Normal sleep

Sleep is composed of two phases: rapid eye movement (REM) and non-rapid eye movement (NREM).^{Reference Gelder3,Reference Carley and Farabi4} NREM sleep is further divided into stages N1, N2 and N3, with stage N3 known as slow-wave sleep (SWS). REM sleep is associated with inhibition of peripheral muscle tone, along with active cortical function and increased blood pressure, heart and respiration rates.^{Reference Mindell, Owens and Carskadon1} The switches that result in the cycling between REM and NREM are facilitated by monoaminergic and cholinergic neurons within the brain-stem (see Figs 1–3).^{Reference Hobson, McCarley and Wyzinski5} Sleep is essential in the development of neurosensory, memory and motor systems in the foetus and neonate, and in the maintenance of brain plasticity over the lifetime.^{Reference Penn and Shatz2} The overall regulation of sleep depends on two major processes: the sleep–wake homeostatic process and the circadian timing system. The homeostatic process ensures that the longer a person stays awake, the more pressure there is to fall asleep.^{Reference Czeisler, Duffy, Shanahan, Brown, Mitchell and Rimmer6}

Fig. 1 Hypnogram showing normal sleep in an adult. Non-rapid eye movement sleep is split into stages N1–N3. Courtesy of Dr Kirstie N. Anderson. REM, rapid eye movement.

Fig. 2 Hypnogram of a 9-year-old boy with a short sleep onset latency and shorter sleep cycle duration. Non-rapid eye movement sleep is split into stages N1–N3. Courtesy of Dr Elizabeth A. Hill, Royal Hospital for Sick Children, Edinburgh.

Fig. 3 Hypnogram of a 14-year-old male. Longer (more adult-like) sleep cycle duration (some sleep fragmentation). Non-rapid eye movement sleep is split into stages N1–N3. Courtesy of Dr Elizabeth A. Hill, Royal Hospital for Sick Children, Edinburgh. REM, rapid eye movement.

Sleep during childhood and adolescence

The duration of sleep, the duration of sleep cycles and the circadian rhythm differ between individuals, and vary during a person's lifetime. Newborns sleep most of the day and toddlers sleep for up to 12 h during the night, as well as having one or two naps during the day.^{Reference Bathory and Tomopoulos7} REM and NREM cycles occur every 50–60 min in children under the age of 8 years, and these cycles gradually change to every 90 min as the child develops. Night waking is normal for all children, but is typically very brief, with rapid return to sleep. The proportion of stage N3 SWS is higher and the amplitude of the delta waves is far larger in children compared with adults; this proportion continuously reduces throughout adolescence and into adulthood (see Figs 2 and 3).^{Reference Mindell, Owens and Carskadon1} During the final stages of puberty, there is a biological change in the circadian system whereby children fall asleep later and increase their total sleep time. This often coincides with a desire to stay up late at night to engage in more adult social activities. This delay in circadian phase leads to a delay in sleep onset,^{Reference Dahl and Lewin8} and may cause a lack of synchronicity between adolescent circadian rhythm and that of the parents and of societal structures (such as school timetables). These changes can lead to insufficient sleep, family discord and subsequent emotional difficulties.^{Reference Carskadon9} Sleep deprivation, particularly REM deprivation, in neonates and infants may have a permanent negative effect on development of neural circuitry of the primary sensory system, and of emotion, social learning and memory.^{Reference Penn and Shatz2}

Affective disorders in adolescents

Major depressive disorder (MDD) in adolescence is characterised by a pervasive and persistent low or irritable mood¹⁰ and/or anhedonia, which can also be accompanied by low self-esteem.¹¹ Adolescents with depression may report certain symptoms more often than adults, such as loss of energy, weight and appetite changes, and sleep difficulties, whereas adults report more anhedonia and concentration difficulties.^{Reference Rice, Riglin, Lomax, Souter, Potter and Smith12} Early and accurate detection and management of MDD is important because early onset can be a potent predictor of lifelong recurrent depression.^{Reference Carballo, Muñoz-Lorenzo, Blasco-Fontecilla, Lopez-Castroman, García-Nieto and Dervic13}

Bipolar disorder is a chronic, typically relapsing condition, characterised by episodes of (hypo)mania and depression with a negative effect on overall functioning. In the DSM-IV¹⁴ and DSM-5,¹⁰ it has been separated into type 1, type 2, cyclothymia and not otherwise specified.^{10,14,Reference Nottelmann15} A number of specifiers have been used for bipolar disorder when it occurs in young people, including prepubescent, juvenile, childhood or adolescent bipolar disorder.¹⁰ Bipolar disorder in adolescence is more likely to exhibit rapid cycling and mixed states, and the symptoms are more likely to have a fluctuating intensity and duration, thereby reducing the likelihood of accurate diagnosis compared with adults.¹⁰

Adolescent-onset MDD and bipolar disorder often have a chronic, episodic course, and can be accompanied by long-term functional impairment.^{Reference Fernandez-Mendoza, Calhoun, Vgontzas, Li, Gaines and Liao16} Those affected have higher rates of substance misuse, poor academic attainment, interpersonal difficulties, poor sleep patterns and higher risk of suicide compared with healthy adolescents. These difficulties can negatively affect physical, emotional, cognitive and social development.^{Reference Fernandez-Mendoza, Calhoun, Vgontzas, Li, Gaines and Liao16–Reference Roberts, Roberts and Duong18}

Sleep disorders

The International Classification of Sleep Disorders, Third Edition (ICSD-3)^{Reference Medicine19} defines more than 70 sleep disorders classified into seven major categories, one of which is insomnia.^{Reference Medicine19} The DSM-5¹⁰ describes ten sleep–wake disorders, and the ICD-10¹¹ groups nonorganic sleep disorders into dyssomnias, parasomnias and nonpsychogenic disorders (see Table 1). When considering disordered sleep, it is worth noting that subjectively reported sleep correlates poorly with objective findings.^{Reference Bertocci, Dahl, Williamson, Iosif, Birmaher and Axelson20,Reference Chen, Burley and Gotlib21} For further information on specific sleep disorders, please refer to Table 2.

Table 1 Diagnostic entities in ICSD-3, DSM-5 and ICD-10

ICSD-3, International Classification of Sleep Disorders, Third Edition; NREM, non-rapid eye movement; REM, rapid eye movement.

Table 2 Sleep disorders

ICSD-3, International Classification of Sleep Disorders, Third Edition; CRD, circadian rhythm disorder; OSA, obstructive sleep apnoea; NREM, non-rapid eye movement; REM, rapid eye movement.

Mood disorders and their relationship with sleep

Disturbed sleep is a core feature of a depressive episode,^{Reference Skapinakis, Rai, Anagnostopoulos, Harrison, Araya and Lewis55} may occur in the prodromal period^{Reference Lovato and Gradisar56,Reference Sivertsen, Harvey, Lundervold and Hysing57} and/or be a residual symptom of depression.^{Reference Skapinakis, Rai, Anagnostopoulos, Harrison, Araya and Lewis55} Conversely, mood disturbance is listed as one of the functional impairment criteria for insomnia disorder.

In the general teenage population, insomnia is associated with mental health difficulties later in life.^{Reference Gregory, Caspi, Eley, Moffitt, O'Connor and Poulton58} Disturbed sleep confers an increased risk of subsequent bipolar.^{Reference Faedda, Baldessarini, Glovinsky and Austin59–Reference Ritter, Marx, Bauer, Lepold and Pfennig63} Wakefulness in bed^{Reference Lovato and Gradisar56} and a shorter duration of sleep^{Reference Sivertsen, Harvey, Lundervold and Hysing57} can be regarded as aetiological factors for depression.^{Reference Buysse, Angst, Gamma, Ajdacic, Eich and Rossler64–Reference McGlinchey, Reyes-Portillo, Turner and Mufson67} Sleep difficulties can precede affective episodes^{Reference Faedda, Baldessarini, Glovinsky and Austin59,Reference Ritter, Höfler, Wittchen, Lieb, Bauer and Pfennig61–Reference Ritter, Marx, Bauer, Lepold and Pfennig63,Reference Baglioni, Battagliese, Feige, Spiegelhalder, Nissen and Voderholzer68,Reference Pancheri, Verdolini, Pacchiarotti, Samalin, Delle Chiaie and Biondi69} and other mental health difficulties^{Reference Gregory, Caspi, Eley, Moffitt, O'Connor and Poulton58} by several years in young people. Aspects of the sleep architecture of insomnia, such as reductions in SWS, low spindle activity or changes in REM, are also associated with internalising behaviours.^{Reference Blake, Trinder and Allen70}

In young people who are currently depressed, disturbed sleep incorporating sleep continuity problems,^{Reference Murray, Murphy, Palermo and Clarke17} wakefulness in bed and increased sleep-onset latency is common. Decreased sleep efficiency, short sleep duration, daytime sleepiness, non-restorative sleep, hypersomnia and irregular sleep–wake rhythm are also common.^{Reference Rice, Riglin, Lomax, Souter, Potter and Smith12,Reference Murray, Murphy, Palermo and Clarke17,Reference Bertocci, Dahl, Williamson, Iosif, Birmaher and Axelson20,Reference Steinsbekk, Berg-Nielsen and Wichstrom29,Reference Lovato and Gradisar56,Reference Sivertsen, Harvey, Lundervold and Hysing57,Reference Liu, Buysse, Gentzler, Kiss, Mayer and Kapornai71–Reference Orchard, Gregory, Gradisar and Reynolds75} Sleep difficulties are associated with higher depression severity and recurrence;^{Reference Emslie G, Armitage, Weinberg, Rush, Mayes and Hoffmann76} greater fatigue, suicidal ideation, physical complaints and pain;^{Reference Murray, Murphy, Palermo and Clarke17} and decreased concentration.^{Reference Urrila, Karlsson, Kiviruusu, Pelkonen, Strandholm and Marttunen73,Reference Emslie, Kennard, Mayes, Nakonezny, Zhu and Tao77} There is evidence that sleep quality, total sleep time and sleep microarchitecture can predict new onset and recurrence of depression.^{Reference Armitage, Hoffmann, Emslie, Weinberg, Mayes and Rush78}

Robust adult data and preliminary data in young people suggests that sleep problems can be encountered in all phases of bipolar disorder,^{Reference Salvatore, Ghidini, Zita, Panfilis, Lambertino and Maggini79} can be associated with psychosocial impairment and can play a role in predicting mania and depression.^{Reference Lunsford-Avery, Judd, Axelson and Miklowitz80–Reference Lewis, Gordon-Smith, Forty, Di Florio, Craddock and Jones82} Sleep disturbances among adolescents with bipolar disorder can have an effect on the severity of depressive^{Reference Lunsford-Avery, Judd, Axelson and Miklowitz80,Reference Gershon and Singh81} and manic symptoms, are associated with academic and social impairment during recovery,^{Reference Lunsford-Avery, Judd, Axelson and Miklowitz80} and can be a poor prognostic indicator. Young people with bipolar disorder show a range of sleep difficulties, including delayed sleep onset, insomnia, hypersomnia and difficulties waking up in the morning. They can also present with difficulties maintaining a regular rhythm, longer naps during the day, more night-time wakings, a greater total time awake during the night and nightmare disorder.^{Reference Lunsford-Avery, Judd, Axelson and Miklowitz80,Reference Stanley, Hom, Luby, Joshi, Wagner and Emslie83–Reference Diler, Goldstein, Hafeman, Merranko, Liao and Goldstein86} Nightmares appear to augment the risk of suicide, and insufficient sleep duration has been associated with self-injurious behaviours.^{Reference Roybal, Chang, Chen, Howe, Gotlib and Singh84} Mania has been associated with decreased need for sleep in young people;^{Reference Harvey, Talbot and Gershon87} however, in adolescents, mania also appears to be linked to a broader sleep disturbance,^{Reference Faedda, Baldessarini, Glovinsky and Austin59} including variable sleep duration and unstable morning routines.^{Reference Lunsford-Avery, Judd, Axelson and Miklowitz80} Sleep disturbance does not appear to differentiate between types of bipolar disorders, being as common, for instance, in bipolar disorder not otherwise specified as in bipolar disorder type 1.^{Reference Baroni, Hernandez, Grant and Faedda88} Disturbed sleep is seen more often in the children of parents with bipolar disorder than in healthy controls.^{Reference Ritter, Höfler, Wittchen, Lieb, Bauer and Pfennig61,Reference Sebela, Novak, Kemlink and Goetz89} One study of children whose parents were diagnosed with bipolar disorder showed shorter time to onset of sleep and longer duration on actigraphy, with opposite subjectively reported experiences.^{Reference Jones, Tai, Evershed, Knowles and Bentall90} This is in accordance with a second study showing that people deemed to be at high risk of developing bipolar disorder, either through family history of bipolar disorder or personal history of severe depression and subthreshold mania symptoms, have shown a longer duration for sleep onset, but also higher levels of sleep specific worries.^{Reference Ritter, Marx, Lewtschenko, Pfeiffer, Leopold and Bauer85} Depressed youth with bipolar disorder are significantly more likely to be affected by daytime sleepiness and hypersomnia than young people with unipolar depression.^{Reference Diler, Goldstein, Hafeman, Merranko, Liao and Goldstein86}

Mechanism of the association between mood disorders and sleep

Maladaptive thinking processes, such as cognitive inflexibility, attention bias by selectively focusing on negative information, misperception of sleep deficit, rumination and worry, may underlie insomnia, depression and bipolar disorder in adolescence.^{Reference Blake, Trinder and Allen70} These processes may trigger autonomic arousal and the reinforcement of negative cognitions. Similarly, abnormalities of circadian rhythm, perhaps mediated by genetic vulnerability via polymorphisms of serotonin, dopamine and circadian clock genes^{Reference Ankers and Jones60,Reference Mansour, Monk and Nimgaonkar91} or dysfunction in the white matter integrity, could also play a role in the aetiology of these conditions.^{Reference Bollettini, Melloni, Aggio, Poletti, Lorenzi and Pirovano92} Disruption of the corticolimbic circuitry^{Reference Blake, Trinder and Allen70} may be a consequence of insomnia, and may impair affective reactivity and regulation.^{Reference Dahl93} Dysregulation of reward/approach related brain function,^{Reference Blake, Trinder and Allen70,Reference Casement, Keenan, Hipwell, Guyer and Forbes94} hypothalamic-pituitary-adrenal axis dysregulation and elevated inflammatory cytokines may also contribute to both psychopathology and sleep disturbance.^{Reference Blake, Trinder and Allen70} Disruptions to the circadian rhythm have been postulated to contribute to the pathophysiology of bipolar disorder.^{Reference Bradley, Webb-Mitchell, Hazu, Slater, Middleton and Gallagher95} Genetic commonality between sleep and impulsivity and anger/frustration has also been described.^{Reference Miadich, Shrewsbury, Doane, Davis, Clifford and Lemery-Chalfant96} Antidepressants, particularly those with noradrenergic or dopaminergic mechanisms, may induce or aggravate insomnia. Electronic media use among adolescents, particularly at night-time, may be related to sleep disturbance and higher levels of depressive symptoms,^{Reference Lemola, Perkinson-Gloor, Brand, Dewald-Kaufmann and Grob97} in which a sleep debt developed over school days is paid back at the weekend.^{Reference Wittmann, Dinich, Merrow and Roenneberg98}

Interventions for mood disorders and sleep problems

Research direction

There are a number of challenges in the field of sleep research in MDD and bipolar disorder. First, most studies are cross-sectional, with an unclear timeline between the onset of the affective disorder, that of sleep difficulties and the short- and long-term complications.^{Reference Stanley, Hom, Luby, Joshi, Wagner and Emslie83} There is still limited objective sleep assessment with polysomnography, and many studies report subjective sleep via self-/parent report or actigraphy without, for example, screening for specific sleep disorders or using polysomnography. The parent–child agreement for psychiatric disorders is low and, wherever possible, the children and young people should be consulted about their sleep directly; depression and bipolar disorder are no exception.^{Reference Roybal, Chang, Chen, Howe, Gotlib and Singh84,Reference Achenbach, McConaughy and Howell126} Terms such as sleep disorders and sleep difficulties are used interchangeably, and there is a need for greater precision with the use of, for example, insomnia disorder or CRD. Additionally, there seems to be a very high reliance on assessment questionnaires for diagnosing sleep problems and sometimes even affective disorders. Over the years, there have been inconsistent findings in the studies for depression in children and young people,^{Reference Ivanenko, Crabtree and Gozal127} as study designs vary a great deal.

The number of patients included in trials for bipolar disorder tends to be relatively low, which might be linked to the relatively low prevalence in children and adolescents and the subsequent recruitment challenge. Some studies seem to focus on one type of bipolar disorder and do not include the full spectrum of disorders, which might leave out important information.^{Reference Baroni, Hernandez, Grant and Faedda88} Findings vary across studies because of differences not only in the sleep measurement, but also sampling; some studies focus on children who have been diagnosed with bipolar disorder, whereas others include children whose parents have an affective disorder.^{Reference Gershon and Singh81} There is also a high variability of the age range, as there is no consensus on the age of onset of paediatric bipolar disorder; some papers have included participants of 5 years of age,^{Reference Baroni, Hernandez, Grant and Faedda88} whereas others have focused on prepubescents^{Reference Lofthouse, Fristad, Splaingard and Kelleher128} or adolescents.^{Reference Lunsford-Avery, Judd, Axelson and Miklowitz80} One has to consider the importance of brain maturation and expected sleep changes that occur in these stages of life, as well as the beliefs around sleep within the family.

An important direction for future research would be developing a robust study design that allows a longer-term evaluation of the symptoms of a bigger cohort, objective and subjective standardised measures of sleep and mood, screening for comorbid sleep disorders, collecting collateral information and focusing on behavioural interventions for sleep, such as CBT-I.