hare, D. J., & Evershed, K. (2005). Actigraphic assessment of circadian activity remitted patients with bipolar disorder. Chronobiology International, 2(315), 698–704. and sleep patterns in bipolar disorder. Bipolar Disorders, 7, 176–186.

Burton, C., McKinstry, B., Szentagotai Tatar, A., Serrano-Blanco, A., Pagliari, C., & Wolters, Jones, S. H., Tai, S., Evershed, K., Knowles, R., & Bentall, R. (2006). Early detection of M. (2013). Activity monitoring in patients with depression: A systematic review. Journal of bipolar disorder: A pilot familial high-risk study of parents with bipolar disorder and their Affective Disorders, 145(1), 21–28. adolescent children. Bipolar Disorders, 8, 362–372.

Cretu, J. B., Culver, J. L., Goffin, K. C., Shah, S., & Ketter, T. a. (2016). Sleep, residual mood Kanady, J. C., Soehnera, A. M., & Harvey, A. G. (2015). A Retrospective Examination of Sleep symptoms, and time to relapse in recovered patients with bipolar disorder. Journal of Affective Disturbance across the Course of Bipolar Disorder. Advances in Psychiatric Treatment, 30(4), Disorders, 190, 162–166. 1000193.

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1997). Structured Clinical Kaplan, K. a, Talbot, L. S., Gruber, J., & Harvey, A. G. (2012). Evaluating sleep in bipolar Interview for DSM-IV Axis I Disorders (SCID-I). New York: Biometric Research An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls disorder: comparison between actigraphy, polysomnography, and sleep diary. Bipolar Department. Disorders, 14, 870–879.

84 85 Karsten, J., Hartman, C. A., Ormel, J., Nolen, W. A., & Penninx, B. W. J. H. (2010). Salvatore, P., Ghidini, S., Zita, G., De Panfilis, C., Lambertino, S., Maggini, C., & Subthreshold depression based on functional impairment better defined by symptom severity Baldessarini, R. J. (2008). Circadian activity rhythm abnormalities in ill and recovered than by number of DSM-IV symptoms. Journal of Affective Disorders, 123, 230–237. bipolar I disorder patients. Bipolar Disorders, 10, 256–265.

Klemfuss, H. (1992). Rhythms and the pharmacology of lithium. Pharmacology & Therapeutics, Seleem, M. A., Merranko, J. A., Goldstein, T. R., Goldstein, B. I., Axelson, D. A., Brent, D. a, 56, 53–78. … Birmaher, B. (2015). The longitudinal course of sleep timing and circadian preferences in adults with bipolar disorder. Bipolar Disorders, 17(4), 392–402. Leverich, G. S., Nolen, W. A., Rush, A. J., McElroy, S. L., Keck, P. E., Denicoff, K. D., … Post, R. M. (2001). The Stanley Foundation Bipolar Treatment Outcome Network: I. Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., … Longitudinal methodology. Journal of Affective Disorders, 67, 33–44. Dunbar, G. C. (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview for DSM-IV Milhiet, V., Etain, B., Boudebesse, C., & Bellivier, F. (2011). Circadian biomarkers, circadian and ICD-10. Journal of Clinical Psychiatry, 59, 22–33. genes and bipolar disorders. Journal of Physiology Paris, 105, 183–189. St-Amand, J., Provencher, M. D., Bélanger, L., & Morin, C. M. (2013). Sleep disturbances in Millar, A., Espie, C. a., & Scott, J. (2004). The sleep of remitted bipolar outpatients: A bipolar disorder during remission. Journal of Affective Disorders, 146, 112–119. controlled naturalistic study using actigraphy. Journal of Affective Disorders, 80, 145–153. Suppes, T., Leverich, G. S., Keck, P. E., Nolen, W. A., Denicoff, K. D., Altshuler, L. L., Ng, T. H., Chung, K. F., Ho, F. Y. Y., Yeung, W. F., Yung, K. P., & Lam, T. H. (2014). Sleep- … Post, R. M. (2001). The Stanley Foundation Bipolar Treatment Outcome Network wake disturbance in interepisode bipolar disorder and high-risk individuals: A systematic - II. Demographics and illness characteristics of the first 261 patients. Journal of Affective review and meta-analysis. Sleep Medicine Reviews, 20, 46–58. Disorders, 67, 45–59. Chapter 3 Nurnberger, J. I., Adkins, S., Lahiri, D. K., Mayeda, A., Hu, K., Lewy, A., … Davis-Singh, D. (2000). Melatonin suppression by light in euthymic bipolar and unipolar patients. Archives of General Psychiatry, 57, 572–579.

Pagani, L., St. Clair, P. A., Teshiba, T. M., Service, S. K., Fears, S. C., Araya, C., … Freimer, N. B. (2016). Genetic contributions to circadian activity rhythm and sleep pattern phenotypes in pedigrees segregating for severe bipolar disorder. Proceedings of the National Academy of Sciences, 113(6), E754–E761.

Pierce, C. A. (2008). Comprehensive Meta-analysis. Organizational Research Methods, 11, 188– 191.

R Development Core Team, R. (2014). R: A Language and Environment for Statistical Computing.

Rocha, P. M. B., Neves, F. S., & Corrêa, H. (2013). Significant sleep disturbances in euthymic bipolar patients. Comprehensive Psychiatry, 54(7), 1003–1008.

Rush, A. J., Carmody, T., & Reimitz, P. (2000). The Inventory of Depressive Symptomatology (IDS): Clinician (IDS-C) and Self-Report (IDS-SR) ratings of depressive symptoms. International Journal of Methods in Psychiatric Research, 9, 45–59.

Sadeh, A., Hauri, P. J., Kripke, D. F., & Lavie, P. (1995). The role of actigraphy in the evaluation of sleep disorders. Sleep, 18, 288–302. An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls

86 87 Supplementary table 2 – Association of sleep with number of manic episodes, depressive Supplementary material episodes and rapid cycling Nr of manic episodes Nr of depr episodes Rapid cycling Supplementary table 1 – Analysis of variability in sleep measures β p β p β p

A Patients Siblings Controls Sleep duration -0.03 0.76 -0.04 0.74 -0.03 0.74 M SD M SD M SD Sleep onset 0.02 0.88 -0.03 0.79 0.09 0.39 Sleep duration 82.49 41.0 67.29 24.83 76.12 33.32 Sleep onset latency -0.21 0.04 -0.03 0.80 -0.10 0.36 Sleep onset 63.74 37.17 54.59 27.21 62.35 31.48 Sleep efficiency 0.18 0.08 0.17 0.11 0.06 0.56 Sleep onset latency 9.81 10.90 10.26 14.99 12.16 29.70 WASO -0.16 0.13 -0.19 0.08 -0.02 0.85 Sleep efficiency 4.60 3.69 5.15 4.66 4.82 4.09 Sleep offset -0.02 0.88 -0.06 0.56 0.05 0.63 WASO 20.26 13.19 19.57 13.75 20.73 18.03 Sleep inertia -0.15 0.15 -0.10 0.34 -0.13 0.20

Sleep offset 71.49 34.90 57.35 24.98 68.39 35.45 Beta (β) and p-values of association between illness characteristics and sleep pattern Sleep inertia 9.42 11.45 10.13 10.53 8.73 14.15 B Patients vs Siblings Patients vs Controls Siblings vs Controls β p β p β p Chapter 3 Sleep duration 13.37 0.02 7.68 0.15 5.69 0.33 Sleep onset 7.12 0.16 3.08 0.53 4.04 0.46 Sleep onset latency -0.51 0.87 -2.26 0.45 1.75 0.59 Sleep efficiency -0.59 0.36 -0.17 0.78 -0.41 0.54 WASO 0.17 0.94 -0.09 0.97 0.26 0.92 Sleep offset 11.42 0.03 5.38 0.27 6.04 0.25 Sleep inertia -0.93 0.62 0.91 0.62 -1.84 0.36 C Patients vs Siblings Patients vs Controls Siblings vs Controls β p β p β p Sleep duration 8.38 0.15 1.93 0.74 6.44 0.26 Sleep onset 4.16 0.45 -0.21 0.97 4.37 0.42 Sleep onset latency -0.51 0.87 -2.26 0.44 1.76 0.59 Sleep efficiency -0.67 0.34 -0.26 0.70 -0.40 0.56 WASO -2.68 0.26 -3.24 0.20 0.56 0.82 Sleep offset 9.14 0.10 2.84 0.59 6.30 0.23 Sleep inertia -0.39 0.85 1.51 0.46 -1.90 0.34

Table A: Means and standard deviations of variability in sleep measures Suppl. Figure 1 – Actograms of a bipolar patient (A), non-affected sibling (B) and healthy Table B: Beta (β) and p-values of patients vs siblings, patients versus controls and siblings versus controls; Random factor: relatedness; Covariates: age and gender control (C). Black lines indicate activity levels per minute. Dark blue intervals indicate sleep and Table C: Beta (β) and p-values of patients vs siblings, patients versus controls and siblings versus controls; lighter blue intervals indicate sleep onset latency (wake prior to sleep) or sleep inertia (wake after An actigraphy study investigating sleep in bipolar I patients, unaffected siblings and controls Random factor: relatedness; Covariates: age, gender and IDS-SR score sleep).

88 89 Chapter 4

The association of sleep and physical activity with integrity of white matter microstructure in bipolar disorder patients and healthy controls

Sanne Verkooijen, Remi Stevelink, Lucija Abramovic, Christiaan H. Vinkers, Roel A. Ophoff, René S. Kahn, Marco P. M. Boks, Neeltje E. M. van Haren

Psychiatry Research (2017), 262, 71-80 Abstract Introduction

We investigate how the sleep disruptions and irregular physical activity levels Bipolar disorder is a severe and recurrent psychiatric disorder, characterized by that are prominent features of bipolar disorder (BD) relate to white matter alternating depressive and manic mood episodes (Belmaker, 2004). Although microstructure in patients and controls. Diffusion tension imaging (DTI) and the aetiology of bipolar disorder remains poorly understood, there is an 14-day actigraphy recordings were obtained in 51 BD I patients and 55 age-and- extensive body of evidence suggesting that brain abnormalities are a core feature gender-matched healthy controls. Tract-based spatial statistics (TBSS) was used of the disorder. Previously, structural imaging studies found marked anatomical for voxelwise analysis of the association between fractional anisotropy (FA) and differences in grey matter volume between patients and healthy controls sleep and activity characteristics in the overall sample. Next, we investigated (Hallahan et al., 2011; Hibar et al., 2016). Moreover, volumes of several areas whether the relation between sleep and activity and DTI measures differed (i.e. amygdala, globus pallidus) were associated with polygenic risk scores for for patients and controls. Physical activity was related to increased integrity of bipolar disorder and are affected in first-degree relatives (Caseras, Tansey, Foley, white matter microstructure regardless of bipolar diagnosis. The relationship & Linden, 2015; Sandoval et al., 2014). In addition to volumetric measures, between sleep and white matter microstructure was more equivocal; we found diffusion tensor imaging (DTI) studies have shown that bipolar disorder is also an expected association between higher FA and effective sleep in controls but characterized by changes in white matter microstructure (Mahon, Burdick, & opposite patterns in bipolar patients. Confounding factors such as antipsychotic Szeszko, 2010). DTI quantifies the integrity and coherence of white matter

medication use are a likely explanation for these contrasting findings and tracts by measuring the diffusion of water parallel and perpendicular along Chapter 4 highlight the need for further study of medication-related effects on white white matter fibre tracts. Water diffuses more easily along white matter tracts matter integrity. than perpendicular to them. This directional dependence is usually quantified with fractional anisotropy (FA), a scalar variable which ranges from 0 to 1, in which higher values indicate a higher directional dependence of diffusion. More organized and myelinated axon will cause water molecules to diffuse in a particular direction rather than randomly in all directions. Therefore, FA is thought to reflect integrity of white matter with lower FA values pointing to sparse, poorly myelinated, or divergent fibres (Beaulieu, 2002). A voxel-based meta-analysis of 18 DTI studies found that bipolar patients have lower FA in the genu of the corpus callosum and left cingulum white matter, suggesting poor integrity of white matter (Wise et al., 2016). This finding has also been established in antipsychotic and mood-stabilizer naïve patients (Yip, Chandler, Rogers, Mackay, & Goodwin, 2013). The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls

92 93 In addition to these brain changes, disturbances in sleep patterns are considered relates to white matter microstructure and that sleep disruptions could lead to as core symptoms of bipolar mood episodes (American Psychiatric Association, changes in oligodendrocytes, myelination, and axonal integrity and organization. 2013). The majority of patients experience a reduced need for sleep during manic phases of the disorder, and hypersomnia or insomnia during depressive episodes Not only sleep, but also altered levels of physical activity are a hallmark of bipolar (Harvey, 2008). Sleep has several properties which are crucial for normal brain mood episodes: while depression is characterized by a marked reduction in overall functioning and integrity (Diekelmann & Born, 2010; Kang et al., 2009; Ooms physical activity, manic episodes stand out for their increased goal-directed et al., 2014; Tononi & Cirelli, 2006; Wang, Grone, Colas, Appelbaum, & behaviour and psychomotor agitation (American Psychiatric Association, 2013). Mourrain, 2011; Xie et al., 2013). In clinical studies of sleep disorders several Physical activity, defined as bodily movements produced by skeletal muscles neurophysiological alterations have been observed. For example, insomnia has (Caspersen, Powell, & Christenson, 1985), is thought to positively affect brain been linked to reductions in orbitofrontal and parietal grey matter and larger characteristics; higher levels of physical activity and physical fitness have rostral anterior cingulate cortex volumes (Altena, Vrenken, Van Der Werf, van been associated with increased grey matter volumes, both cross-sectionally, den Heuvel, & Van Someren, 2010; Winkelman et al., 2013), and narcolepsy and longitudinally, and after a physical exercise intervention (Bherer, Erickson, & obstructive sleep apnoea are associated with widespread reductions in grey matter Liu-Ambrose, 2013; Erickson, Leckie, & Weinstein, 2014). Indeed, a meta- volume and thickness (Joo et al., 2010, 2011; Joo, Tae, Kim, & Hong, 2009). analysis of 29 studies on the association between white matter and physical Furthermore, animal studies have demonstrated that sleep deprivation interferes activity in healthy older adults tentatively concluded that physical exercise leads

with long term synaptic potentiation and neurogenesis (Kopp, Longordo, to both global and local improvements in white matter (micro)structure (Sexton Chapter 4 Nicholson, & Lüthi, 2006; Kreutzmann, Havekes, Abel, & Meerlo, 2015; et al., 2015). Possible underlying mechanisms involve the increased expression Mirescu, Peters, Noiman, & Gould, 2006). Mechanisms of the link between and release of brain-derived neurotrophic factor and oxygen supply as a result sleep and integrity of white matter remain largely unknown. Animal studies have of increased physical activity (Burzynska et al., 2014). Results from an exercise shown that oligodendrocyte proliferation and myelin-related gene expression intervention study in schizophrenia patients indicated that after six months of strongly increase during sleep and reduce during wake time and experimental sleep training, the connectivity in white matter fibre tracts associated with motor deprivation (Bellesi et al., 2013). Conversely, chronic sleep loss has been associated functioning improved equally in the schizophrenia patients and a control group with down-regulation of the expression of two myelin-related genes coding for (Svatkova et al., 2015). Until now, similar studies in bipolar patients are absent. plasmolipin and a membrane protein expressed in the mature myelin sheath (Cirelli, Faraguna, & Tononi, 2006). Human studies using DTI measures found The current study investigates the link between white matter microstructure that variability in sleep duration during weekdays was associated with reduced FA and objectively measured physical activity and sleep patterns of bipolar patients in the parietocortical, frontocortical and frontostriatal tracts (Telzer, Goldenberg, and controls. Both physical activity and sleep are measured using actigraphy. Fuligni, Lieberman, & Gálvan, 2015). Furthermore, a sleep deprivation study in We hypothesize that both sleep and physical activity are positively related to adult males found widespread changes in DTI measures after just one day of sleep increased integrity of white matter, and we further explore whether these

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls deprivation (Elvsåshagen et al., 2015). These findings suggest that effective sleep associations are similar for patients and controls.

94 95 Methods International Neuropsychiatric Interview (M.I.N.I) (First, Spitzer, Gibbon, & Williams, 1997; Sheehan et al., 1998). Sample Participants were recruited from the Dutch Bipolar Cohort (DBC) study, which Actigraphy recordings and analyses is a collaboration between the University Medical Centre Utrecht (UMCU), Circadian rhythmicity and sleep-wake measurements were recorded with various health care institutes in the Netherlands and the University of California an Actiwatch (the Actiwatch 2; Philips Respironics Inc, Murrysville, PA, Los Angeles (UCLA). The objective of the DBC study is to collect a deep USA). The Actiwatch has a solid-state piezo-electric accelerometer and a phenotype characterization of bipolar I patients, their first-degree relatives and lithium rechargeable battery. It records wrist movements and the sum of wrist controls. After completion of the DBC protocol, we invited subjects by telephone movements is scored in epochs of 1 min. All participants were instructed to and mail who had indicated that they were available for future research wear the Actiwatch for a period of 14 consecutive days on their non-dominant and invited them to participate in two cross-sectional actigraphy and MRI wrist and only to remove it when exposed to water for long periods of time (e.g. measurements. The medical ethical committee of the UMCU approved the DBC swimming). During the 14-day recording period, participants kept a sleep diary study and the actigraphy and MRI protocols. Written informed consent was in which bed times, nap times and off-wrist periods were noted. obtained from all participants prior to participation. The majority of participants underwent MRI scanning prior to the two-week actigraphy protocol, with the To calculate the sleep and activity patterns, a series of algorithms in R

exception of 6 participants who started with actigraphy measurements. The statistical package (R Development Core Team, 2014) were used according to Chapter 4 mean time difference between MRI and actigraphy measurements was 1.4 years. the procedure developed by Pagani et al. (2015). These algorithms resulted in 7 sleep measures: sleep duration, timing of sleep onset (i.e. minutes prior or All participants had a minimum age of 18 years old. The inclusion criterion for after midnight), timing of sleep offset (i.e. minutes after midnight), sleep onset patients was a diagnosis of bipolar disorder type I, in order to create a clinically latency, sleep efficiency (i.e. minutes asleep divided by minutes in bed), wake homogeneous sample. Controls with a bipolar or psychotic diagnosis were after sleep onset (WASO) and sleep inertia (i.e. wake interval between sleep excluded, as were control subjects who had a first or second degree relative with offset and time out of bed). Mean activity level was calculated over 24 hours, a bipolar or psychotic diagnosis. Any other type of psychiatric disorder was followed by mean activity levels in four 6-h periods corresponding to morning, not an exclusion criterion in order to avoid the recruitment of a ‘super-normal afternoon, evening, night time: midnight to 6 AM (24-6 h), 6 AM to noon (6 – control sample’, that is not representative of the general population. Exclusion 12 h), noon to 6 PM (12 – 18 h), and 6 PM to midnight (18 – 24 h). To measure criteria for all participants were sleep apnoea, pregnancy, head trauma and the degree of day-to-day stability of the 24-h activity rhythm, we calculated neurological disorders. An independent radiologist blindly evaluated all MRI interdaily stability (IS), which is the ratio between the variance of the average scans. Participants with any clinical findings were excluded from further 24-h pattern around the mean and the overall variance (Van Someren et al., analyses. Patient diagnoses were verified using the Structured Clinical Interview 1996). IS ranges from 0 to 1, with higher values indicating more stable activity

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls for DSM-IV (SCID) and controls subjects were assessed using the Mini- rhythms.

96 97 MRI acquisition of all subjects were non-linearly aligned to a standard FA template (FMRIB55_ Structural Magnetic Resonance Images were acquired on a 3 Tesla Philips FA; the average of 55 good quality healthy subjects) after which the average Achieva scanner (Philips Healthcare, Best, the Netherlands), equipped with was computed. The resulting average FA map was thinned to create a white an 8-channel SENSE-headcoil. Fast field echo scans with 200 contiguous matter ‘skeleton’, representing the centre of white matter tracts common to all sagittal slices (TE=4.6 ms, TR=10 ms, flip angle=8˚, FOV=240 mm, 0.75 x 0.75 subjects. This white matter skeleton was thresholded at FA>0.2 to exclude non- x 0.80 mm³ voxels) were obtained. Processing was done on the neuroimaging white matter and voxels in extremities where there is too much cross-subject computer network of the University Medical Centre Utrecht - Brain Centre variability in alignment, resulting in a white matter skeleton of 141,230 voxels. Rudolf Magnus, Utrecht, the Netherlands. All MRI analyses were conducted Next, each subject’s FA data was projected onto the white matter skeleton; for using the FMRIB Software Library (FSL v5.0) (Jenkinson, Beckmann, each subject the highest FA value perpendicular to each voxel of the skeleton Behrens, Woolrich, & Smith, 2012). Diffusion weighted images were pre- (i.e. the individual’s local white matter tract centre) was projected onto the white processed using FMRIB’s Diffusion Toolbox (FDT). First,topup was used to matter skeleton. correct for susceptibility induced distortions (distortions caused by the magnetic susceptibility inhomogeneities in the subject’s head), using two non-diffusion Statistical analyses weighted (b-value=0) images with opposite phase-encoding directions (anterior Whole-brain voxelwise analyses were carried out using ‘randomise’ in FSL to posterior and opposite), which thus have distortions going in opposite v5.0, a tool for permutation-based non-parametric testing using the general

directions. The susceptibility-induced off-resonance field (distortions caused by linear model (Nichols and Holmes, 2002; https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/ Chapter 4 inhomogeneities to magnetic susceptibility of the subject’s head) was estimated GLM). All statistical tests were performed with 5000 permutations. Our using a method similar to that described in Andersson et al. (2003) and the two statistical threshold was a threshold-free cluster enhancement (TFCE) and images were combined into a single corrected one. Next, eddy was used to correct family-wise error corrected P-value of 0.05. (Smith and Nichols, 2009). First, for eddy-current distortion and head movements, and bet (brain-extraction tool) we investigated the difference in FA between patients and controls for each was used to exclude any non-brain tissue (Smith, 2002). Finally, dtifit was used voxel. The MNI coordinates of the centre of gravity of significant clusters were to fit a diffusion tensor model at each voxel. calculated and the anatomical regions of these clusters were identified using the ‘JHU ICBM-DTI-81 White-Matter Labels’ atlas, similar to Liu et al. (2016). TBSS analysis Next, we analysed the association between actigraphy measures and FA for Tract-based spatial statistics (TBSS) was used for voxelwise comparison each voxel, in the overall sample of patients and healthy controls. Furthermore, of fractional anisotropy (FA), between subjects (Smith et al., 2007, 2006). we investigated whether actigraphy measures were differently associated with Compared to the other DTI metrics (i.e. axial diffusivity, radial diffusivity and DTI measures in patients and controls. For this, we analysed the association mean diffusivity), FA is the most elaborate studied diffusion parameter and between an interaction term for Group*actigraphy measures with FA, again correlates well with microscopic changes in white matter such as myelination, using the general linear model as implemented in FSL. In case of significant

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls axonal diameter, density and organization (Beaulieu, 2009). First, the FA maps voxels in the overall sample, we created scatterplots to visualize the association

98 99 Table 1 – Sample characteristics between average FA of the significant clusters and actigraphy measures in Patients Controls IBM SPSS Statistics 21.0. For variables with significant interaction effects, we N=51 N=55 F / χ2 p-value

created stratified scatterplots to determine the direction of association between Age M (sd) 49.5 (11.4) 45.5 (15.8) 2.18 0.14 actigraphy measures and the average FA of significant clusters. We analysed Gender M/F (% male) 28/23 (54.9%) 25/30 (45.5%) 0.44 0.22 the confounding effect of illness-specific variables (number of episodes, history Handedness R/L/B (% right) 42/6/3 (82.4%) 44/11/0 (80.0%) 4.37 0.11 of psychotic symptoms, illness duration and use of psychotropic medication) IQ M (sd) 99.6 (15.7) 105.7 (14.4) 4.18 0.04* Level of education n (%) 1.28 0.53 by testing the correlation with average FA of the significant cluster and the 1. Low education 13 (25.5%) 9 (16.7%) actigraphy measures in SPSS Statistics 21.0. In case these variables showed 2. Intermediate education 16 (31.4%) 20 (37.0%) significant associations with both FA and actigraphy measures, they were 3. Higher education 22 (43.1%) 25 (46.3%) considered as potential confounders and separately entered as covariates in the Inventory of Depressive Symptoms M (sd) 14.3 (10.1) 5.2 (3.9) 35.60 <0.001* stratified patient regression analyses, using the Enter method, with all variables Altman Self-Rating Mania Scale M (sd) 1.9 (1.8) 1.7 (2.3) 0.41 0.53 entered in one block. Body Mass Index M (sd) 27.4 (4.3) 25.2 (3.3) 8.45 0.01* Nr of workdays M (sd) 4.2 (3.5) 5.7 (3.7) 4.11 0.05* Illness duration/ years M (sd) 16.5 (14.2) Results Number of episodes M (sd) 10.0 (10.8) Age at onset M (sd) 31.7 (12.2) Participants Chapter 4 History of psychotic symptoms N (%) 35 (70.0%)

A total of 106 participants were included (51 patients, 55 controls). See table * Significant group difference at p<0.05 1 for sample characteristics. No differences between patients and controls were found for age, gender or handedness. Patients worked on average fewer Patients showed widespread reductions in FA compared to controls (Figure 1 days during the actigraphy measurement period and had higher BMI levels as and Table 2). These differences were located in the corpus collosum, cerebral compared with controls. Although none of the patients reported being in a mood peduncle, internal capsule, corona radiata, thalamic radiation, external capsule episode, patients reported significantly more current depressive symptoms. and the posterior limb of internal capsule.

With the exception of 1, all patients used between 1-4 types of psychotropic Mean scores and standard deviations of the 7 sleep measures and 5 activity medication during the actigraphy measurements; 24 patients used lithium and 17 measures are shown in table 3. Patients had on average a longer sleep duration patients used other mood stabilizing agents. Furthermore, 15 patients received (F[1,102]=6.6, p=0.01, a longer wake time after sleep onset (F[1,102]=5.1, p=0.03) antipsychotic medication, 9 patients used antidepressants and 19 patients used and lower activity counts between 6 PM and midnight (F[1,102]=6.1, p=0.02). benzodiazepines. Of the control subjects, 1 person received an antidepressant However, after corrections for multiple testing, none of the group difference and 1 a benzodiazepine. remained significant at a Bonferroni-corrected significance level of 0.004. The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls

100 101 Table 2 – Cluster size (number of voxels), p-value, t-value, MNI coordinates of centre of gravity and anatomical location of significant differences in FA between patients and controls associated with increased FA (Supplementary figure 1 and supplementary table

MNI 3). These correlations were found in largely overlapping regions of the brain, Cluster Corrected coordinates most of all in the genu and body of the corpus callosum and the right anterior Analysis size p-value T-value (x; y; z) Anatomical region

Significant 44232 <0.001 3.08 2.37; -21.4; 16.8 Genu of corpus callosum corona radiate (Table 4). There were no significant interactions between group group difference Body of corpus callosum and activity measures with any of the DTI measures, suggesting that the in FA between patients and Splenium of corpus callosum association between activity measures and DTI measures was not significantly controls R cerebral peduncle different for patients and controls. L cerebral peduncle R posterior limb of internal capsule Sleep measures L posterior limb of internal capsule L retrolenticular part of internal Voxelwise correlations in the overall sample (corrected for age, gender, and capsule patient/control group) revealed no significant correlations between sleep R anterior corona radiata measures and FA. Analyses of the interaction between group and sleep on L anterior corona radiata R superior corona radiata FA revealed that the association of three sleep measures (i.e. sleep duration, L superior corona radiata sleep onset, and sleep inertia) with FA differed between patients and healthy Right posterior corona radiata controls in several regions of the brain (Figure 3). These interaction effects of L posterior corona radiata Chapter 4 R posterior thalamic radiation (include optic radiation) L posterior thalamic radiation (include optic radiation) R external capsule L external capsule L cingulum (cingulate gyrus) R superior longitudinal fasciculus L superior longitudinal fasciculus

Activity measures Figure 1 – TBSS was used to assess the difference in FA between patients and controls. Significant correlations are shown in yellow to red, depending on the significance level. The Voxelwise correlations in the overall sample (corrected for age and gender, and white matter skeleton is shown in green and overlaid on the MNI152 template brain. Images are patient/control group) revealed positive associations between FA and mean shown in radiological convention (the left is displayed on the right) and corresponding Montreal activity throughout the day, activity between noon and 6 PM, and between Neurological Institute (MNI) coordinates are displayed below. P-values are TFCE and FWE

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls 6 PM and midnight (Figure 2). Also, a more stable 24-h activity rhythm was corrected for multiple comparisons across space, and corrected for age and gender.

102 103 sleep duration and sleep inertia with FA were primarily localized in the genu, average FA did not alter the results (Supplementary table 1B). Antipsychotic body and splenium of the corpus collosum and the bilateral corona radiata drug use was associated with sleep duration and average FA (Supplementary (Table 4). The interaction effect of sleep onset with group on FA was localized table 2). Subsequent addition of antipsychotic drug use as covariate to the in a small cluster in the left superior longitudinal fasciculus. After stratifying model rendered the previous association between sleep duration and average for patients and controls, we found that mean FA of the voxels with a significant FA in patients non-significant (Supplementary 1B). When the TBSS analysis interaction effect was positively correlated with sleep duration in controls, while for sleep duration was restricted to patients not using antipsychotic medication mean FA was negatively associated with sleep duration in patients (Figure 4A). (N=25 patients and 55 controls), the interaction effect between sleep duration Similar contrasting relations were found for sleep onset and sleep inertia, with and group was no longer significant providing further support for a confounding positive associations in patients and negative associations in controls (Figure effect of medication. The interaction effect between sleep duration and group 4B and 4C). We analysed the possibility of a U-shaped relation between sleep remained significant when the TBSS analysis was performed in patients using duration and FA and found that squaring, or centring plus squaring the sleep antipsychotic medication, albeit in a smaller area (Supplementary figure 2 and duration variable yielded similar results (Supplementary table 1C). We also supplementary table 3). tested whether there is a difference in association between the 22 patients that could be classified as long sleepers (> 8 hours) vs. 25 patients classified as normal sleepers (6-8 hours). We found that the association between sleep duration and Table 3 – Difference between sleep and activity measures for patients and controls

FA in ‘normal sleepers’ (β=-0.27, p=0.10, R2=0.53) was lower compared to long Patients Controls p-value Chapter 4 sleepers (β=-0.43, p=0.06, R2 =0.21), but still in the same direction, indicating Sleep duration M (sd) 475.2 (86.3) 447.2 (40.4) 0.01 Sleep onset M (sd) -0.50 (75.3) 17.6 (73.0) 0.22 that the relation between sleep duration and FA is linear. Sleep onset latency M (sd) 5.0 (5.7) 4.9 (5.2) 0.89 Sleep offset M (sd) 474.7 (88.5) 464.8 (74.8) 0.33 Confounder analyses Sleep inertia M (sd) 6.8 (7.2) 5.4 (6.3) 0.37 Stratified analyses of the sleep parameters in patients only allowed us to examine Wake after sleep onset M (sd) 59.6 (27.0) 48.5 (19.5) 0.03 Sleep efficiency M (sd) 84.8 (8.1) 87.3 (4.9) 0.12 the possible confounding effects of medication and illness characteristics. First, we analysed the association between the sleep parameters with average FA of Mean Activity M (sd) 204.7 (89.3) 234.4 (54.5) 0.08 the significant clusters resulting from the TBSS analysis (Supplementary table Activity 0 to 6 h M (sd) 42.1 (36.6) 45.0 (41.5) 0.82 1A). Next, we analysed the association of the possible confounders with both Activity 6 to 12 h M (sd) 227.3 (127.8) 260.8 (91.6) 0.12 sleep parameter and white matter integrity outcomes (Supplementary table 2). Activity 12 to 18 h M (sd) 333.4 (143.1) 361.8 (93.9) 0.34 Sleep onset and average FA were significantly associated with the number of Activity 18 to 24 h M (sd) 215.6 (109.6) 270.4 (84.9) 0.02 Means (M) and standard deviations (sd) on sleep and activity measures. Models are corrected for age and episodes. History of psychotic symptoms was also significantly associated with gender. Sleep duration, sleep onset latency, sleep inertia and WASO are measured in minutes. Sleep onset sleep onset and average FA. However, adding number of episodes and history and sleep offset describe minutes prior or past midnight. Sleep efficiency is a percentage ranging from 1-100. All activity variables are activity counts averaged per minute. The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls of psychotic symptoms as covariate to the association between sleep onset and

104 105 Table 4 – Cluster size (number of voxels), p-value, t-value, MNI coordinates of centre of gravity and anatomical location of significant associations between sleep and activity parameters with FA

Cluster Corrected MNI coordinates Analysis size p-value T-value (x; y; z) Anatomical region Significant Group*Sleep 651 0.04 3.74 10.2; 20.5; 19.5 Genu of corpus callosum duration interaction with FA 16 0.05 2.34 -6.25; 12.7; 22 Body of corpus callosum Significant Group*Sleep 292 0.039 4.36 37.6; -11.9; 29.4 R superior longitudinal fasciculus onset interaction with 266 0.038 4.85 -37; -15.6; 27.2 L superior longitudinal fasciculus FA 134 0.047 3.39 2.29; 24.9; 13.2 Genu of corpus callosum 75 0.046 3.78 18.1; -51.6; 30.4 Splenium of corpus callosum R posterior corona radiata 45 0.047 3.63 12.2; 37.9; 21 Splenium of corpus callosum 33 0.047 4.96 38; -40.5; 24.8 R superior longitudinal fasciculus Significant Group*Sleep 21826 0.006 3.46 12.2; -13.3; 17.2 Genu of corpus callosum inertia interaction Body of corpus callosum with FA Splenium of corpus callosum 227 0.047 3.10 -11.6; 8.42; -7.77 Cerebral peduncle Posterior limb of left internal capsule 202 0.049 2.19 8.01; -26; -28.8 R corticospinal tract Pontine crossing tract Chapter 4 R cerebral peduncle 162 0.046 3.96 -7.18; -28.3; -30.7 L corticospinal tract Pontine crossing tract L cerebral peduncle 104 0.049 2.63 -29.4; -23.4; -1.7 Retrolenticular part of left Figure 2 – Daytime activity is positively associated with integrity of white matter internal capsule Sagittal stratum microstructure. TBSS was used to assess the association of activity counts (per minute) with Fornix (cres) / Stria terminalis integrity of white matter microstructure. In the combined sample (patients and healthy controls) 78 0.049 2.58 -37.9; -47.6; 1.23 Posterior thalamic radiation we found significant positive correlations of mean activity of the whole day (A), activity between Sagittal stratum noon and 6PM (B) and activity between 6PM and midnight (C). Significant correlations 38 0.05 3.57 -35.1; -58.4; -5.16 Posterior thalamic radiation Sagittal stratum are shown in yellow to red, depending on the significance level. The white matter skeleton is 15 0.05 1.93 -40.6; -25.7; -4.6 Sagittal stratum shown in green and overlaid on the MNI152 template brain. Images are shown in radiological 13 0.05 2.39 -32; -203; 0.769 L external capsule convention (the left is displayed on the right) and corresponding Montreal Neurological Institute Significant correlation 3153 0.019 2.84 9.01; 19.2; 18.7 Genu of corpus callosum of mean activity with Body of corpus callosum (MNI) coordinates are displayed below. P-values are TFCE and FWE corrected for multiple FA R anterior corona radiata comparisons across space and corrected for age, gender and (patient/control) group. Significant correlation 2466 0.028 3.24 7.37; 17; 19.7 Genu of corpus callosum of activity from 12h to Body of corpus callosum 18h with FA R anterior corona radiata Significant correlation 333 0.044 3.46 3.64; 25.7; 10.9 Genu of corpus callosum

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls of activity from 18h to Body of corpus callosum 24h with FA

106 107 Chapter 4

Figure 3 – The association of sleep measures with integrity of white matter microstructure differs Figure 4 – Sleep measures are associated with integrity of white matter microstructure within between patients and controls. TBSS was used to assess the interaction effects of sleep variables the patient group as well as within the control group. Scatterplots showing the association of and group (Group * Sleep) with integrity of white matter microstructure. There is a significant sleep measures with average FA of the voxels with a significant interaction effect. Sleep duration negative interaction between sleep duration*group and FA (A), a positive interaction between (measured in minutes) is negatively associated with FA in patients and positively associated with sleep onset*group and FA (B), a positive interaction between sleep inertia and FA (C). Positive FA in controls (A). Sleep onset (minutes prior or post midnight) is positively associated with FA interaction effects are defined as interactions where the slope between sleep variables and DTI in patients and negatively associated with FA in controls (B). Similarly, sleep inertia (measured measures is higher for patients than for controls. Significant correlations are shown in yellow in minutes) is positively associated with FA in patients and negatively associated with FA in to red, depending on the significance level. The white matter skeleton is shown in green and controls (C). overlaid on the MNI152 template brain. Images are shown in radiological convention (the left is displayed on the right) and corresponding Montreal Neurological Institute (MNI) coordinates

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls are displayed below. P-values are TFCE and FWE corrected for multiple comparisons across space, and corrected for age and gender.

108 109 Discussion disorder patients. This is relevant as integrity of white matter is hypothesized to be a factor in the aetiology of bipolar disorder (Heng, Song, & Sim, 2010; The current study investigated the association between sleep and physical Nortje, Stein, Radua, Mataix-Cols, & Horn, 2013). It raises the question activity patterns and integrity of white matter microstructure in bipolar whether symptoms and disease progression of bipolar disorder can be disorder. Bipolar patients were characterized by widespread reductions in ameliorated by improving integrity of white matter through physical fitness fractional anisotropy (FA) in several major white matter tracts, including the therapy or behavioural activation. Although the effect of such therapies has corpus collosum, cerebral peduncle, corona radiata, internal and external not yet been studied in a bipolar sample, a first study in schizophrenia patients capsule, thalamic radiation and superior longitudinal fasciculus. Furthermore, found that an exercise intervention increased the level of white matter integrity we found that daytime and evening physical activity correlated with higher (Svatkova et al., 2015). FA in bipolar patients and controls. In controls, a longer sleep duration, earlier sleep onset and shorter sleep inertia were related to higher FA, which is in line A longer sleep duration, earlier onset of sleep and shorter sleep inertia are with previous studies reporting on the effects of sleep on cell proliferation, all indicative of a more healthy sleep pattern (Merikanto et al., 2012), so the synaptic plasticity and neurogenesis (Bellesi, et al., 2013; Kopp, et al., 2006; finding that this relates to better integrity of white matter in controls confirms Kreutzmann, et al., 2015; Mirescu, et al., 2006). However, several measures of the hypothesis that sleep and white matter integrity are positively associated. sleep showed opposing associations with integrity of white matter in patients. Previous studies that measured the link between sleep and human brain

These differences between patients and controls are most likely the result of function concluded that sleep is associated with synaptic homeostasis, clearance Chapter 4 confounding factors, such as use of antipsychotic drugs in the patient group. of amyloid beta, grey matter volume and cortical thickness (Altena et al., 2010; Diekelmann & Born, 2010; Joo et al., 2010, 2011, 2009; Kang et al., 2009; Our results are in keeping with evidence suggesting that physical activity is Ooms et al., 2014; Tononi & Cirelli, 2006; Wang et al., 2011; Winkelman et al., positively related to measures of white matter integrity in healthy subjects 2013; Xie et al., 2013). Our study is one of the first to expand these findings to (Sexton et al., 2015; Voelcker-Rehage & Niemann, 2013). This may be partly better integrity of white matter microstructure. A recent study in middle-aged explained by an up-regulation of the expression of brain-derived neurotrophic adults found that short sleep duration (i.e. ≤6h) correlated with higher mean factor (BDNF) by physical activity. This neurotrophin has consistently diffusivity and lower fractional anisotropy although the latter was no longer been related to increases in grey matter volume and evidence also points to a significant after adjustments for confounding variables (Yaffe et al., 2016). As a relation with myelinogenesis and white matter microstructure (Du, Fischer, clinical study previously demonstrated, sleep positively affects both the number Lee, Lercher, & Dreyfus, 2003; Tost et al., 2013). Alternatively, higher levels and proliferation rate of oligodendrocytes, while spontaneous and forced wake of cardiorespiratory fitness lead to improvements in the cerebral blood flow, states reduced oligodendrocyte proliferation by half (Bellesi et al., 2013). Longer resulting in increased brain volume and white matter integrity (Chen, Rosas, sleep duration, earlier sleep onset and shorter sleep inertia could therefore & Salat, 2013; Zhu et al., 2015). Here, we extend these findings, and show increase oligodendrocyte proliferation and myelination, ultimately leading to

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls that physical activity and white matter integrity are similarly related in bipolar increased integrity of white matter. On the contrary, a longer sleep duration,

110 111 exceeding 9 h per night, has been associated with white matter hyperintensities were now left undetected. However, the time between MRI and actigraphy and decreased cognitive functioning, suggesting that sleep duration shows measurements was similar for patients and controls, and there is no reason to a U-shaped relation with brain functioning (Lo, Groeger, Cheng, Dijk, & assume that such variation was systematic and has biased the results. Also, we Chee, 2015; Ramos et al., 2014). However, in our data bipolar patients show need to point out that we did not exclude participants diagnosed with conditions a linear, but reversed association between sleep and integrity of white matter other than sleep apnoea, head trauma or neurological disorders. compared to controls. We show that a likely explanation for this counterintuitive association is the use of psychotropic medication. Several widely prescribed In conclusion, the current study found that higher levels of physical activity types of medication affect sleep-wake patterns; lithium normalizes the 24-h were related to better integrity of white matter, regardless of bipolar diagnosis. circadian cycle, whereas benzodiazepines and antipsychotic medication can In controls, less disturbed sleep was associated with better integrity of white promote sleep onset and/or maintenance (Klemfuss, 1992; Monti and Monti, matter, but opposing associations were found in bipolar patients, most likely due 2004). We found that in bipolar patients who received antipsychotic medication to extraneous variables such as use of antipsychotic medication. longer sleep duration was associated with lower FA. A previous study already reported that chronic use of antipsychotic medication without mood stabilizing effect decreases myelin/oligodendrocyte related gene expression in white matter (Narayan, Kass, & Thomas, 2007). Moreover, a recent study suggests that

sleep may be influenced by antipsychotics via mTORC1- proteins synthesis Chapter 4 that in turn is involved in neuronal function, but further studies are required (Bowling et al., 2014). Our findings do warrant further study of the effect per medication type on sleep and white matter microstructure. Randomized clinical trials studying the effect of medication on white matter microstructure could address this by incorporating sleep measures, taking into account the relevant white matter tracts resulting from our analyses. We found no relationship with other types of medication, but given that the majority of our bipolar sample used more than one type of medication, we cannot reliably distinguish between the effects of medication types (Brambilla, Bellani, Yeh, & Soares, 2009). Also, confounding by indication may be at play here, whereby more severely affected patients may receive particular types of medication. A further limitation of the current study is the average time difference between the MRI acquisition and actigraphy measurement of 1.4 years. Changes in exercise behaviour and sleep

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls patterns, might have resulted in variation in white matter microstructure that

112 113 Caspersen, C. J., Powell, K. E., & Christenson, G. M. (1985). Physical activity, exercise, and References physical fitness : definitions and distinctions for health-related research. Public Health Reports, 100(2), 126–131.

Altena, E., Vrenken, H., Van Der Werf, Y. D., van den Heuvel, O. A., & Van Someren, E. J. W. Chen, J. J., Rosas, H. D., & Salat, D. H. (2013). The relationship between cortical blood flow (2010). Reduced orbitofrontal and parietal gray matter in chronic insomnia: A voxel-based and sub-cortical white-matter health across the adult age span. PLoS ONE, 8(2), e56733. morphometric study. Biological Psychiatry, 67(2), 182–185. Cirelli, C., Faraguna, U., & Tononi, G. (2006). Changes in brain gene expression after long- American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: term sleep deprivation. Journal of Neurochemistry, 98(5), 1632–1645. DSM-5. Washington DC. Diekelmann, S., & Born, J. (2010). The memory function of sleep. Nature Reviews Neuroscience, Andersson, J. L. R., Skare, S., & Ashburner, J. (2003). How to correct susceptibility distortions 11(2), 114–126. in spin-echo echo-planar images: Application to diffusion tensor imaging. NeuroImage, Du, Y., Fischer, T. Z., Lee, L. N., Lercher, L. D., & Dreyfus, C. F. (2003). Regionally specific 20(2), 870–888. effects of BDNF on oligodendrocytes. Developmental Neuroscience, 25(2–4), 116–126. Beaulieu, C. (2002). The basis of anisotropic water diffusion in the nervous system - a technical Elvsåshagen, T., Norbom, L. B., Pedersen, P. Ø., Quraishi, S. H., Bjørnerud, A., Malt, U. F., review. NMR in Biomedicine, 15, 435–455. Westlye, L. T. (2015). Widespread changes in white matter microstructure after a day of Beaulieu, C. (2009). The biological basis of diffusion anisotropy. In Diffusion MRI (pp. 105–126). waking and sleep deprivation. PLoS ONE, 10(5), 1–15. Elsevier Inc. Erickson, K. I., Leckie, R. L., & Weinstein, A. M. (2014). Physical activity, fitness, and gray Bellesi, M., Pfister-Genskow, M., Maret, S., Keles, S., Tononi, G., & Cirelli, C. (2013). Effects matter volume. Neurobiology of Aging, 35(SUPPL.2), S20–S28. of sleep and wake on oligodendrocytes and their precursors. The Journal of Neuroscience, First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1997). Structured Clinical 33(36), 14288–14300. Interview for DSM-IV Axis I Disorders (SCID-I). New York: Biometric Research Chapter 4 Belmaker, R. H. (2004). Bipolar disorder. The New England Journal of Medicine, 351, 476–486. Department.

Bherer, L., Erickson, K. I., & Liu-Ambrose, T. (2013). A review of the effects of physical activity Hallahan, B., Newell, J., Soares, J. C., Brambilla, P., Strakowski, S. M., Fleck, D. E., … and exercise on cognitive and brain functions in older adults. Journal of Aging Research, 2013. McDonald, C. (2011). Structural magnetic resonance imaging in bipolar disorder: An international collaborative mega-analysis of individual adult patient data. Biological Bowling, H., Zhang, G., Bhattacharya, A., Pérez-Cuesta, L. M., Deinhardt, K., Hoeffer, C. A., Psychiatry, 69(4), 326–335. Chao, M. V. (2014). Antipsychotics activate mTORC1-dependent translation to enhance neuronal morphological complexity. Science Signaling, 7(308), ra4. Harvey, A. G. (2008). Sleep and circadian rhythms in bipolar disorder: Seeking synchrony, harmony, and regulation. American Journal of Psychiatry, 165(7), 820–829. Brambilla, P., Bellani, M., Yeh, P. H., & Soares, J. C. (2009). Myelination in bipolar patients and the effects of mood stabilizers on brain anatomy. Current Pharmaceutical Design, 15(22), Heng, S., Song, A. W., & Sim, K. (2010). White matter abnormalities in bipolar disorder: 2632–2636. insights from diffusion tensor imaging studies.Journal of Neural Transmission, 117(5), 639– 654. Burzynska, A. Z., Chaddock-Heyman, L., Voss, M. W., Wong, C. N., Gothe, N. P., Olson, E. A., … Kramer, A. F. (2014). Physical activity and cardiorespiratory fitness are beneficial for Hibar, D. P., Westlye, L. T., van Erp, T. G. M., Rasmussen, J., Leonardo, C. D., Faskowitz, white matter in low-fit older adults. PLoS ONE, 9(9), e107413. J., … Andreassen, O. A. (2016). Subcortical volumetric abnormalities in bipolar disorder. Molecular Psychiatry, 21(12), 1–7. Caseras, X., Tansey, K. E., Foley, S., & Linden, D. (2015). Association between genetic risk scoring for schizophrenia and bipolar disorder with regional subcortical volumes. Jenkinson, M., Beckmann, C. F., Behrens, T. E. J., Woolrich, M. W., & Smith, S. M. (2012). Translational Psychiatry, 5(12), e692. FSL. NeuroImage, 62(2), 782–790. The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls

114 115 Joo, E. Y., Jeon, S., Lee, M., Kim, S. T., Yoon, U., Koo, D. L., … Hong, S. B. (2011). Analysis Narayan, S., Kass, K. E., & Thomas, E. A. (2007). Chronic haloperidol treatment results in of cortical thickness in narcolepsy patients with cataplexy. Sleep, 34(10), 1357–1364. a decrease in the expression of myelin/oligodendrocyte-related genes in the mouse brain. Journal of Neuroscience Research, 85(4), 757–765. Joo, E. Y., Tae, W. S., Kim, S. T., & Hong, S. B. (2009). Gray matter concentration abnormality in brains of narcolepsy patients. Korean Journal of Radiology, 10(6), 552–558. Nichols, T. E., & Holmes, A. P. (2002). Nonparametric permutation tests for functional neuroimaging: A primer with examples. Human Brain Mapping, 15(1), 1–25. Joo, E. Y., Tae, W. S., Lee, M. J., Kang, J. W., Park, H. S., Lee, J. Y., … Hong, S. B. (2010). Reduced brain gray matter concentration in patients with obstructive sleep apnea syndrome. Nortje, G., Stein, D. J., Radua, J., Mataix-Cols, D., & Horn, N. (2013). Systematic review and Sleep, 33(2), 235–241. voxel-based meta-analysis of diffusion tensor imaging studies in bipolar disorder.Journal of Affective Disorders, 150(2), 192–200. Kang, J.-E., Lim, M. M., Bateman, R. J., Lee, J. J., Smyth, L. P., Cirrito, J. R., … Holtzman, D. M. (2009). Amyloid-b dynamics are regulated by orexin and the sleep-wake cycle. Sleep, Ooms, S., Overeem, S., Besse, K., Rikkert, M. O., Verbeek, M., & Claassen, J. A. H. R. (2014). 326(5955), 1005–1007. Effect of 1 night of total sleep deprivation on cerebrospinal fluidβ -amyloid 42 in healthy middle-aged men: A randomized clinical trial. JAMA Neurology, 71(8), 971–977. Klemfuss, H. (1992). Rhythms and the pharmacology of lithium. Pharmacology & Therapeutics, 56, 53–78. Pagani, L., St. Clair, P. a., Teshiba, T. M., Service, S. K., Fears, S. C., Araya, C., … Freimer, N. B. (2016). Genetic contributions to circadian activity rhythm and sleep pattern phenotypes Kopp, C., Longordo, F., Nicholson, J. R., & Lüthi, A. (2006). Insufficient sleep reversibly in pedigrees segregating for severe bipolar disorder. Proceedings of the National Academy of alters bidirectional synaptic plasticity and NMDA receptor function. Journal of Neuroscience, Sciences, 113(6), E754–E761. 26(48), 12456–12465. R Development Core Team, R. (2014). R: A Language and Environment for Statistical Kreutzmann, J. C., Havekes, R., Abel, T., & Meerlo, P. (2015). Sleep deprivation and Computing. hippocampal vulnerability: Changes in neuronal plasticity, neurogenesis and cognitive function. Neuroscience, 309, 173–190. Ramos, A. R., Dong, C., Rundek, T., Elkind, M. S. V, Boden-Albala, B., Sacco, R. L., & Chapter 4 Wright, C. B. (2014). Sleep duration is associated with white matter hyperintensity volume Liu, D., Wang, Z., Shu, H., & Zhang, Z. (2016). Disrupted white matter integrity is associated in older adults: The Northern Manhattan Study.Journal of Sleep Research, 23(5), 524–530. with cognitive deficits in patients with amnestic mild cognitive impairment: An atlas-based study. SAGE Open Medicine, 4, 1–8. Sandoval, H., Soares, J. C., Mwangi, B., Asonye, S., Alvarado, L. A., Zavala, J., … Escamilla, M. A. (2014). Confirmation of MRI anatomical measurements as endophenotypic markers Lo, J. C., Groeger, J. A., Cheng, G. H., Dijk, D.-J., & Chee, M. W. L. (2015). Self-reported for bipolar disorder in a new sample from the NIMH Genetics of Bipolar disorder in Latino sleep duration and cognitive performance in older adults: a systematic review and meta- Populations study. Psychiatry Research - Neuroimaging, 247, 34–41. analysis. Sleep Medicine, 17, 87–98. Sexton, C. E., Betts, J. F., Demnitz, N., Dawes, H., Ebmeier, K. P., & Johansen-Berg, H. Mahon, K., Burdick, K. E., & Szeszko, P. R. (2010). A role for white matter abnormalities in (2015). A systematic review of MRI studies examining the relationship between physical the pathophysiology of bipolar disorder. Neuroscience and Biobehavioral Reviews, 34(4), 533– fitness and activity and the white matter of the ageing brain.NeuroImage, 131, 81–90. 554. Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., … Merikanto, I., Kronholm, E., Peltonen, M., Laatikainen, T., Lahti, T., & Partonen, T. (2012). Dunbar, G. C. (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.): The Relation of chronotype to sleep complaints in the general Finnish population. Chronobiology development and validation of a structured diagnostic psychiatric interview for DSM-IV International, 29(3), 311–317. and ICD-10. Journal of Clinical Psychiatry, 59, 22–33.

Mirescu, C., Peters, J. D., Noiman, L., & Gould, E. (2006). Sleep deprivation inhibits adult Smith, S. M. (2002). Fast robust automated brain extraction. Human Brain Mapping, 17(3), 143– neurogenesis in the hippocampus by elevating glucocorticoids. Proceedings of the National 155. Academy of Sciences, 103(50), 19170–19175.

The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls Monti, J. M., & Monti, D. (2004). Sleep in schizophrenia patients and the effects of antipsychotic drugs. Sleep Medicine Reviews, 8(2), 133–48.

116 117 Smith, S. M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T. E., Mackay, C. E., Xie, L., Kang, H., Xu, Q., Chen, M. J., Liao, Y., Thiyagarajan, M., … Nedergaard, M. (2013). … Behrens, T. E. J. (2006). Tract-based spatial statistics: voxelwise analysis of multi-subject Sleep drives metabolite clearance from the adult brain. Science, 342(6156), 373–7. diffusion data.NeuroImage, 31(4), 1487–14505. Yaffe, K., Nasrallah, I., Hoang, T. D., Lauderdale, D. S., Knutson, K. L., Carnethon, M. R., Smith, S. M., & Nichols, T. E. (2009). Threshold-free cluster enhancement: Addressing … Sidney, S. (2016). Sleep duration and white matter quality in middle-aged adults. Sleep, problems of smoothing, threshold dependence and localisation in cluster inference. 39(9), 1743–1747. NeuroImage, 44(1), 83–98. Yip, S. W., Chandler, R. A., Rogers, R. D., Mackay, C. E., & Goodwin, G. M. (2013). White Smith, S. M., Smith, S. M., Johansen-berg, H., Johansen-berg, H., Jenkinson, M., Jenkinson, matter alterations in antipsychotic- and mood stabilizer-naïve individuals with bipolar II/ M., … Behrens, T. E. J. (2007). Acquisition and voxelwise analysis of multi-subject diffusion NOS disorder. NeuroImage: Clinical, 3, 271–278. data with tract-based spatial statistics. Nature Protocols, 2(3), 499–503. Zhu, N., Jacobs, D. R. J., Schreiner, P. J., Launer, L. J., Whitmer, R. A., Sidney, S., … Bryan, Svatkova, A., Mandl, R. C. W., Scheewe, T. W., Cahn, W., Kahn, R. S., & Hulshoff Pol, H. E. R. N. (2015). Cardiorespiratory fitness and brain volume and white matter integrity: The (2015). Physical exercise keeps the brain connected: Biking increases white matter integrity CARDIA Study. Neurology, 84(23), 2347–2353. in patients with schizophrenia and healthy controls. Schizophrenia Bulletin, 41(4), 869–878.

Telzer, E. H., Goldenberg, D., Fuligni, A. J., Lieberman, M. D., & Gálvan, A. (2015). Sleep variability in adolescence is associated with altered brain development. Developmental

Cognitive Neuroscience, 14, 16–22.

Tononi, G., & Cirelli, C. (2006). Sleep function and synaptic homeostasis. Sleep Medicine Reviews, 10(1), 49–62.

Tost, H., Alam, T., Geramita, M., Rebsch, C., Kolachana, B., Dickinson, D., … Marenco, S. Chapter 4 (2013). Effects of the BDNF Val66Met polymorphism on white matter microstructure in healthy adults. Neuropsychopharmacology, 38(3), 525–32.

Van Someren, E. J. W., Hagebeuk, E. E. O., Lijzenga, C., Scheltens, P., de Rooij, S. E. J. A., Jonker, C., … Swaab, D. F. (1996). Circadian rest-activity rhythm disturbances in Alzheimer’s Disease. Biological Psychiatry, 40, 259–270.

Voelcker-Rehage, C., & Niemann, C. (2013). Structural and functional brain changes related to different types of physical activity across the life span.Neuroscience and Biobehavioral Reviews, 37(9), 2268–2295.

Wang, G., Grone, B., Colas, D., Appelbaum, L., & Mourrain, P. (2011). Synaptic plasticity in sleep: Learning, homeostasis and disease. Trends in Neurosciences, 34(9), 452–463.

Winkelman, J. W., Plante, D. T., Schoerning, L., Benson, K., Buxton, O. M., O’Connor, S. P., … Gonenc, A. (2013). Increased rostral anterior cingulate cortex volume in chronic primary insomnia. Sleep, 36(7), 991–998.

Wise, T., Radua, J., Nortje, G., Cleare, A. J., Young, A. H., & Arnone, D. (2016). Voxel-based meta-Analytical evidence of structural disconnectivity in major depression and bipolar disorder. Biological Psychiatry, 79(4), 293–302. The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls

118 119 Supplementary material

Suppl. Table 1 – A: Association between sleep measures and average FA of the significant clusters (with age and gender as covariates). B: Association between sleep onset and average FA of the significant clusters, while controlling for history of psychotic symptoms or number of episodes (with age and gender as covariates); Association between sleep duration and average FA of the significant clusters, while controlling for antipsychotic medication (with age and gender as covariates) C:Association between squared sleep duration and average FA of the significant clusters, and centred plus squared sleep duration and average FA of the significant clusters(with Supplementary figure 1 – TBSS was used to assess the association of interdaily stability with age and gender as covariates). integrity of white matter microstructure. We found positive significant correlations in the combined sample (patients and healthy controls) which are shown in yellow to red, depending on β R2 p-value N the significance level. The white matter skeleton is shown in green and overlaid on the MNI152 A Sleep duration -0.39 0.32 0.004 51 template brain. Images are shown in radiological convention (the left is displayed on the right) Sleep onset 0.49 0.37 <0.001 51 and corresponding Montreal Neurological Institute (MNI) coordinates are displayed below. Sleep inertia 0.34 0.33 0.008 51 P-values are TFCE and FWE corrected for multiple comparisons across space and corrected for B Sleep onset + psychotic symptoms 0.41 0.35 0.002 50 age, gender and (patient/control) group. Sleep onset + number of episodes 0.35 0.40 0.02 41 Sleep duration + antipsychotics -0.26 0.35 0.11 40

C Sleep duration squared -0.38 0.32 0.004 51 Chapter 4 Sleep duration centred + squared -0.39 0.32 0.004 51

Suppl. Table 2 – Correlations between illness-specific confounders, sleep measures and average FA of the significant clusters. P<0.05 indicated with one asterisk, p<0.001 indicated with double asterisk.

Sleep FA sleep Sleep FA sleep Sleep FA sleep duration duration onset onset inertia inertia Suppl. Figure 2 – The association of sleep duration with integrity of white matter microstructure

Lithium 0.10 0.34 * -0.10 0.28* 0.17 0.38** differs between patients using antipsychotics and controls. TBSS was used to assess the Other mood stabilizers 0.12 -0.33 * -0.04 -0.15 -0.14 -0.34* interaction effects of sleep duration and group (Group * Sleep) with integrity of white matter Benzodiazepines 0.07 -0.03 0.13 0.20 0.01 0.05 microstructure. There is a significant negative interaction between sleep duration*group and FA. Antipsychotics 0.47* -0.31* -0.40** -0.26 -0.23 -0.17 A negative interaction is defined as interactions where the slope between sleep variables and Nr of episodes -0.02 -0.35* -0.26* -0.39** -0.16 -0.30* DTI measures is lower for patients than for controls. Significant interactions are shown in yellow History of psychosis -0.19 0.29* 0.27* -0.35** 0.04 0.22 to red, depending on the significance level. The white matter skeleton is shown in green and Illness duration -0.12 -0.34** 0.01 -0.28* -0.12 -0.43** overlaid on the MNI152 template brain. Images are shown in radiological convention (the left is displayed on the right) and corresponding Montreal Neurological Institute (MNI) coordinates are displayed below. P-values are TFCE and FWE corrected for multiple comparisons across The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls space, and corrected for age and gender.

120 121 Suppl. Table 3 - Cluster size (number of voxels), p-value, t-value, MNI coordinates of centre of gravity and anatomical location of significant associations between interdaily stability and sleep duration (latter analysis restricted to patients using antipsychotics) with FA.

Cluster Corrected MNI coordinates Analysis size p-value T-value (x; y; z) White matter tracts

Significant 9935 0.018 3.12 6.22; -25; 28.6 Genu of corpus callosum correlation Body of corpus callosum of interdaily stability with Splenium of corpus callosum FA Right anterior corona radiata Left anterior corona radiata Right superior corona radiata Left superior corona radiata Right posterior corona radiata Left posterior corona radiata Right posterior thalamic radiation (include optic radiation) Left cingulum (cingulate gyrus) Right superior longitudinal fasciculus 989 0.043 3.07 -40.2; -5.46; 27.3 Left superior longitudinal fasciculus

480 0.044 3.83 38.1; -22.4; -4.51 Retrolenticular part of right internal capsule Chapter 4 Sagittal stratum Right external capsule Fornix (cres) / Stria terminalis 311 0.043 3.81 48.4; -32.8; -10.6 Sagittal stratum Significant 53 0.048 3.95 -16.9; 27.2; 21.7 Genu of corpus callosum Group*Sleep Left anterior corona radiata duration interaction with FA in patient using antipsychotics and controls The associationof sleep physicaland activity with integrityof white matter microstructure bipolarin disorder patients and healthy controls

122 123 Chapter 5

Sleep disturbances, psychosocial difficulties and health risk behaviour in 16,781 Dutch adolescents

Sanne Verkooijen, Nelleke de Vos, Betty Bakker, Susan J.T. Branje, René S. Kahn, Roel A. Ophoff, Carolien M. Plevier, Marco P.M. Boks Abstract Introduction

Objective To investigate the prevalence of adolescent sleep disturbances Alterations in sleep-wake pattern are among the many developmental changes and their relation with psychosocial difficulties and health risk behaviours we that occur during adolescence. Several studies report that adolescents have an analysed data of a province-wide health survey (n=16,781). increased need for sleep, with the average adolescent requiring around 9 hours of sleep per night (Carskadon, 1990; Warner, Murray, & Meyer, 2008). However, Methods Assessments included self-reported sleep disturbances, the it is also generally accepted that this age group is specifically vulnerable for Strength and Difficulties Questionnaire, suicidality, current use of tobacco, experiencing sleep disturbances (Dahl & Lewin, 2002). Data from several alcohol and drugs, physical inactivity, and compulsive use of multimedia. We industrialized countries revealed that one in four adolescents acquire less than 6 used multi-level analyses to investigate the relations including differences hours of sleep, and prevalence rates of insomnia range between 4-24% (Hysing, between boys and girls, as well as the mediating role of emotional problems. Pallesen, Stormark, Lundervold, & Sivertsen, 2013; Johnson, Roth, Schultz, & Breslau, 2006; Ohayon, Roberts, Zulley, Smirne, & Priest, 2000; Roberts, Results Just under 20% of adolescents reported sleep disturbances in Roberts, & Duong, 2009). the previous month. These sleep disturbances were associated with psychosocial problems (OR: 6.42, p<0.001), suicidality (OR: 3.90 - 4.14, p<0.001) and all Studies focusing on the consequences of impaired sleep in adolescents have health risk behaviour (OR: 1.62-2.66, p<0.001) but not with physical inactivity. found associations with a wide range of psychosocial problems and general We found moderation by gender for the relations between sleep and suicide health issues. Adolescents who report sleep disturbances have an increased attempts (OR: 0.38, p<0.002) and sleep and cannabis use (OR: 0.52, p=0.002), risk for depressed mood, anxiety disorders and suicide attempts (Fredriksen,

indicating attenuated relations in girls compared to boys. Emotional problems Rhodes, Reddy, & Way, 2004; Liu & Buysse, 2005; Roberts, Roberts, & Chapter 5 partially mediated the relations between sleep disturbances and multimedia use. Duong, 2008). Moreover, lack of sleep was associated with a wide range of health risk behaviours, such as increased substance use, electronic multimedia Conclusions The current study reiterates the high incidence of sleep use and physical inactivity (Cain & Gradisar, 2010; Johnson & Breslau, 2001; disturbances during adolescence. These sleep disturbances were strongly related McKnight-Eily et al., 2011). As these health risk behaviours frequently co- to psychosocial problems and a wide range of health risk behaviours. Although occur with emotional problems, the question remains whether the link between the direction of causality cannot be inferred, the current study emphasizes sleep disturbances and health risk behaviour is (partly) mediated by emotional the need for awareness of impaired sleep in adolescents. Moreover, the gender problems (Shochat, Cohen-Zion, & Tzischinsky, 2014). differences in associated suicide attempts and cannabis use call for further research into tailored intervention strategies. In addition to the role of emotional problems, another important modifier is of note. Whereas the psychosocial consequences of sleep disturbances have

Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents gained increasing attention, the presence of gender differences has gone largely

126 127 unnoticed. However emerging evidence suggests that girls have a higher 9,747 fourth year students from 55 schools were invited to participate in the incidence of sleep disturbances, a finding that most likely emerges after the survey. The response rate for second year students was 91% and for fourth year onset of menses (Johnson et al., 2006). Gender differences are also prevalent for students 83%. Reasons for not participating were refusal by the student or their psychosocial problems and health risk behaviours. Boys more frequently report parents or absence due to illness, doctor visits or internships. In 3 schools use of drugs, multimedia and tobacco, while girls have a higher incidence of data were only available of a small selection of students, because screening for suicidal behaviour (Kann et al., 2014; Mentzoni et al., 2011). Therefore, in order health difficulties was done face-to-face. Data from these schools were excluded to adequately assess the sleep disturbances in adolescents and their relation to from further analyses (n=22). The remaining 53 schools comprised 82.8% of all psychosocial problems and health risk behaviours, it is necessary to investigate schools in the greater Utrecht area, with an average of 418 included students gender differences to see whether not only mean levels but also associations per school (range 63-616). In 22 schools, face-to-face screening was preferred differ. over questionnaires with regard to pre-vocational level students. Data from these schools only included higher pre-vocational, higher secondary and pre- The current study investigates the prevalence of sleep disturbances in a large university students. This resulted in a relative underrepresentation of pre- sample of secondary school adolescents from the Netherlands. We examine vocational students of 10% in the total sample (see Table 1). the association between reported sleep disturbances, psychosocial difficulties and health risk behaviours, including substance use, physical inactivity and Measures compulsive multimedia use. We subsequently asses differences in these Sleep disturbances associations between boys and girls and investigate the potential mediating role We determined the presence of sleep disturbances with the question: “In the of emotional problems. past 4 weeks, how often did you experience trouble sleeping?” , which was answered

with “never”, “almost never”, “sometimes”, “often” or “very often”. This question was Chapter 5 Methods dichotomized into (almost) never/sometimes versus (very) often.

Sample Strengths and Difficulties Questionnaire The current study used data from a Youth Health Care questionnaire by the We assessed the presence of psychosocial problems with the self-report version Dutch Community Health Service of the greater Utrecht area. The general of the Strength and Difficulties Questionnaire (SDQ ; Goodman et al., 1998). aim of this questionnaire was to identify adolescents at risk for developing This questionnaire is designed as a brief behavioural screening measure for psychosocial and general health problems. Schools were provided with a digital children between ages 11 and 16. It encompasses 25 items, all answered on a health check questionnaire for second and fourth grade students. All students three-point scale ranging from “not true”, “somewhat true” to “certainly true”. The completed the questionnaire in school, under the supervision of a teacher, but items are divided into 5 subscales: emotional problems, behavioural problems, participation was not mandatory. We used anonymised data that were acquired hyperactivity-inattention, peer problems and pro-social behaviour, of which

Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents in the academic year of 2014-2015. A total of 9,535 second year students and the first four subscales are summed to generate a total difficulties score. To aid

128 129 clinical interpretation of the results, we used validated categorization of the inquired how many days respondents were physically active for at least an hour questionnaire into ‘normal’, ‘borderline’ and ‘clinical’ (Goodman et al., 1998), in the previous week. This resulted in two activity variables. First, following and focussed on clinical outcomes relative to normal and borderline. Validation the Dutch guidelines for moderate physical activity (Kemper, Ooijendijk, & of the Dutch version of the SDQ indicated that the total difficulties score has Stiggelbout, 2000), students were rated as inactive when they were not physically sufficient internal consistencyα ( =0.70-0.78). The subscale emotional difficulties active for at least one hour on a daily basis. For the second activity variable we which was used as mediator, was found to have similar sufficient internal rated students as inactive when they were not physically active for an hour on consistency (α=0.63-0.71) (Muris, Meesters, & van den Berg, 2003; Van any of the previous seven days. Widenfelt, Goedhart, Treffers, & Goodman, 2003). Statistical analyses Suicidality All analyses were performed using R statistical software (R Core Team, Suicidal ideation was assessed with the following question: “In the previous year, 2015) and SPSS 23.0. We had 14 missing values for the sleep disturbances have you seriously considered ending your own life?” Students rated this question questionnaire, 82 for the SDQ , 55 for suicidality, and between 8-115 for the on a five-point scale (“never”, “once or twice”, “sometimes”, “often”, “very often”). health risk behaviours. All missing data were deleted listwise. Differences in We dichotomized suicidal ideation into present versus absent. Suicide attempts characteristics between adolescents with and without reported sleep disturbances were assessed with the question: “In the previous year, have you tried to commit were analysed using Mann-Whitney U rank sum test. The association between suicide?”, on which either “yes” or “no” was answered. sleep disturbances and psychosocial difficulties were analysed using multilevel logistic regression analyses with total SDQ score or suicidality as dependent Health risk behaviours variables, sleep disturbances as independent variable and school and grade as

A short version of the Compulsive Internet Use Scale (CIUS) (Meerkerk, Van random factors. Similarly, the associations between sleep disturbances and Chapter 5 Den Eijnden, Vermulst, & Garretsen, 2009) was used to inquire compulsive use health risk behaviours were analysed with separate multilevel logistic regression of social media and games. All six questions were answered on a 5-point scale analyses for all types of health risk behaviour as dependent variables and sleep ranging from “never” to “very often” and provided information regarding risky disturbances as independent variable. behaviour, including the inability to spent less time using multimedia, whether multimedia use was preferred over face-to-face contact, whether it affected their For all health risk behaviour variables that were significantly associated with mood and if homework was neglected due to multimedia use. Average scores of sleep disturbances, we investigated whether emotional problems mediated this 3 or higher are indicative of compulsive use of social media and games. relationship. As both the mediator (emotional problems) and the outcome (health risk behaviours) were dichotomous variables we used a method based on the Additionally, the following substances were examined: current use of tobacco, analyses of Mackinnon and Dwyer (Mackinnon & Dwyer, 1993). To account daily consumption of energy drinks, use of alcohol and cannabis in the past 4 for the fact that predictor and outcome variables have different scales in logistic

Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents weeks and life-time use of drugs other than cannabis (hard drugs). Lastly, we regression, we standardized the coefficients (for details see supplementary

130 131 materials). To test whether the indirect effect of emotional problems on health Association between sleep disturbances and psychosocial problems risk behaviour was significantly different from zero we calculated Sobel Z. In case See Table 2 for a summary of the multilevel logistic regression analyses of the of significant mediation, we reported the corresponding proportion mediated. associations between sleep disturbances and psychosocial difficulties. We found that the presence of sleep disturbances was associated with a borderline-clinical Lastly, the interaction between sleep disturbances and gender was tested score on the SDQ , as well as previous suicide attempts and suicidal ideation. by adding an interaction-term for sleep disturbance X gender to the previous psychosocial and health risk regression analyses. In case of significant Association between sleep disturbances and health risk behaviour interaction-terms, stratified analyses by gender were performed. Table 3 shows the results of the multilevel logistic regression analyses testing the association between sleep disturbances and health risk behaviours. Except All regression analyses were adjusted for age, gender and educational level, with for physical inactivity, all measures of health risk behaviour were related to a Bonferroni-corrected p-value set at 0.0045 to adjust for the separate analyses sleep disturbances. Analysing no days of physical activity versus at least one day of psychosocial difficulties and health risk behaviours. of physical activity we found that sleep disturbances were related to physical inactivity with an odds ratio of 1.45 (95% CI: 1.02-2.05) and p=0.036. However, Results this p-value did not exceed the Bonferroni-corrected p-value of 0.0045. Emotional problems were a significant mediator in all health risk behaviour, Sample characteristics although it only survived corrections for multiple testing in the relation between Table 1 shows the characteristics of the complete sample which consisted of sleep disturbances and compulsive use of social media and gaming (Table 4). 8479 boys (50.5%) and 8302 girls (49.5%). Age ranged between 12 and 18 years, Given that the direct effect (c’) remained significant in both the relation between

with a mean age of 14.5 (sd = 1.2). About half of the students (n=8666 , 51.6%) sleep disturbances and compulsive use social media (OR: 0.14, p<0.001) and Chapter 5 attended the second grade of secondary education and 8112 students (48.3%) compulsive gaming (OR: 0.16, p<0.001), emotional problems are considered as a attended the fourth grade. In total, 3337 adolescents (19.9%) reported sleep partial mediator. disturbances. Age, grade and ethnicity did not differ between adolescents with and without sleep disturbances (U=22204789.0, p=0.40; U=22252384.0, p=0.49; Gender differences U=21837543.5, p=0.31) but significantly more girls reported sleep disturbances The interaction analyses (Figure 1A) showed that the association between sleep compared to boys (U=19120672.0, p<0.001). Also, level of education differed disturbances and suicide attempts is more pronounced for girls (OR: 5.71 [3.78- between the two groups, with more sleep disturbed adolescents attending lower 8.63], p<0.001) than for boys (OR: 2.32 [1.47-3.65], p<0.001). Additionally, we educational levels (U=20615952.5, p<0.001). found that girls with sleep disturbances more often report cannabis use (OR: 3.66 [2.67-5.02], p<0.001) compared to boys (OR: 1.93 [1.50-2.47], p<0.001) (Figure 1B). Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents

132 133 Table 2 – Association between sleep disturbances, psychosocial problems and suicidality Table 1 – Sample characteristics Sleep disturbances No sleep disturbances (N=3,337) (N=13,430) n=16,781 N / % N / % OR 95% CI p-value Age M / SD 14.5 (1.2) Psychosocial difficulties 848 / 25.4 689 / 5.1 6.42 5.32-7.76 <0.001* Male gender N / % 8479 (50.5) Suicide attempts 88 / 2.6 56 / 0.4 4.14 2.97-5.77 <0.001* Level of education N / % pre-vocational 1686 (10.0) Suicidal ideation 817 / 24.5 949 / 7.1 3.90 3.51-4.33 <0.001* higher pre-vocational 4455 (26.5) Odds ratio (OR) and 95% confidence interval (CI) age, gender and level of education as covariates and higher secondary 5994 (35.7) school and grade as level 2 variables (no sleep disturbances as reference category). pre-university 4642 (27.7) * Significant at Bonferroni-corrected p-value of 0.0045 Year of education (2nd) 8666 (51.6) Ethnicity (Dutch) N / % 15306 (91.2) Sleep disturbances N / % 3337 (19.9%) Table 3 - Association between sleep disturbances and health risk behaviour Emotional problems N / % 1290 (7.7) Conduct problems N / % 1260 (7.5) Sleep disturbances No sleep disturbances (N=3,337) (N=13,430) Hyperactivity N / % 4189 (25.0) N / % N / % OR 95% CI p-value Antisocial problems N / % 1415 (8.4) Peer problems N / % 1168 (7.0) Current smokers 396 / 11.9 784 / 5.8 2.13 1.85-2.44 <0.001* General difficulties N / % 1539 (9.2) Energy drinks (daily) 66 / 2.0 124 / 0.9 1.95 1.48-2.56 <0.001* Suicide attempt N / % 144 (0.9) Alcohol (< 4 weeks) 780 / 23.4 2255 / 16.8 1.62 1.45-1.80 <0.001* Suicidal ideation N / % 1766 (10.5) Cannabis (<4 weeks) 166 / 5.0 272 / 2.0 2.42 1.96-2.98 <0.001* Current smokers N / % 1180 (7.0) Hard drugs (ever) 66 / 2.0 76 / 0.6 2.66 1.92-3.69 <0.001* Energy drinks (daily) N/% 190 (1.1) Physical inactivity 2189 / 65.6 8802 / 65.5 1.02 0.94-1.11 0.578 Alcohol (< 4 weeks) N / % 3035 (18.1) Chapter 5 Compulsive social media use 1281 / 38.4 2726 / 20.3 2.20 2.02-2.40 <0.001* Cannabis (<4 weeks) N / % 438 (2.6) Compulsive gaming 628 / 18.8 1908 / 14.2 2.16 1.92-2.42 <0.001* Hard drugs (ever) N / % 142 (0.8) Physical inactivity (Dutch guidelines) 10992 (65.5) Odds ratio (OR) and 95% confidence interval (CI), age, gender and level of education as covariates and school and grade as level 2 variables (no sleep disturbances as reference category). Physical inactivity (<1 hr. per week) 123 (0.7) * Significant at Bonferroni-corrected p-value of 0.0045 Compulsive social media use 4008 (23.9) Compulsive gaming 2636 (15.1) Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents

134 135 Table 4 – Mediation effect of emotional problems on health risk behaviour stronger relation between sleep disturbances and suicide attempts compared to Mediated effect Sobel Z p-value Proportion Mediated Current smokers 0.01 2.35 0.02 - boys. Furthermore, sleep disturbances were related to nearly all inquired health Energy drinks (daily) 0.03 2.10 0.04 - risk behaviours, including the use of substances (i.e. alcohol, tobacco, cannabis Alcohol (< 4 weeks) 0.00 -0.39 0.69 - and other illicit drugs) and compulsive multimedia use. We found that girls who Cannabis (<4 weeks) 0.04 2.16 0.03 - reported sleep disturbances more frequently used cannabis compared to boys. Hard drugs (ever) 0.02 3.06 0.002 - Physical inactivity Ɨ - - - - Compulsive social media use 0.03 10.80 <0.001* 0.24 The incidence of sleep disturbances corroborates previous studies that estimated Compulsive gaming 0.04 4.68 <0.001* 0.16 the prevalence of adolescent sleep disturbances between 7-36% (for a review see * Significant at Bonferroni-corrected p-value of 0.0045 Gradisar et al., 2011). This age dependent vulnerability for sleep disturbances Ɨ Mediation analysis was not performed for physical inactivity since it was unrelated to sleep disturbances is thought to originate from a multitude of interacting factors. For one, the preferred bed times delay markedly during the early teen years (Crowley, Acebo, & Carskadon, 2007). However, since the majority of school times follow strict morning schedules, total sleep time reduces significantly during week days. To compensate for the acquired ‘sleep debt’, adolescents often increase the total sleep time and delay their bed times during weekends, thereby shifting their sleep pattern which increases subsequent difficulties initiating and maintaining sleep during the following week days (Dahl & Lewin, 2002). These delayed sleep patterns have been linked to circadian phase delays, such as self-report

phase preference and delayed Dim Light Melatonin Onset (Carskadon, Chapter 5 2011). Additionally, developmental and psychosocial changes that accompany adolescence (i.e. increased late-night social activities, screen time and preferred Figure 1 - a: The relation between sleep disturbances and suicide attempts, separately displayed autonomy regarding bed times) interact and further aggravate the disruptions of for boys and girls; b: Relation between sleep disturbances and cannabis use, separately displayed the sleep-wake patterns (Carskadon, 2011). for boys and girls

We showed that the presence of sleep disturbances is strongly associated with Discussion psychosocial difficulties and suicidality. These results are consistent with various studies relating adolescent sleep disturbances to mood and anxiety disorders, In this large sample of Dutch secondary school adolescents (n=16,781) we found behavioural difficulties and suicidal behaviour (Fredriksen et al., 2004; Liu & a prevalence of sleep disturbances just under 20%. These sleep disturbances were Buysse, 2005; Roberts et al., 2008). We also showed that sleep disturbances

Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents associated with psychosocial difficulties and suicidality, with girls showing a are related to several health risk behaviours. A possible explanation lies in

136 137 the impairments in inhibition and emotional regulation that accompany sleep Our cross-sectional design limits inferences of causality, so it is also possible disturbances and fatigue resulting in increased engagement in health risk that the health risk behaviours may have (partially) caused the reported sleep behaviours (Wong, Brower, Nigg, & Zucker, 2010). disturbances. The excessive use of electronic media and the daily consumption of energy drinks are in particular factors that predispose to impaired sleep We found that emotional problems partly mediate the relation between sleep (Calamaro, Mason, & Ratcliffe, 2009). Another limitation of the current study disturbances and compulsive use of multimedia and gaming. However, sleep concerns the fact that we inquired sleep disturbances with one question (i.e. disturbances were still associated with a two-fold increased risk for these health whether respondents had trouble sleeping). Although this is the most prominent risk behaviours after adjustment for emotional problems. Also, emotional symptom of a diagnosis of insomnia, we were not able to separate specific sleep problems did not mediate the relation between sleep disturbances and substance impairments (e.g. difficulties initiating versus maintaining sleep). Also, it is use. These results are consistent with suggestions that sleep disturbances in possible that the self-report nature of our study has led to biased responses. itself may lead to enhanced engagement in health risk behaviours (Catrett & However, a comparison between survey, sleep diary and actigraphy measures Gaultney, 2009). Although the underlying mechanism for this relation remains showed that in adolescents, these methods corresponded with medium to large elusive, a study by Gujar et al. (2011) found that sleep deprived individuals correlations, supporting the validity of self-reported sleep estimates (Wolfson et show an amplified reactivity in the mesolimbic network after being exposed al., 2003). Lastly, several schools did not include lower educated students and to pleasure-invoking stimuli, perhaps making them more susceptible to the since no data was available from adolescents not participating in the survey, we rewarding experiences following health risk behaviour. were unable to examine whether non-response was in any way selective. At the same time, the large sample and 82.8% coverage of the greater Utrecht area also Another finding of note is that we did not establish a link between inactivity and does not suggest selection bias.

sleep disturbances. This contradicts the meta-analysis by Bartel et al. (2015) that Chapter 5 concluded otherwise. A possible reason for this discrepancy is that, according In conclusion, our study found evidence for strong associations between sleep to sleep hygiene guidelines, engaging in vigorous sports a few hours prior bed disturbances and psychosocial problems. Moreover, we found that these sleep time can delay sleep onset (Stepanski & Wyatt, 2003). Since we did not inquire disturbances were related to an increased risk for health risk behaviours that the timing of activities, it is possible that this may have affected our results. could only partially be explained by the presence of emotional problems. Both However, several studies failed to find a link between activity close to bedtime the increased prevalence of psychosocial problems and health risk behaviour and disrupted sleep (Chennaoui, Arnal, Sauvet, & Léger, 2015). Moreover, a warrant special attention from parents and school health care staff when sleep previous study in a large sample of Estonian and Swedish adolescents also found disturbances are reported by adolescents. Our study is one of the first to identify no association between sleep duration and activity levels when age and Tanner gender differences in the association between sleep disturbances and mental stages were accounted for (Ortega et al., 2011). health measures and more research is warranted to determine tailored strategies in the prevention and treatment of sleep disturbances in boys and girls. Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents

138 139 Gujar, N., Yoo, S.-S., Hu, P., & Walker, M. (2011). Sleep deprivation amplifies reactivity of References brain reward networks, biasing the appraisal of positive emotional experiences. The Journal of Neuroscience, 31(12), 4466–74.

Bartel, K. A., Gradisar, M., & Williamson, P. (2015). Protective and risk factors for adolescent Hysing, M., Pallesen, S., Stormark, K. M., Lundervold, A. J., & Sivertsen, B. (2013). Sleep sleep: A meta-analytic review. Sleep Medicine Reviews, 21, 72–85. patterns and insomnia among adolescents: A population-based study. Journal of Sleep Research, 22(5), 549–556. Cain, N., & Gradisar, M. (2010). Electronic media use and sleep in school-aged children and adolescents: A review. Sleep Medicine, 11(8), 735–742. Johnson, E. O., & Breslau, N. (2001). Sleep problems and substance use in adolescence. Drug and Alcohol Dependence, 64, 1–7. Calamaro, C. J., Mason, T. B. A., & Ratcliffe, S. J. (2009). Adolescents living the 24/7 lifestyle: Effects of caffeine and technology on sleep duration and daytime functioning. Pediatrics, Johnson, E. O., Roth, T., Schultz, L., & Breslau, N. (2006). Epidemiology of DSM-IV 123(6), e1005–e1010. insomnia in adolescence: lifetime prevalence, chronicity, and an emergent gender difference. Pediatrics, 117(2), e247-56. Carskadon, M. A. (1990). Patterns of sleep and sleepiness in adolescents. Pediatrician, 17(1), 5–12. Kann, L., Kinchen, S., Shanklin, S. L., Flint, K. H., Hawkins, J., Harris, W. A., … Zaza, S. (2014). Youth risk behavior surveillance — United States, 2013. Surveillance Summaries, Carskadon, M. A. (2011). Sleep in adolescents: The perfect storm.Pediatric Clinics of North 63(4), 1–47. America, 58(3), 637–647. Kemper, H., Ooijendijk, W., & Stiggelbout, M. (2000). Consensus over de Nederlandse norm Catrett, C. D., & Gaultney, J. F. (2009). Possible insomnia predicts some risky behaviors among voor gezond bewegen. Tijdschrift Voor Gezondheidswetenschappen, 87(3), 180–183. adolescents when controlling for depressive symptoms. The Journal of Genetic Psychology, 170(4), 287–309. Liu, X., & Buysse, D. J. (2005). Sleep and youth suicidal behavior: a neglected field.Current Opinion in Psychiatry, 19, 288–293. Chennaoui, M., Arnal, P. J., Sauvet, F., & Léger, D. (2015). Sleep and exercise: A reciprocal issue? Sleep Medicine Reviews, 20, 59–72. Mackinnon, D. P., & Dwyer, J. H. (1993). Estimating mediated effects in prevention studies. Evaluation Review, 17(2), 144–158. Crowley, S. J., Acebo, C., & Carskadon, M. A. (2007). Sleep, circadian rhythms, and delayed

phase in adolescence. Sleep Medicine, 8(6), 602–612. McKnight-Eily, L. R., Eaton, D. K., Lowry, R., Croft, J. B., Presley-Cantrell, L., & Perry, G. Chapter 5 S. (2011). Relationships between hours of sleep and health-risk behaviors in US adolescent Dahl, R. E., & Lewin, D. S. (2002). Pathways to adolescent health: Sleep regulation and students. Preventive Medicine, 53(4–5), 271–273. behavior. Journal of Adolescent Health, 31(6 SUPPL.), 175–184. Meerkerk, G.-J., Van Den Eijnden, R. J. J. M., Vermulst, A. A., & Garretsen, H. F. L. (2009). Fredriksen, K., Rhodes, J., Reddy, R., & Way, N. (2004). Sleepless in Chicago: Tracking the The Compulsive Internet Use Scale (CIUS): some psychometric properties.Cyberpsychology effects of adolescent sleep loss during the middle school years.Child Development, 75(1), & Behavior, 12(1), 1–6. 84–95. Mentzoni, R. A., Brunborg, G. S., Molde, H., Myrseth, H., Skouverøe, K. J. M., Hetland, J., & Goodman, R., Meltzer, H., & Bailey, V. (1998). The strengths and difficulties questionnaire: A Pallesen, S. (2011). Problematic video game use: estimated prevalence and associations with pilot study on the validity of the self-report version. European Child and Adolescent Psychiatry, mental and physical health. Cyberpsychology, Behavior and Social Networking, 14(10), 591–596. 7(3), 125–130. Muris, P., Meesters, C., & van den Berg, F. (2003). The Strengths and Difficulties Questionnaire Gradisar, M., Gardner, G., & Dohnt, H. (2011). Recent worldwide sleep patterns and problems (SDQ ): Further evidence for its reliability and validity in a community sample of Dutch during adolescence: A review and meta-analysis of age, region, and sleep. Sleep Medicine, children and adolescents. European Child & Adolescent Psychiatry, 12, 1–8. 12(2), 110–118. Ohayon, M. M., Roberts, R. E., Zulley, J., Smirne, S., & Priest, R. G. (2000). Prevalence and

Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents patterns of problematic sleep among older adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 39(12), 1549–56.

140 141 Ortega, F. B., Ruiz, J. R., Labayen, I., Kwak, L., Harro, J., Oja, L., … Sjöström, M. (2011). Sleep duration and activity levels in Estonian and Swedish children and adolescents. Supplementary material European Journal of Applied Physiology, 111(10), 2615–2623.

R Core Team. (2015). R: A Language and Environment for Statistical Computing. The mediation equations

Roberts, R. E., Roberts, C. R., & Duong, H. T. (2008). Chronic insomnia and its negative Y’ = cX + E1

consequences for health and functioning of adolescents: A 12-month prospective study. M’ = aX + E2 Journal of Adolescent Health, 42(3), 294–302. Y’’ = bM + c’X + E3 Roberts, R. E., Roberts, C. R., & Duong, H. T. (2009). Sleepless in adolescence: Prospective data on sleep deprivation, health and functioning. Journal of Adolescence, 32(5), 1045–1057. Standardization of coefficients Shochat, T., Cohen-Zion, M., & Tzischinsky, O. (2014). Functional consequences of inadequate Standardized a = a * SD (X) / SD (M’) sleep in adolescents: A systematic review. Sleep Medicine Reviews, 18(1), 75–87. Standardized b = b * SD (M) / SD (Y’’) Stepanski, E. J., & Wyatt, J. K. (2003). Use of sleep hygiene in the treatment of insomnia. Sleep Standardized c = c * SD (X) / SD (Y’) Medicine Reviews, 7(3), 215–225. Standardized c’ = c’ * SD (X) / SD (Y’’) Van Widenfelt, B. M., Goedhart, A. W., Treffers, P. D. A., & Goodman, R. (2003).Dutch version of the Strengths and Difficulties Questionnaire (SDQ European). Child and Adolescent Psychiatry, 12(6), 281–289. Equations used to calculate SD for Y’, M’, and Y’’ 2 2 Wolfson, A. R., Carskadon, M. A., Acebo, C., Seifer, R., Fallone, G., Labyak, S. E., & Martin, Var(Y’) = c * Var(X) + p /3 J. L. (2003). Evidence for the validity of a sleep habits survey for adolescents. Sleep, 26(2), Var(M’) = a2 * Var(X) + p2/3 213–216. Var(Y”) = c’2 * Var(X) + b2 * Var(M) + 2*b*c’*Cov(X,M) + p2/3 Wong, M. M., Brower, K. J., Nigg, J. T., & Zucker, R. A. (2010). Childhood sleep problems,

response inhibition, and alcohol and drug outcomes in adolescence and young adulthood. Equations used to calculate SE for standardized a, b, c and c’ Chapter 5 Alcoholism: Clinical and Experimental Research, 34(6), 1033–1044. SE(stand a) = SE(a) * SD(X)/SD(M’) SE(stand b) = SE(b) * SD(M)/SD(Y”) SE(stand c) = SE(c) * SD(X)/SD(Y’) SE(stand c’) = SE(c’) * SD(X)/SD(Y”) Sleep disturbances, psychosocial difficulties and health risk behaviour 16,781 in Dutch adolescents

142 143 Chapter 6

Summary and general discussion In this thesis we addressed the association between sleep disturbances and of the disorder (i.e. age at onset, psychotic symptoms, number of episodes and psychopathology in bipolar disorder patients and adolescents from the general psychotic symptoms) were related to the sleep pattern of bipolar patients. population. This chapter will provide an overview of this thesis’ main findings, describe how these findings relate to the research questions from the first Chapter 4 focused on the association between integrity of white matter and sleep chapter, discuss our methodological considerations and provide implications for and activity patterns of bipolar patients and controls. We showed that bipolar future research. disorder is characterized by wide-spread reductions in fractional anisotropy (FA), including the corpus collosum, cerebral peduncle, corona radiata, internal Summary and external capsule, thalamic radiation and superior longitudinal fasciculus. In chapter 2 we gave an overview of the characteristics of bipolar phenotype The daytime and evening activity levels of both bipolar patients and controls and described why studying phenotypes such as circadian rhythms, sleep were related to higher FA. When analysing the association between FA and disturbances and the neuroanatomy of bipolar patients is of relevance. Such a sleep pattern we found opposing associations in bipolar patients compared to deep phenotype characterization of the disorder in combination with genome- controls. In controls we found the expected positive relation between healthy wide genetic data provides an opportunity to study the interplay between overt sleep patterns and FA. However, in patients we found that a longer sleep clinical characteristics and possible underlying biological mechanisms and could duration, earlier sleep onset and shorter sleep inertia all related to lower FA. ultimately lead to novel intervention strategies and treatment opportunities. We showed that these unexpected results were most likely due to confounding factors, such as use of antipsychotic medication. In chapter 3 we focussed on the sleep-wake phenotype of bipolar patients, their non-affected siblings and controls. We found that, compared to their non- In chapter 5 we investigated adolescent sleep disturbances in a large sample affected siblings, bipolar patients slept longer, most notably in the morning. of Dutch secondary school students. We found that an estimated twenty per However, these findings were mainly attributable to differences in mood cent of adolescents reported problems sleeping. These sleep disturbances were symptomatology. When we selectively compared euthymic patients to their strongly related to psychosocial problems, suicide ideation and suicide attempts. siblings, we did not find any differences in sleep characteristics. Moreover, Moreover, we found associations with a wide range of health risk behaviours, Chapter 6 we found an association between sleep pattern and depressive symptoms. including current use of alcohol, tobacco and drugs and compulsive multimedia Although our sample of patients and controls did not differ on any of the sleep use. We found that the relation between sleep disturbances and suicide attempts characteristics, our meta-analysis revealed that bipolar patients slept longer, had and cannabis use was more attenuated in girls. Also, emotional problems a longer sleep onset latency, woke up more often and had a lower sleep efficiency partly mediated the relation between sleep disturbances and compulsive use of compared to controls. These differences were, however, small and the total sleep multimedia and gaming. However, we found that sleep disturbances were still time never exceeded 9 hours, which still falls within the recommended amount associated with a two-fold increased risk for these health risk behaviours after of sleep for adults. This suggested that these difference in sleep pattern may not adjustment for emotional problems and emotional problems did not mediate the

Summary and general discussion be of clinical relevance. Lastly, we found that none of the clinical characteristics relation between sleep disturbances and substance use.

146 147 Conclusions episodes. Our findings do not point to stable deregulations in circadian biology, One of the most striking findings in this thesis is the fact that, despite although future research, as suggested in chapter 2 (i.e. circadian oscillators in several studies concluding otherwise, we did not find convincing support for the fibroblasts of bipolar patients) could further address this question. Also, the hypothesis that bipolar disorder patients are characterized by trait-like more large-scale longitudinal actigraphy studies could focus on the temporal dysfunctions in their sleep-wake pattern. Although two meta-analyses concluded association between sleep disturbances and emotional deregulation to explore that euthymic bipolar patients do show variations in sleep-wake pattern which of these two factors is the first to derail. compared to controls, as of yet, most of these variations are small and show inconsistencies over different measurement instruments (Geoffroy et al., 2015; The association between sleep disturbances and emotional deregulation is Ng et al., 2015). Being the largest actigraphy study to date, we question whether furthermore supported in our non-clinical sample of adolescents. Here we found the previously found case-control differences point to abnormalities that are of strong associations between adolescent sleep disturbances and psychosocial clinical relevance. If the sleep-wake pattern of bipolar patients indeed appears problems and suicidality. Moreover, we found associations between sleep normal, this may suggest that bipolar disorder is not characterized by disrupted disturbances and health risk behaviour that were partly mediated by emotional circadian rhythms, as previously concluded from studies finding irregularities problems, suggesting that sleep disturbances and the related emotional in core body temperature, dim light melatonin onset and cortisol secretion deregulation not only results in psychiatric mood symptomatology, but could (Cervantes, Gelber, Ng Ying Kin, Nair, & Schwartz, 2001; Nurnberger et al., also lead to behavioural problems, such as substance use and compulsive use of 2000; Wehr, 1982). In concordance, despite a few studies reporting correlations multimedia. Interestingly, these behavioural problems were not fully accounted between circadian gene variants and clinical characteristics of bipolar disorder, for by emotional problems. This suggests that not only the relation between none of the recent genome-wide association studies in bipolar disorder have emotional deregulation and sleep disturbances plays a role in the occurrence identified circadian genes (Dallaspezia & Benedetti, 2009; Hou et al., 2016; of behavioural problems, but that sleep deprivation in itself may lead to risky Ikeda et al., 2017). On the other hand, it is also possible that our sample of behaviour. As discussed in chapter 5, a previous study suggests that sleep bipolar patients has circadian rhythm disturbances but showed no overt sign deprived individuals show an amplified reactivity in the mesolimbic network of sleep-wake disturbances due to, for example, effective pharmacological after exposure to pleasure-invoking stimuli (Gujar, Yoo, Hu, & Walker, 2011). Chapter 6 treatment or social rhythm therapy (Hirschfeld, 2005). The direct association This could make adolescents with sleep disturbances more susceptible to the between sleep and mood symptomatology implies that bipolar sleep rewarding experiences that follow health risk behaviours and thus increase the disturbances are mainly state dependent, rather than a trait of the disorder. This chance of these adolescents engaging in this type of behaviour. is corroborated by the fact that, while being heritable traits, the sleep pattern of non-affected siblings is similar to that of controls (Pagani et al., 2016). When Another addition to the model suggested in chapter 1 is the interaction between we revert back to the model suggested in chapter 1, our findings imply that in sleep and white matter microstructure. In chapter 4 we concluded that a more bipolar disorder, the most noteworthy finding appears in the relation between healthy sleep pattern is related to higher fractional anisotropy (FA), suggesting

Summary and general discussion emotional deregulation and sleep disturbances, that could result in bipolar mood better integrity of white matter. This could subsequently imply that sleep

148 149 disturbances lead to impairments in white matter microstructure. Indeed, sleep disturbances. However, the relation may also be reversed, in which previous studies found that sleep promotes myelination and oligodendrocyte psychosocial problems and health risk behaviour predispose to impaired sleep, precursor cell proliferation (Bellesi, 2015). Whether impairments in white perhaps because adolescents were more emotionally, cognitively or physically matter microstructure that follow sleep disturbances are, in turn, involved in aroused around bedtime. emotional deregulation cannot be concluded from our results. However, a study by Zou et al. (2008) hypothesized that depressive symptoms may result from A second limitation in chapter 5 is the use of self-report questions regarding alterations in white matter microstructure in the frontothalamic circuits. Of sleep disturbances. In the previous chapters we used objective measures of note is that we did not establish a similar relation between sleep pattern and sleep, which is generally preferred over self-report measures, given the general FA in bipolar patients. On the one hand this could suggest that, although problem of reliably recalling sleep from previous nights (Lockley, Skene, & bipolar patients show wide-spread deviations in white matter microstructure, Arendt, 1999). However, the down-side of objective measurements is that large, these alteration are unrelated or inversely related to the sleep pattern in this population-wide screenings are generally unfeasible. In chapter 5 we used the specific population. On the other hand, we showed that the use of antipsychotic question whether the respondents had trouble sleeping in the previous month. medication is a likely explanation for this counterintuitive association. Future Although this question covers the essential problem of insomnia, there was no research could address the issue of confounding by medication of the association way we could differentiate between problems initiating sleep or maintaining between sleep pattern and white matter microstructure in bipolar disorder. sleep.

Limitations A third, and recurring problem in the field of sleep research in psychiatric There are several limitations that need to be taken into consideration when populations, is the issue of pharmacological treatments. Although we know interpreting the results of this thesis. The first is the cross-sectional design of that several types of medications affect sleep-wake patterns, we were unable to our studies. This means that the associations that we identified cannot provide fully account for these effects in our statistical analyses. For one, since almost firm conclusions regarding causality. This is specifically relevant in chapter 4, all of our bipolar patients used some form of pharmacological treatment at where we found associations between integrity of white matter microstructure time of testing, we could not use medication as covariate in our case-control Chapter 6 and several parameters of sleep patterns. The options that sleep disturbances comparisons. Also, confounding by indication can be at play, in which specific cause alterations in white matter microstructure or that brain abnormalities types of patients receive specific types of medication. For example, more severely cause problems in sleep-wake patterns are both plausible, although experimental ill patients may receive sedating antipsychotic medication while more sleep- studies point towards the first option (Bellesi et al., 2013; Elvsåshagen et deprived patients more often use benzodiazepines. Lastly, a large group of al., 2015). Similarly, in chapter 5 we found associations between reported patients used more than one type of medication, so differentiating the specific sleep disturbances on the one hand and psychosocial problems, suicidality effect per medication type is nearly impossible. and behavioural problems on the other. In this case, one could argue that

Summary and general discussion psychosocial problems and engagement in health risk behaviour resulted from

150 151 Lastly, the measurements in chapter 4 (i.e. actigraphy and MRI data) were References collected with an average time difference of 1.4 years between both protocols.

The possibility exists that activity patterns or sleep behaviour changed during Bellesi, M. (2015). Sleep and oligodendrocyte functions. Current Sleep Medicine Reports, 1(1), those months, resulting in variation in white matter microstructure that was 20–26. not measured at time of MRI scanning. However, the time between both Bellesi, M., Pfister-Genskow, M., Maret, S., Keles, S., Tononi, G., & Cirelli, C. (2013). Effects of sleep and wake on oligodendrocytes and their precursors. The Journal of Neuroscience, measurements was similar for bipolar patients and controls, and there is no 33(36), 14288–300. reason to assume that such variation was systematic and has biased the results. Cervantes, P., Gelber, S., Ng Ying Kin, F. N. K., Nair, V. N. P., & Schwartz, G. (2001). Circadian secretion of cortisol in bipolar disorder. Journal of Psychiatry and Neuroscience, General conclusion 26(5), 411–416.

In this thesis we found strong support for the relation between sleep disturbances Dallaspezia, S., & Benedetti, F. (2009). Melatonin, circadian rhythms, and the clock genes in and psychopathology in two separate populations: bipolar disorder patients and bipolar disorder. Current Psychiatry Reports, 11(6), 488–93.

adolescents from the general population. Although we cannot conclude that Elvsåshagen, T., Norbom, L. B., Pedersen, P. Ø., Quraishi, S. H., Bjørnerud, A., Malt, U. F., bipolar disorder is characterized by trait-like dysfunctions in sleep-wake pattern, Westlye, L. T. (2015). Widespread changes in white matter microstructure after a day of waking and sleep deprivation. PLoS ONE, 10(5), 1–15. we found relations with mood symptoms suggesting that sleep is an important factor in the (recurrence of) mood episodes. In adolescents we found a similar Geoffroy, P. A., Scott, J., Boudebesse, C., Lajnef, M., Henry, C., Leboyer, M., … Etain, B. (2015). Sleep in patients with remitted bipolar disorders: a meta-analysis of actigraphy pattern in which sleep disturbances in the previous month were related to a studies. Acta Psychiatrica Scandinavica, 131(2), 89–99. wide range of psychosocial and behavioural problems. All in all these findings Gujar, N., Yoo, S.-S., Hu, P., & Walker, M. (2011). Sleep deprivation amplifies reactivity of indicate that both bipolar patients and adolescents are impaired by sleep brain reward networks, biasing the appraisal of positive emotional experiences. The Journal of disturbances and that in both populations increased vigilance from physicians, Neuroscience, 31(12), 4466–74.

mental health care workers and school staff is warranted when disturbances in Hirschfeld, R. M. A. (2005). Guideline watch: Practice guideline for the treatment of patients with sleep-wake patterns are presented. bipolar disorder, 2nd edition. Arlington, VA: American Psychiatric Association. Chapter 6 Hou, L., Bergen, S. E., Akula, N., Song, J., Hultman, C. M., Landén, M., McMahon, F. J. (2016). Genome-wide association study of 40,000 individuals identifies two novel loci associated with bipolar disorder. Human Molecular Genetics, 25(15), 3383–3394.

Ikeda, M., Takahashi, A., Kamatani, Y., Okahisa, Y., Kunugi, H., Mori, N., … Iwata, N. (2017). A genome-wide association study identifies two novel susceptibility loci and trans population polygenicity associated with bipolar disorder. Molecular Psychiatry.

Lockley, S. W., Skene, D. J., & Arendt, J. (1999). Comparison between subjective and actigraphic measurement of sleep and sleep rhythms. Journal of Sleep Research, 8(3), 175–183.

Ng, T. H., Chung, K. F., Ho, F. Y. Y., Yeung, W. F., Yung, K. P., & Lam, T. H. (2015). Sleep-

Summary and general discussion wake disturbance in interepisode bipolar disorder and high-risk individuals: A systematic review and meta-analysis. Sleep Medicine Reviews, 20, 46–58.

152 153 Nurnberger, J. I., Adkins, S., Lahiri, D. K., Mayeda, A., Hu, K., Lewy, A., … Davis-Singh, D. (2000). Melatonin suppression by light in euthymic bipolar and unipolar patients. Archives of General Psychiatry, 57, 572–579.

Pagani, L., St. Clair, P. a., Teshiba, T. M., Service, S. K., Fears, S. C., Araya, C., … Freimer, N. B. (2016). Genetic contributions to circadian activity rhythm and sleep pattern phenotypes in pedigrees segregating for severe bipolar disorder. Proceedings of the National Academy of Sciences, 113(6), E754–E761.

Wehr, T., Goodwin, F., Wirz-Justice, A., Breitmaier, J., & Craig, C. (1982). 48-hour sleep- wake cycles in manic-depressive illness: naturalistic observations and sleep deprivation experiments. Archives of General Psychiatry, 39(5), 559–565.

Zou, K., Huang, X., Li, T., Gong, Q., Li, Z., Ou-yang, L., … Sun, X. (2008). Alterations of white matter integrity in adults with major depressive disorder: a magnetic resonance imaging study. Journal of Psychiatry & Neuroscience, 33(6), 525–30. Chapter 6 Summary and general discussion

154 155 Chapter 7

Nederlandse samenvatting Slaap is een onderwerp dat velen al eeuwenlang fascineert. Inmiddels heeft de kenmerken van de ziekte en mogelijk onderliggende biologische mechanismen. wetenschap zich hier in toenemende mate in verdiept en zijn we veel te weten Dit zou uiteindelijk kunnen leiden tot de ontwikkeling van nieuwe interventies gekomen over bijvoorbeeld de functie van slaap en de mechanismen in de en behandelmogelijkheden. hersenen die verantwoordelijk zijn voor het dagelijkse patroon van inslapen en wakker worden. Het merendeel van de lezers zal proefondervindelijk hebben In hoofdstuk 3 hebben we ons gericht op het slaap-waak fenotype van ondervonden dat slaap cruciaal is voor het dagelijks functioneren. Een breed bipolaire patiënten, hun niet-bipolaire broers en zussen en controles. Hierbij scala aan studies heeft aangetoond dat acute slaapdeprivatie kan leiden tot hebben we gebruik gemaakt van activiteitsmeters die rond de pols worden problemen met controle over ons bewegingsapparaat, verstoring van cognitieve gedragen (actiegrafie) en gedurende 2 weken de gemiddelde activiteit per processen, stofwisselingsproblemen, veranderingen in de hormoonhuishouding, minuut registreerden van onze deelnemers. Aan de hand van deze gegevens, in problemen met het immuunsysteem, verstoorde breinactiviteit en verminderd combinatie met een slaapdagboek dat werd bijgehouden, konden wij kenmerken welbevinden (zie voor een overzichtsartikel Orzeł-Gryglewska, 2010). Het van het slaappatroon berekenen, zoals de totale slaapduur, het moment van gegeven dat slaapproblemen de laatste jaren steeds meer lijken toe te nemen inslapen en ontwaken, het aantal keer ontwaken, sluimeren in de ochtend en de in onze moderne 24-uurs maatschappij is derhalve een reden tot zorg en slaapefficiëntie (de slaapslaapduur/de totale tijd in bed). We vonden dat bipolaire tevens de reden dat wij in dit proefschrift het verband tussen slaapproblemen patiënten vergeleken met hun broers en zussen langer sliepen, en dan met en psychopathologie hebben onderzocht. Hierbij hebben we ons gericht name in de ochtend. Dit resultaat leek echter verklaard te kunnen worden door op patiënten met een bipolaire stoornis en adolescenten uit de algemene verschillen in stemming tijdens de meetperiode. Wanneer we ons beperkten tot populatie. De reden hiervoor is dat beide groepen gekenmerkt worden door euthyme patiënten (dus patiënten zonder depressieve of manische symptomen), veranderingen in stemming en kwetsbaar zijn voor veranderingen in het vonden we geen enkel verschil in slaappatroon. We vonden daarnaast een slaap-waak ritme. In dit hoofdstuk geven we een overzicht van de algemene lineair verband tussen het slaappatroon en depressieve symptomen. Vervolgens bevindingen uit het proefschrift, beschrijven we hoe deze bevinding relateren hebben wij onze gegevens samengevoegd met eerdere onderzoeksgegevens tot de onderzoeksvragen, worden de kanttekeningen van de studies besproken uit het veld, en vonden toen, in tegenstelling tot onze eigen bevindingen, dat en geven we suggesties voor toekomstig onderzoek. bipolaire patiënten wél langer sliepen, langer wakker lagen voor het slapen, vaker wakker werden en een lagere slaapefficiëntie hadden dan controles. Een Samenvatting belangrijke kanttekening hierbij is dat de meeste verschillen klein waren en de Chapter 7 In hoofdstuk 2 gaven we een overzicht van de karakteristieken, ook wel totale slaaptijd nooit langer was dan 9 uur, wat nog steeds binnen de marges fenotypes genoemd, van de bipolaire stoornis, en beschrijven we wat de van gezonde slaap voor volwassenen valt. Deze bevindingen suggereren dat klinische relevantie is van fenotypes, zoals circadiane (dag-nacht) ritmiek, er mogelijk wel verschillen zijn in slaappatroon tussen patiënten en controles, slaapproblemen en neuroanatomie. Een dergelijke gedetailleerde beschrijving maar dat deze verschillen wellicht niet heel klinisch relevant zijn. Onze laatste van het fenotype van de stoornis, in combinatie met genoom-wijde genetische bevinding in dit hoofdstuk is dat geen van de klinische karakteristieken

Nederlandse samenvatting data, geeft ons de mogelijkheid om de interactie te bestuderen tussen zichtbare van de stoornis (leeftijd waarop de ziekte zich voor het eerst openbaarde, het

158 159 aantal stemmingsepisodes en psychotische symptomen) samenhing met het gebruik van alcohol, tabak, drugs en compulsief gebruik van multimedia. We slaappatroon van bipolaire patiënten. vonden dat de relatie tussen slaapproblemen enerzijds en suïcidepogingen en cannabisgebruik anderzijds sterker was bij meisjes dan bij jongens. Verder Hoofdstuk 4 richtte zich op het verband tussen integriteit van de zenuwbanen hebben we gekeken naar de invloed van emotionele problemen op de relatie in witte stof en het slaap- en activiteitenpatroon van bipolaire patiënten en tussen slaapproblemen en risicogedrag. Het bleek dat emotionele problemen de controles. De integriteit van deze witte stof banen hebben we gemeten met relatie tussen slaapproblemen en compulsief multimediagebruik gedeeltelijk kon behulp van een MRI-scan waaruit we de fractionele anisotropie (FA) hebben verklaren. Toch vonden we ook dat slaapproblemen nog steeds samenhingen kunnen afleiden, wat een maat is voor diffusie van watermoleculen in het brein. met een 2 keer zo grote kans op dit risicogedrag wanneer emotionele problemen Watermoleculen zullen zich in het brein namelijk niet vrij bewegen, maar waren meegenomen in de analyses. Ten slotte vonden we dat emotionele zullen zich meer in de richting van de aanwezige witte stof banen verplaatsen. problemen de relatie tussen slaapproblemen en middelengebruik in het geheel Hierbij nemen we aan dat een hogere FA-waarde wijst op betere integriteit niet leek te beïnvloeden. van de witte stof verbindingen. Een hogere integriteit van deze verbindingen zou onder andere verklaard kunnen worden door betere isolatie van de axonen Conclusies (myelinisatie) (Beaulieu, 2002). Onze studie liet allereerst een afname in FA Een van de meest opvallende bevindingen in dit proefschrift is dat we zien bij bipolaire patiënten, verspreid over het gehele brein. Een hogere mate geen overtuigend bewijs hebben kunnen vinden voor de hypothese dat van activiteit overdag en in de avond was verder gerelateerd aan hogere FA, slaapproblemen een kenmerk zijn van de bipolaire stoornis, buiten manisch en zowel bij bipolaire patiënten als controles. De associatie tussen slaap en FA gaf depressieve stemmingsepisodes. Deze conclusie wijkt af van twee eerdere meta- een wat tegenstrijdiger beeld. Bij controles vonden we het verwachte positieve analyses die concludeerden dat euthyme bipolaire patiënten wel afwijkingen verband tussen gezonde slaap en FA, waarbij gezondere slaap samenhing met in het slaappatroon laten zien ten opzichte van controles. Belangrijk is echter hogere FA. Bij bipolaire patiënten vonden we echter het tegenovergestelde: dat de tot nu toe gevonden verschillen vaak klein waren en inconsistent over een langere slaapduur, eerder tijdstip van inslapen en korter wakker liggen in verschillende meetmethodes (Geoffroy et al., 2015; Ng et al., 2015). Als de ochtend waren gerelateerd aan hogere FA. Onze resultaten laten zien dat grootste actigrafiestudie tot nu toe trekken wij daarom de conclusie in ziekte-gerelateerde factoren, zoals het gebruik van antipsychotica, hier zeer twijfel dat deze verschillen wijzen op klinisch relevante afwijkingen in het waarschijnlijk de oorzaak van zijn. slaappatroon. Als het slaap-waakpatroon van euthyme bipolaire patiënten Chapter 7 inderdaad normaal is, dan zou dat erop kunnen wijzen dat de bipolaire stoornis In hoofdstuk 5 hebben we slaapproblemen van adolescenten onderzocht in een niet wordt gekenmerkt door verstoorde circadiane ritmiek zoals eerder werd grote groep middelbare scholieren. Hier vonden we dat ongeveer 20% van de geconcludeerd uit studies die afwijkingen vonden in lichaamstemperatuur, adolescenten problemen met slapen rapporteert. Deze slaapproblemen hingen melatonineproductie en cortisolafgifte (Cervantes, Gelber, Ng Ying Kin, Nair, sterk samen met psychosociale problemen, suïcidegedachtes en suïcidepogingen. & Schwartz, 2001; Nurnberger et al., 2000; Wehr, Goodwin, Wirz-Justice,

Nederlandse samenvatting Ook vonden we associaties met een breed scala aan risicogedrag, zoals Breitmaier, & Craig, 1982). Er zijn inderdaad, behalve een aantal correlationele

160 161 verbanden tussen circadiane gen-variaties en de klinische karakteristieken gedragsproblemen kan leiden, zoals middelengebruik en compulsief gebruik van de bipolaire stoornis, vooralsnog geen genoom-wijde associatie studies van multimedia. Een opvallend punt daarbij is dat deze gedragsproblemen niet geweest die circadiane genen hebben geïdentificeerd (Dallaspezia & Benedetti, volledig verklaard werden door emotionele problemen. Dit zou kunnen betekenen 2009; Hou et al., 2016; Ikeda et al., 2017). Aan de andere kant zou het ook dat niet alleen de relatie tussen slaapproblemen en emotionele deregulatie mogelijk kunnen zijn dat ons sample van bipolaire patiënten toch onderliggende uiteindelijk kan leiden tot gedragsproblemen, maar dat slaapdeprivatie zelf ook afwijkingen heeft in het circadiane systeem, maar geen zichtbare ontregeling risicogedrag tot gevolg kan hebben. Zoals we ook in hoofdstuk 5 beschrijven van het slaap-waak ritme laten zien, bijvoorbeeld als gevolg van medicamenteuze concludeerde een eerdere fMRI-studie dat proefpersonen met slaaptekort behandeling of sociaal ritme therapie (Hirschfeld, 2005). Het directe verband een verhoogde reactie hadden in het mesolimbische netwerk nadat ze waren tussen slaap en stemmingssymptomatologie impliceert echter dat bipolaire blootgesteld aan plezierige stimuli (Gujar, Yoo, Hu, & Walker, 2011). Deze slaapproblemen met name episode-afhankelijk zijn. Daar komt nog bovenop dat verhoogde reactie zou adolescenten met slaapproblemen gevoeliger maken voor niet-bipolaire broers en zussen hetzelfde slaappatroon laten zien als controles, de belonende ervaring die volgt op risicovol gedrag, waardoor ze mogelijk sneller terwijl eerder onderzoek heeft aangetoond dat deze slaapkarakteristieken wel geneigd zijn dergelijk gedrag te laten zien. degelijk erfelijke eigenschappen zijn (Pagani et al., 2016). Onze opvallendste resultaten bevinden zich derhalve op het niveau van de relatie tussen emotionele Ten slotte hebben we de relatie tussen slaap en de witte stof verbindingen deregulatie en slaapproblemen welke uiteindelijk kunnen resulteren in bipolaire onderzocht, waarbij we concludeerden dat een gezonder slaappatroon stemmingsklachten. Ons onderzoek wijst niet op een stabiele deregulatie in de samenhangt met hogere FA-waardes en dat slaapproblemen zouden kunnen circadiane biologie, al zou toekomstig onderzoek, met behulp van bijvoorbeeld leiden tot verminderde integriteit van de witte stof. Een verklaring hiervoor is circadiane oscillatoren in fibroblasten van patiënten, hier meer uitsluitsel over de eerdere bevinding dat slaap myelinisatie stimuleert, waardoor een gebrek kunnen geven. Ook zouden meer grootschalige longitudinale actigrafiestudies aan slaap negatieve gevolgen kan hebben voor de microstructuur van witte stof het temporele verband tussen slaapproblemen en emotionele deregulatie (Bellesi, 2013). Of de verstoorde witte stof verbindingen die voortkomen uit kunnen onderzoeken om vast te stellen welke van deze twee factoren het eerst slaapproblemen vervolgens weer een rol spelen bij emotionele deregulatie kan afwijkingen laat zien. uit onze resultaten niet worden geconcludeerd. Een studie van Zou en collega’s (2008) hebben hier echter wel in meer detail naar gekeken en concludeerden De associatie tussen slaapproblemen en emotionele deregulatie werd verder dat depressieve symptomen inderdaad het gevolg kunnen zijn van afwijkingen Chapter 7 ondersteund in ons sample van niet-klinische adolescenten. Hier vonden we in witte stof verbindingen van frontothalamische circuits. Belangrijk om sterke associaties tussen slaapproblemen enerzijds en psychosociale problemen te vermelden is dat we niet een positief verband tussen slaap en FA konden en suïcidaliteit anderzijds. Ook vonden we associaties tussen slaapproblemen aantonen bij bipolaire patiënten. Dit zou aan de ene kant kunnen betekenen en risicogedrag die gedeeltelijk werden verklaard door emotionele problemen. dat de witte stof van bipolaire patiënten, welke juist globale afwijkingen laat Dit suggereert dat slaapproblemen en de daarmee samenhangende emotionele zien ten opzichte van controles, geen verband of een tegengesteld verband heeft

Nederlandse samenvatting deregulatie niet alleen tot psychiatrische symptomen, maar ook tot met het slaappatroon in deze populatie. Aan de andere kant hebben we in onze

162 163 studie laten zien dat een waarschijnlijke verklaring voor dit verband te maken objectieve slaapmetingen is echter dat grotere bevolkingsonderzoeken doorgaans heeft met het gebruik van antipsychotica. Verder onderzoek zou verder kunnen niet haalbaar zijn. In hoofdstuk gebruikten we de vraag of de respondenten ingaan op de invloed van medicatie op dit contra-intuïtieve verband tussen slaap problemen hadden met slapen in de afgelopen maand. Hoewel deze vraag het en witte stof bij bipolaire patiënten. essentiële probleem van slapeloosheid omvat, konden we geen onderscheid maken tussen in- of doorslaapproblemen. Een volgend, en veel voorkomend Limitaties probleem binnen het onderzoeksveld van slaap in de psychiatrie, is het gebruik Er zijn een aantal limitaties die in overweging moeten worden genomen bij van medicatie. Hoewel we weten dat verschillende types medicatie het slaap- de interpretatie van de resultaten van dit proefschrift. Allereerst het cross- waak patroon beïnvloeden, was het niet mogelijk om deze effecten overal sectionele design van onze studies, waarbij alle proefpersonen slechts eenmalig mee te nemen in onze statistische analyses. Allereerst gebruikten vrijwel alle gemeten zijn. Dit betekent dat de verbanden die we hebben gevonden niets bipolaire patiënten medicatie tijdens onze meting, waardoor we medicatie kunnen zeggen over causaliteit. Dit is met name relevant in hoofdstuk 4, waar niet konden meenemen in onze patiënt-controle vergelijking. Ook is het in de we het verband tussen witte stof verbindingen en het slaappatroon hebben regel vaak zo dat bepaalde patiënten bepaalde soorten medicatie vaker krijgen onderzocht. Dit zou kunnen betekenen dat slaapproblemen zorgen voor voorgeschreven. Patiënten die ernstiger ziek zijn zullen bijvoorbeeld vaker een afnames in de kwaliteit van deze verbindingen, maar evengoed dat afwijkingen sederend antipsychoticum krijgen, terwijl patiënten met slaapproblemen vaker in het brein er juist voor zorgen dat er problemen in het slaap-waak patroon een benzodiazepine krijgen. Ten slotte gebruikten veel patiënten meer dan 1 ontstaan, al wijzen experimentele studies in de richting van de eerste optie soort medicatie, waardoor we de effecten per type medicatie onmogelijk van (Bellesi et al., 2013; Elvsåshagen et al., 2015). Een vergelijkbare kanttekening elkaar konden onderscheiden. De laatste kanttekening die we willen noemen kan geplaatst worden bij hoofdstuk 5, waarin we een associatie vonden tussen heeft te maken met de metingen in hoofdstuk 4, namelijk de actigrafie en MRI enerzijds gerapporteerde slaapproblemen en anderzijds psychosociale problemen, gegevens. Deze zijn namelijk verzameld met een gemiddeld tijdsverschil van 1.4 suïcidaliteit en gedragsproblemen. Hier zou men kunnen beargumenteren dat jaar tussen beide metingen. Het zou mogelijk kunnen zijn dat het activiteiten- of psychosociale problemen en risicogedrag resulteren uit het slaaptekort. Echter, slaappatroon veranderd is tijdens deze maanden, waardoor variaties in witte stof deze relatie zou ook omgedraaid kunnen zijn, waarbij psychosociale problemen ontstonden die niet gemeten zijn tijdens de MRI-scan. Echter, de tijd tussen de en risicogedrag juist kwetsbaar maken voor verstoorde slaap, mogelijk omdat twee metingen was gelijk voor bipolaire patiënten dus is er geen reden om aan te adolescenten als gevolg van deze psychosociale problemen of risicogedrag nemen dat de variatie onze resultaten heeft beïnvloed. Chapter 7 emotioneel, cognitief of fysiek geprikkeld zijn rond bedtijd. Een tweede limitatie in hoofdstuk 5 is het feit dat slaapproblemen middels vragenlijsten zijn uitgevraagd. In de voorgaande hoofdstukken gebruikten we objectieve metingen van slaap, welke over het algemeen geprefereerd worden ten opzichte van vragenlijsten, gezien de problemen met het betrouwbaar terughalen van

Nederlandse samenvatting slaap in voorgaande nachten (Lockley, Skene, & Arendt, 1999). Het nadeel van

164 165 Algemene conclusies Referenties In dit proefschrift vonden we sterk bewijs voor de relatie tussen slaapproblemen

en psychopathologie in twee populaties: bipolaire patiënten en adolescenten Beaulieu, C. (2002). The basis of anisotropic water diffusion in the nervous system - a technical uit de algemene bevolking. Hoewel we niet kunnen concluderen dat review. NMR in Biomedicine, 15, 435–455. bipolaire stoornis wordt gekenmerkt door structurele problemen in het slaap- Bellesi, M., Pfister-Genskow, M., Maret, S., Keles, S., Tononi, G., & Cirelli, C. (2013). Effects of sleep and wake on oligodendrocytes and their precursors. The Journal of Neuroscience, waakpatroon, hebben we wel relaties gevonden met stemminsymptomen, 33(36), 14288–300. wat suggereert dat slaap een belangrijke factor is bij (de terugkeer van) Cervantes, P., Gelber, S., Ng Ying Kin, F. N. K., Nair, V. N. P., & Schwartz, G. (2001). stemminsepisodes. We vonden een vergelijkbaar patroon bij adolescenten, Circadian secretion of cortisol in bipolar disorder. Journal of Psychiatry and Neuroscience, waarbij slaapproblemen in de voorgaande maand samenhingen met een breed 26(5), 411–416.

scala aan psychosociale en gedragsproblemen. Deze bevindingen wijzen erop dat Dallaspezia, S., & Benedetti, F. (2009). Melatonin, circadian rhythms, and the clock genes in zowel bipolaire patiënten als adolescenten kwetsbaar zijn voor de gevolgen van bipolar disorder. Current Psychiatry Reports, 11(6), 488–93.

slaapproblemen en dat het voor beide groepen noodzakelijk is dat artsen, GGZ- Elvsåshagen, T., Norbom, L. B., Pedersen, P. Ø., Quraishi, S. H., Bjørnerud, A., Malt, U. F., medewerkers en onderwijspersoneel waakzaam zijn wanneer slaapproblemen Westlye, L. T. (2015). Widespread changes in white matter microstructure after a day of waking and sleep deprivation. PLoS ONE, 10(5), 1–15. worden gerapporteerd. Geoffroy, P. A., Scott, J., Boudebesse, C., Lajnef, M., Henry, C., Leboyer, M., … Etain, B. (2015). Sleep in patients with remitted bipolar disorders: a meta-analysis of actigraphy studies. Acta Psychiatrica Scandinavica, 131(2), 89–99.

Gujar, N., Yoo, S.-S., Hu, P., & Walker, M. (2011). Sleep deprivation amplifies reactivity of brain reward networks, biasing the appraisal of positive emotional experiences. The Journal of Neuroscience, 31(12), 4466–74.

Hirschfeld, R. M. A. (2005). Guideline watch: Practice guideline for the treatment of patients with bipolar disorder, 2nd edition. Arlington, VA: American Psychiatric Association.

Hou, L., Bergen, S. E., Akula, N., Song, J., Hultman, C. M., Landén, M., McMahon, F. J. (2016). Genome-wide association study of 40,000 individuals identifies two novel loci associated with bipolar disorder. Human Molecular Genetics, 25(15), 3383–3394.

Ikeda, M., Takahashi, A., Kamatani, Y., Okahisa, Y., Kunugi, H., Mori, N., … Iwata, N. Chapter 7 (2017). A genome-wide association study identifies two novel susceptibility loci and trans population polygenicity associated with bipolar disorder. Molecular Psychiatry.

Lockley, S. W., Skene, D. J., & Arendt, J. (1999). Comparison between subjective and actigraphic measurement of sleep and sleep rhythms. Journal of Sleep Research, 8(3), 175–183.

Ng, T. H., Chung, K. F., Ho, F. Y. Y., Yeung, W. F., Yung, K. P., & Lam, T. H. (2015). Sleep-

Nederlandse samenvatting wake disturbance in interepisode bipolar disorder and high-risk individuals: A systematic review and meta-analysis. Sleep Medicine Reviews, 20, 46–58.

166 167 Nurnberger, J. I., Adkins, S., Lahiri, D. K., Mayeda, A., Hu, K., Lewy, A., … Davis-Singh, D. (2000). Melatonin suppression by light in euthymic bipolar and unipolar patients. Archives of General Psychiatry, 57, 572–579.

Orzeł-Gryglewska, J. (2010). Consequences of sleep deprivation. International Journal of Occupational Medicine and Environmental Health, 23(1), 95–114.

Pagani, L., St. Clair, P. A., Teshiba, T. M., Service, S. K., Fears, S. C., Araya, C., … Freimer, N. B. (2016). Genetic contributions to circadian activity rhythm and sleep pattern phenotypes in pedigrees segregating for severe bipolar disorder. Proceedings of the National Academy of Sciences, 113(6), E754–E761.

Wehr, T., Goodwin, F., Wirz-Justice, A., Breitmaier, J., & Craig, C. (1982). 48-hour sleep- wake cycles in manic-depressive illness: naturalistic observations and sleep deprivation experiments. Archives of General Psychiatry, 39(5), 559–565.

Zou, K., Huang, X., Li, T., Gong, Q., Li, Z., Ou-yang, L., … Sun, X. (2008). Alterations of white matter integrity in adults with major depressive disorder: a magnetic resonance imaging study. Journal of Psychiatry & Neuroscience, 33(6), 525–30. Chapter 7 Nederlandse samenvatting

168 169 List of Publications Curriculum Vitae Dankwoord Knapen, S.E., Verkooijen, S., Riemersma-van der Lek, R.F., Kahn, R.S., Ophoff, R.A., Boks, List of publications M.P.M., Schoevers, R.A. Circadian rhythm disturbances in bipolar disorder: not a trait phenomenon.

Publications van Bergen, A.H., Verkooijen, S., Vreeker, A., Abramovic, L., Hilligers, M.H., Spijker, A.T., Hoencamp, E., Regeer, E.J., Riemersma - van der Lek, R.F., Schroevers, R.A., Stevens, Verkooijen, S., van Bergen, A.H., Knapen, S.E., Vreeker, A., Abramovic, L., Pagani, L., A.W., Schulte, R.F., Vonk, R., Hoekstra, R., van Beveren, N.J., Kupka, R.W., Sommer, Jung, Y., Riemersma-van der Lek, R., Schoevers, R.A., Takahashi, J.S., Kahn, R.S., Boks, I.E.C., Ophoff, R.A., Kahn, R.S., Boks, M.P.M. The characteristics of psychotic features in M.P.M., Ophoff, R.A. (2017). An actigraphy study investigating sleep in bipolar I patients, Bipolar Disorder. (under review). unaffected siblings and controls.Journal of Affective Disorder (208)248-254. Stevelink, R., Abramovic, L., Verkooijen, S., Ophoff, R.A., Begemann, M.J.H., Cahn, W., Verkooijen, S., Stevelink, R., Abramovic, L., van Haren, N.E.M., Vinkers, C.H., Ophoff, Kahn, R.S., Sommer, I.E.C., Boks, M.P.M., Mandl, R., van Haren, N.E.M., Vinkers, C.H. R.A., Boks, M.P.M., Kahn, R.S. (2017). The association of sleep and physical activity with The relation between childhood abuse and white matter integrity in bipolar disorder patients integrity of white matter microstructure in bipolar disorder patients and healthy controls. and healthy controls. (under review). Psychiatry Research, 262, 71-80. Abramovic, L., Boks, M.P.M., Vreeker, A., Verkooijen, S., van Bergen, A.H., Ophoff, R.A., Daalman, K., Verkooijen, S., Derks, E.M., Aleman, A., Sommer, I. E. C. (2012) The influence Kahn, R.S., van Haren, N.E.M. White matter disruptions in patients with bipolar disorder. of semantic top-down processing in auditory verbal hallucinations. Schizophrenia Research (under review). (139), 82-86.

Vreeker, A., Boks, M.P.M., Abramovic, L., Verkooijen, S., van Bergen, A.H., Spijker, A.T., Hoencamp, E., Regeer, E.J., Riemersma-van der Lek, R.F., Stevens, A.W.M.M., Schulte, P.J.F., Vonk, R., Hoekstra, R., van Beveren, N.J.M., Kupka, R.W., Brouwer, R.M., Bearden, C.E., MacCabe, J.H., Ophoff, R.A., GROUP investigators. (2016). High educational performance is a distinctive feature of bipolar disorder; a study on cognition in bipolar disorder, schizophrenia patients and controls. Psychological Medicine, 46(4), 807-818.

Vreeker, A., Abramovic, L., Boks, M.P.M., Verkooijen, S, van Bergen, A.H., Ophoff, R.A., Kahn, R.S., van Haren, N.E.M. (2017). The relationship between brain volume and intelligence in bipolar disorder. Journal of Affective Disorder, 223, 59-64.

Abramovic, L., Boks M.P.M., Vreeker, A., Bouter, D.C., Kruiper, C., Verkooijen S, van Bergen, A.H., Ophoff, R.A., Kahn, R.S., van Haren N.E.M. (2016). The association of antipsychotic medication and lithium with brain measures in patients with bipolar disorder. European Neuropsychopharmacology, 26(11), 1741-1751.

Planned Publications

Verkooijen, S., de Vos, N., Bakker, B.,Branje, S.J.T., Kahn, R.S., Ophoff, R.A.,Plevier, C.M., Boks, M.P.M. Sleep disturbances, psychosocial difficulties and health risk behaviour in 16,781 Dutch adolescents. (under review).

172 173 Curriculum Vitae

Susanne Verkooijen was born on September 4th 1987 in Middelburg. She Susanne Verkooijen is geboren op 4 september 1987 te Middelburg. In finished her secondary education at the Stedelijke Scholengemeenschap 2005 rondde ze haar Atheneum af aan de Stedelijke Scholengemeenschap Nehalennia in Middelburg in 2005, after which she started pursuing her Nehalennia, waarna ze startte met de bachelor Psychologie aan de Universiteit Bachelor’s degree in Psychology at Utrecht University. After graduating in 2009, Utrecht. Nadat ze haar diploma behaalde in 2009 stopte ze een jaar met she took a year off from studying and travelled to India where she did voluntary studeren en vertrok naar India waar ze onder andere vrijwilligerswerk deed work in a school for mentally disabled children. In 2010 she started the Master in een school voor gehandicapte kinderen. In 2010 begon ze aan de master program Clinical Health Psychology and did both her research and clinical Klinische en Gezondheidspsychologie waarvoor ze zowel haar onderzoeksstage internships at the Department of Psychiatry at University Medical Center als haar klinische stage volbracht op de afdeling Psychiatrie van het Universitair Utrecht. During this year, she wrote a thesis on auditory verbal hallucinations Medisch Centrum Utrecht. Hier schreef ze haar scriptie over auditieve verbale under the supervision of dr. Daalman and prof. dr. Sommer. After graduating, hallucinaties onder begeleiding van dr. Daalman and prof. dr. Sommer. Na het she continued her career at the University Medical Center Utrecht to work as behalen van het masterdiploma vervolgde ze haar carrière aan het UMC Utrecht a research assistant at Bipolar Genetics (Dutch Bipolar Cohort Study). A year als onderzoeksassistent van Bipolar Genetics (Dutch Bipolar Cohort Study). later, in 2012, she got the opportunity to start a PhD-program under the Een jaar later kreeg ze de kans om te starten met een promotieonderzoek onder supervision of prof. dr. Kahn, prof. dr. Ophoff and dr. Boks, focusing on sleep begeleiding van prof. dr. Kahn, prof. dr. Ophoff and dr. Boks waar ze zich and psychopathology. During this program, she also took part in a US exchange richtte op het verband tussen slaap en psychopathologie. Ook nam ze deel aan program, where she worked at the laboratories of prof. dr. Ophoff (University of een uitwisselingsprogramma met de VS, waar ze werkte op de laboratoria van California Los Angeles, California) and prof. dr. Takahashi (UT Southwestern, prof. dr. Ophoff (University of California Los Angeles, Californië) en prof. dr. Texas). After her years as a PhD-student, she got increasingly motivated to Takahashi (UT Southwestern, Texas). Na jaren gewerkt te hebben binnen de use her knowledge on sleep and psychopathology in a clinical setting and has wetenschap werd ze in toenemende mate gemotiveerd om haar kennis op het since developed e-health content for mental health professionals at ZorgIQ gebied van slaap en psychopathologie toe te passen in een klinische setting. and worked as a psychologist in a sleep clinic (Chronomed Wassenaar) and at a Sindsdien heeft ze e-healthprogramma’s voor de GGZ helpen ontwikkelen general practitioner’s office (Huisartsenpraktijk Schuurman en Franke). bij ZorgIQ en gewerkt als psycholoog bij een slaapkliniek (Chronomed) en huisartsenpraktijk Schuurman en Franke.

174 175 Dankwoord Riemersma-van der Lek, Stefan Knapen, onderzoeksassistentes en stagiaires die in het hoge Noorden hebben bijgedragen aan Bipolar Genetics in zijn geheel Het is een zonnige dag in september 2017 en voor mij ligt het stapeltje papier en mijn actigrafieproject in het bijzonder. dat ik inmiddels met gepaste trots mijn proefschrift mag noemen. Om maar even in het nautische thema te blijven; het was een bijzondere en soms barre tocht, That brings me to people in the US with whom I collaborated during my waarbij regelmatig even alle zeilen bijgezet moesten worden. Gelukkig heb ik bij PhD. First en foremost, Lucia Pagani and Professor Takahashi. Thank you de totstandkoming van dit proefschrift een heel stel mensen om me heen gehad for having me in your lab in Dallas and your invaluable contribution to my die, elk op hun eigen manier, een belangrijke bijdrage hebben geleverd en ik wil actigraphy analyses. And a big shout out to the UCLA-crew: Anil, Loes, ze in dit laatste hoofdstuk heel graag bedanken. Merel, Catherine, Tim, Kim and Yoon. I had an amazing time being part of your group and have been plotting my return ever since I got back. Allereerst Marco. Met eindeloos enthousiasme heb je jarenlang de dagelijkse leiding gegeven aan het ‘bipo-team’. Daarnaast heb ik ontzettend veel van je Daarnaast hebben er ook in Utrecht een aantal mensen heel hard gewerkt aan mogen leren de afgelopen jaren: schrijven, analyseren, niet de moed op te het project. Allereerst Eveline, Ellen, Diandra en Diane, dank voor jullie geven als R de zoveelste error geeft en pittige reviewer-kritieken vol goede bijdrage. Door de jaren heen hebben ook ontelbaar veel stagiaires hun bijdrage moed tegemoet te treden. Dank voor je inzet en je betrokkenheid, het is van geleverd, maar hier zou ik in het bijzonder Anouk, Tetske, Kris, Ditta en grote waarde geweest. Beste René, als hoofdonderzoeker van Bipolar Genetics Laura willen bedanken voor alle inclusies van mijn studie. stimuleerde je ons om tot dit mooie cohort te komen. Ook tijdens mijn schrijfproces kon ik van je leren. Hartelijk dank voor de samenwerking. Beste Dan de rest van het bipo-team. Wat was het een avontuur. Jarenlang hebben Roel, we zagen elkaar slechts sporadisch in Nederland, maar hebben gelukkig we lief en leed gedeeld, sloten koffie gedronken, geroddeld en gelachen en enkele weken zeer intensief kunnen samenwerken tijdens mijn bezoek aan het tussendoor nog de tijd gevonden om al die deelnemers te includeren. Annabel, UCLA-lab. Ontzettend bedankt voor de inspirerende en leerzame tijd. jouw enthousiaste woorden trokken me anno 2011 het onderzoek in. Je bent een bevlogen wetenschapper en staat met veel warmte open voor iedereen om je Hartelijk dank ook aan de leescommissie: Prof. dr. Susan Branje, prof. dr. heen. Lucija, de diesel van de club. Je laat je door werkelijk niets of niemand gek Wiepke Cahn, prof. dr. Floortje Scheepers, prof. dr. Hilleke Hulshoff Pol en maken en ziet altijd nog ergens de humor van in. En ten slotte Annet, ik heb prof. dr. Robert Schoevers voor het lezen en beoordelen van mijn proefschrift. veel geleerd van je nuchterheid, je stressbestendigheid en je positieve houding.

We hebben met het team van Bipolar Genetics het genoegen gehad samen te Na het vetrek van mijn bipo-cluppie heb ik mijn kamer te mogen delen met werken met een lange lijst GGZ-instellingen in Nederland, maar daarbij zou Lotte. Dank voor je gezelligheid en je hulp bij ingewikkelde analyses die ik in het bijzonder het UMC Groningen willen bedanken voor hun waardevolle je achteloos uit je mouw schudde. En na jaren alleen dames om me heen was toevoeging aan ons project. Hartelijk dank aan professor Schoevers, Rixt Marc een heerlijk nuchtere kamergenoot. Onwijs met je gelachen en ooit, in

176 177 een grijs verleden, veel van je boulderkunsten geleerd. Na de toevoeging van Buiten de wetenschap heb ik ook een aantal fijne en belangrijke mensen om me Edwin aan kamer B01.145 was de oude damesdynamiek ver te zoeken maar heen. Sommigen heb ik zo gek gekregen om twee weken met mijn actiwatches de slappe lach voerde hoogtij. Buiten de kamer waren er op de gangen van de rond te lopen, waarvoor uiteraard enorm veel dank: mama, Noor, oma, Esther afdeling nog een heel aantal fantastische lui te vinden, waarmee zeer fanatiek Drent, Janna, Peter, Mattijs, Louise en Viola. werd gevoetbald (nogmaals mijn excuses voor alle blauwe schenen door de jaren heen), geborreld, gedanst in de Chin Chin en, zeker niet onbelangrijk, zeer Daarnaast mijn vaste clubje: Janna, Tessa, Louise, Nynke en Laura. Je succesvol gewintersport. Dank voor de leuke tijd Sonja, Bart, Merel, Judith, waren en zijn er altijd wanneer het tijd is voor een glas wijn en een berg eten, Remmelt, Vincent, Branko, Chantal, Pascal, Wendy en Caitlyn. En een bij voorkeur uit het Midden Oosten. Dank dat jullie al die jaren mijn verhalen speciale vermelding voor Martijn (alias Martino), die altijd op precíes het goede geduldig hebben aangehoord. Lieve Sevinç, je staat altijd voor me klaar met moment langskwam voor een babbeltje en mijn harde schijf uit zeer kritieke eerlijk en oprecht advies. Jeroen, mijn statistiekhulplijn-extraodiniaire en toestand weer tot leven heeft gebracht. koffie maatje, ik ben blij dat je weer in Nederland woont. Dehockeydames , inmiddels mogen we onszelf dames 5 noemen. Met jullie kan ik al jaren met veel Een aantal collega’s heeft het pand al ruimschoots verlaten ten tijde van plezier uitrazen op het veld onder bezielende leiding van Sam. Verder, Team schrijven maar mogen in dit lijstje niet ontbreken. De oude spectrum-garde Groningen, want water is altijd plus 10. Uiteraard mag klimgroep ‘de harde waar ik mijn dagen in het UMCU begon: Remko, Kelly, Willemijn, Florian, boulder’ ook niet ongenoemd blijven want na een dag wetenschap gaat er niks en in het bijzonder Kirstin, die er voor zorgde dat ik na mijn afstuderen mocht boven een avondje Sterk. Rolf, ontzettend bedankt voor je fantastische bijdrage blijven als onderzoeksassistent. Sanne Kemner en Esther Mesman, bipolaire aan de lay-out van mijn boekje en voor Eva’s onvoorwaardelijke knuffels.And matties. Dat er met de rest van de 5BL nog maar vele dinertjes mogen volgen. finally, thanks toBig Blue Diving, for helping me pursue my back-up career in diving. Dan de UH-groep. Iris, ik begon in 2011 bij jou als groen studentje en maakte de cirkel rond in Biddinghuizen. Het was niet voor niets dat ik na mijn En dan de eindeloze dank aan mijn familie die tijdens dit proces gedeeld stageperiode graag wilde blijven hangen in het UMCU en heb met zeer veel hebben in de leuke en minder leuke momenten. Lieve mama, hoe tof was het plezier meegedraaid in jullie Lowlands-project. Ook met de rest van de groep om met jou aan de keukentafel inhoudelijk over mijn onderzoek te praten. En heb ik een enorm leuke tijd beleefd. Sanne Koops, mijn eerste zomer in het dank voor je geduld Jan, als het weer eens over die ongein ging. Maar jullie UMCU, twee brave studenten op een verder verlaten afdeling. Ik heb met beiden ook bedankt voor de momenten dat ik aan de keukentafel veel liever bewondering al je projecten gadegeslagen. Maya, Schuttie, wat een doldwaze over alles behalve mijn onderzoek wilde praten. We spraken het tijdens jullie avonturen met jou. Bij jou is het glas altijd meer dan halfvol. Marieke, Buuf, ik bruiloft al naar elkaar uit, maar ik zou hier graag nogmaals, zwart op wit, willen ben blij dat je slechts 3x struikelen van me vandaan woont. Ten slotte ook de rest benadrukken dat jullie de allerbesten zijn. Mama, het doet me zo goed dit van de UH-groep, Sophie, Kim, Lyliana, Margot, Janna en Willemijn; dank proces af te ronden, wetende dat jij dit proefschrift leest met de trots van een dat jullie me adopteerde na het vertrek van mijn bipo-team. papa en een mama tegelijk.

178 179 En dan Noor. Mijn partner-in-crime wanneer er ergens schaaldieren gegeten dat je verzet en vind het bijzonder zoveel van mezelf in je te herkennen. Jullie moeten worden en al 27 jaar het leukste zusje dat ik me kan wensen. Ik ben zijn het perfecte team om naast me te hebben staan. super trots op je. Ik ken niemand met zo’n groot doorzettingsvermogen en met je brede interesses leer ik altijd wel weer iets nieuws van je. Jon, je bent Tobias, de laatste alinea is voor jou. Tijdens dit proces ben je echt van voor een perfecte toevoeging aan de familie. Gelukkig zet jij het PhD-neuroticisme tot achter mijn onvoorwaardelijke steun en toeverlaat geweest, met een binnen de familie nog even voort. engelengeduld van bovenmenselijke proporties. Ik ben blij en dankbaar dat je er altijd voor me bent, als mijn beste maatje, mijn lievelingsduikbuddy en mijn Viola, technisch gezien niet familie, maar na meer dan 2 decennia vriendschap veilige haven. mag je niet ontbreken in dit rijtje. We delen onze Zeeuwse roots en waren samen tijdens een aantal zeer bijzondere momenten door de jaren heen. Ik kijk uit naar alle mijlpalen die we, al dan niet in glitterkostuum, nog samen gaan aantikken.

Eric en Matthijs, ik heb heel veel respect voor jullie veerkracht die ik de afgelopen jaren heb mogen aanschouwen. Jullie zijn een mooi team en een warme familie.

Lieve oma, dit boekje is niet zomaar aan jou opgedragen. Uiteraard ben je, zonder twijfel, de allerleukste oma. Maar daarnaast ben je ook van grote waarde geweest bij mijn onderwijscarrière, waarvan dit proefschrift het eindstation is. Je hebt er voor gezorgd dat ik daarin mijn doelen hebben kunnen bereiken en daar kan ik je niet genoeg voor bedanken. Maar naast alle praktische ondersteuning vind ik het inspirerend om te zien hoe jij uit werkelijk alles om je heen onbevooroordeeld een les weet te trekken en boven jezelf weet uit te stijgen. Onze gesprekken daarover zijn me heel dierbaar.

Dan zijn daar nog de beste paranimfen, Tessa en Mascha. Tes, met jou is het al heel veel jaren feest. Je voelt me perfect aan en maakt me altijd aan het lachen, ook als ik even door de bomen het bos niet meer zie. En Mascha, wat ben ik blij dat jij je intrek nam in de kamer naast me. Ik heb bewondering voor al het werk

180 181