Why sleep matters for young people who may get depressed

Abstract

Depression and anxiety are negative emotional states familiar to us all through personal experience. Less familiar are severe states of depression, in particular, which can actually shorten the lives of sufferers by over a decade. The relationship of these very severe states of illness to the milder cases more common earlier in development is important. Most patients who have suffered from depression will suffer from further episodes during their lifetime, and an early onset may make recurrence more likely. A number of factors increase the risk for depression, including family history, stressful life events, early life experiences, personality (particularly the traits of neuroticism and perfectionism) and mood lability (marked ups and downs). Sleep disturbance may both provoke and/or signal the onset of mood disorder. Sleep is therefore doubly important as a gateway to treatment. Understanding more about how sleep interacts with the established risk factors would allow vulnerable young people to be identified earlier for more effective intervention. Early identification of sleep disorder and depression allows psychological treatments to be used, which are less effective once a full depressive episode and a cascade of neurobiological and psychological effects have occurred.

1. Introduction

Depression is an everyday word. In his memoir, Darkness Visible, the novelist William Styron described it disparagingly as having ‘slithered innocuously through our language like a slug, leaving little trace of its intrinsic malevolence, and preventing by its very insipidity, a general awareness of the horrible intensity of the illness when it is out of control’. Styron himself experienced severe depression, which resulted in his own hospitalization [1]. Men and women who have depression severe enough to be referred to secondary care have a remarkably reduced life expectancy. From the age of 15, men will lose 14 years, women 12 years. This excess mortality is partly due to suicide but also to physical illness, particularly cardiovascular disease [2]. These stark numbers illustrate why it seems reasonable to call depression in its severest forms a disease and also how bad we still are at preventing its worst outcome. If depression is a disease, then it follows that a medical approach may help us understand what it is and how we might reduce its impact. That means to adopt a methodology that is scientific and has worked for other conditions like epilepsy, whose essence in the past may have seemed more of the spirit than the body.

Depression occurs with different severities and different degrees of persistence. To capture these properties requires measurement. The most obvious approach to such measurement is to describe what we see in clinical practice and what patients tell us. The picture that emerges is that depression is characterized by experiences, which we choose to call symptoms. The analogy with physical illness seems reasonable because the experiences are unpleasant, somewhat stereotyped and persist over time. They resolve and disappear when patients recover, as they usually do, and their mood and cognition become what is normal for them. These characteristics are the justification for treating patient experience as an object for study. This approach does not seek to understand the meaning of the patients' symptoms. So, we do not offer an immediate interpretation of why a patient may feel a particular way; if we did, it is very likely that we would find many different potential explanations for the same symptom. Worse still, we would probably find that different explanations would be offered for the same person's symptom by different observers. The reason we concentrate on what is usually called phenomenology is that it is the price we pay for a common framework, which is reliable and allows us to make explanatory conjectures. I would argue that it is what makes a science of mental illness possible, and it is therefore fallible in the sense that we must frame hypotheses that are falsifiable. There are other traditions that see mental illness quite differently and are currently dominated by postmodern theories of oppression, which I will not address because they seem to me to lead to pointless nihilism.

Depression is accordingly conceptualized by psychiatrists as a cluster of particular symptoms that persist together over an interval of time. ‘Major depression’ is defined as five or more of the symptoms shown in table 1. They must have persisted for at least two weeks. This approach is now adopted in both the Diagnostic and Statistical Manual (DSM-IV) of the American Psychiatric Association and the International Classification of Diseases (ICD) of the World Health Organization. The use of symptoms as criteria in this way dates from 1980. Prior to that, descriptions of psychiatric disorders were based on interpretation. Thus, in DSM-II, depressive symptoms would have been described as manifestations of neurosis. In other words, depression was a defence mechanism against ‘anxiety’ and its mechanism would have been explored through a psychotherapy derived from psychoanalysis. Adopting a classification based on symptoms allowed reliable research into patient populations and the impact of treatments of all kinds.

Table 1.

DSM-IV criteria for a major depressive episode. There must be depressed mood or a loss of interest or pleasure in daily activities for more than two weeks. Specific symptoms, at least five of these nine, must be present nearly every day. There is impaired function: social, occupational and educational. All compared with a normal baseline and present for at least two weeks.

(1) depressed mood

(2) markedly diminished interest or pleasure

(3) significant weight loss

(4) insomnia or hypersomnia

(5) psychomotor agitation or retardation

(6) fatigue or loss of energy

(7) worthlessness or excessive/inappropriate guilt

(8) diminished ability to think or concentrate, or indecisiveness

(9) recurrent thoughts of death, suicidal ideation

The most striking finding from such research in the general population is that major depression as defined is quite common and its onset is most often in the age group 15–25 years [3]. In our own survey of students in Oxford and Stanford, 6% described a depressive episode at or before the age of 19 years [4]. Many of these episodes resolve without treatment, but some do not. Indeed, some will be the precursor to the more severe illnesses that psychiatrists so often find difficult to treat. Something like a consensus has emerged in formal consultations by the Medical Research Council (https://mrc.ukri.org/documents/pdf/mrc-mental-health-research-report-2010/) and the Wellcome Trust that depression in young people should be prioritized in the allocation of research funding. The Wellcome Trust have recently initiated, as a Mental Health Priority Area, a 5- to 10-year project to understand and improve psychological approaches for treating young people with mood and anxiety disorder.

Our understanding of the causes of major depression is reasonably well established at a phenomenological level. That is to say there are clear associations between an episode of depression and identifiable antecedent risk factors. The predisposing factors include a family history of serious depression, an anxious temperament and early abuse and/or neglect. Family history and temperament are significantly determined by genes. Most mental illness and many of our most robustly measured traits are to some degree heritable on the basis of twin studies. Such findings are replicable and consistent but inconvenient to some schools of contemporary academic thought. Blank slateism––the insistence that the mind has no innate traits––is most famously associated with ideas about general intelligence [5] but there are many echoes of it in criticisms of psychiatry. The usual approach to deconstructing such data is to deny the validity of the measured variable, claim that twin studies are ‘fatally flawed’ and, if all else fails, impugn the motives of the investigator. More respectably, correlational studies are all threatened by the risk that we fail to understand the causal direction or the origin of the relationship. But, if we accept the most parsimonious explanation of the current literature, major depression, anxiety disorders and associated anxious temperament are all moderately heritable and can often be observed to run in stable non-abusive families in consequence.

The argument for heritability has shifted with the development of methods for detecting sequences of DNA in the human genome at scale. An association between a psychological measure and variation in the structure of a gene can have only one plausible direction of effect. A psychological measure could never cause a genetic variation. However, it has required very large samples for the small effects of individual genes to be teased out. Consequently, we remain uncertain about the mechanisms involved in translating these genomic differences into measureable brain mechanisms. Nonetheless, most scientists working in psychiatric genetics are now optimistic that something material will emerge from current studies [6]. Very recently, for example, it was shown in a very large population sample that depression is associated with rare but obvious DNA sequence errors in individual genes known to play a part in neurodevelopment [7]. This moves us closer to understanding mechanisms than previous work that has studied small, common variations in genes of very small effect. Implicating genes with a known biology and potentially large effects on brain function could be game changing.

The impact of temperament is worth emphasizing and will be further illustrated below. The tendency to anxious worrying is often described as neuroticism and can be estimated on a simple scale, first developed by Hans Eysenck. This poses 12 simple yes/no questions like ‘do you often need kind friends to cheer you up’ and ‘are your feelings easily hurt?’. Women on average have higher scores on this scale than men. This observation means that women with very high scores will significantly out-number men. This partly explains the consistent observation that the risk of depression is about twice as high in women compared with men [8].

The precipitation of an actual episode of depression in a vulnerable individual usually follows an adverse life event, characterized by a significant loss, disappointment or humiliation. Life events that are dangerous or threatening are more likely to precipitate anxiety [8]. There is reasonably good evidence that such events correlate with personality. In other words, the effects of one's genetic endowment may be amplified by choices and preferences made during one's development [9]. However, they may also obviously be a result of misfortune. The impact of life events on the risk of a depressive episode is cumulative and depends very strongly on neuroticism [10]. This is illustrated in figure 1. Notice the dramatic difference in risk mediated by neuroticism: low neuroticism means that one is quite resilient in the face of severe life events, while high neuroticism predicts high rates of major depression when exposed to a notionally similar stress. Men and women appear to behave very similarly in this regard (figure 1).

Figure 1.

Hazard ratios indicating risk of onset of major depression for a population-based sample (N= 7517) classified by sex, neuroticism and stressful life events (adapted from [8]).

Any individual with depression is very likely to seek an explanation: why me? It may not be very satisfactory, but the best description we can give is that it will be an amalgam of genetic predisposition and environmental adversity. When the usual adversity is absent and patients have grown up happily in a united family, one may speculate that genetic factors will have been more important. Where a person has experienced considerable adversity, depression is more likely, whether or not they have a strong genetic predisposition, but some such predisposition is quite likely. The metaphor that is sometimes used is that genetic and environmental risk factors are like marbles filling a container of fixed volume. When it is full, there is a state of depression. Current treatment can be thought of as adding volume to the jar so that the marbles no longer fill it. Cure would mean to take the marbles out of the jar entirely. We do not know how to do that. To be able to do so would require a more fundamental understanding of the brain mechanisms that underlie the appearance of a depressive episode.

The brain systems which are the most likely candidates for the substrate of depression are those concerned with the regulation of stress. Stress was originally conceived as a physiological mechanism [11]. Extremes of environmental stress or infection have adverse effects on the body and there are a series of homeostatic mechanisms that seem to react to preserve vital organs. The steroid hormones and monoamines secreted by glands situated above the kidneys (adrenal glands) are necessary for a normal stress response and are usually regarded as prime movers in integrating the response. In addition, the immune system can be seen as orchestrating the response to infection. Anyone who has had a flu-like illness can testify to the major impact infection can have on cognition and motivation. Hence there is no doubt that stress can have a direct impact on the function of the brain including effects on mood. In recent years, infection and inflammation have emerged as possible contributors to the risk of depression. It may eventually provide an alternative way of treating depression or even preventing depression, but currently the evidence is just not there.

It has also long seemed possible that there may be an overlap between these physiological effects and the consequences of purely psychological stress. In animals, exposure to early stress (such as the removal of cubs from their dams) results in abnormal stress regulation in adulthood and part of the vulnerability to this effect is genetically determined because some strains are resistant to this treatment [12]. The mechanism in rodents has been quite well characterized at a molecular level, but it does not translate easily, if at all, to the human case.

In depression, there appears to be hypersecretion of the glucocorticoid hormones in a sub-group of patients, and this is associated with reduced tissue sensitivity to the hormone's effects. There are dense expressions of glucocorticoid receptors in the hippocampus and hypothalamus. This finding has long appeared to be a major clue to how depression arises in the brain, but it has proved very difficult to map the mechanism onto patient phenotypes in a way that either explains what we see clinically or provides a means for novel treatment [13].

2. Sleep and depression

Table 1 illustrates that sleep is usually disturbed in major depression. Sleep may also be relevant to stress regulation and infection. It is still difficult to give a fully convincing account of why sleep is important for humans, but it is simply a fact that all the major organs, and particularly the brain, have a very striking diurnal clock that regulates physiological states of all kinds [14]. At a purely subjective level, it is plausible from ordinary experience that emotional distress or worry that builds up over a day can be banished by the proverbial good night's sleep. Poor or unrefreshing sleep could contribute to the persistence of a negative state. There is also the possibility that memory requires adequate sleep.

According to one influential hypothesis, sleep serves to normalize synaptic strength in the brain. The idea is that during wakefulness synaptic strength increases, because we are continually learning new associations, and the function of sleep is to reset the system. This is the synaptic homeostasis hypothesis [15]. It is supported by molecular and electrophysiological studies of the brain (refer to Foster [14]). According to this view, sleep need reflects a plastic change in synapses during the awake stage and slow wave sleep is important for the restoration of brain physiology. Poor sleep could then impair the restoration required following a period of being awake in a depressed state. How this occurs remains uncertain. The brain consumes 25% of all our energy intake (as much as 35% in children). Energy exchange produces potentially dangerous chemical by-products that must be continuously mopped up and moved out of brain tissue. This is probably a cumulative problem in the awake state. One novel idea invokes the glymphatic system, which is the arrangement of fluid-filled spaces making up the interstitial space around the small blood vessels that serve the brain. Astroglial cells operate as house cleaners for toxins produced in these perivascular structures which increase in volume substantially during sleep [16]. Whatever the precise mechanism, slow wave sleep is reduced in depression, so this reduction may contribute to maintaining the depressed state.

It is an important additional idea that poor sleep may be antecedent to depression and contribute to its onset. Insomnia appears often to be present before depression in individual cases and may simultaneously reflect the impact of an environmental stimulus and be the mediating cause for a state of depression to become enduring and troublesome [17]. It is possible that the energetic balance described previously is particularly vulnerable to disturbance in people at risk for depression. If correct, this idea places poor sleep in a pivotal position at the onset of new and notably first episodes of depression. This is of interest because poor sleep may be obvious to a person before a sustained low mood is noticed. Moreover, poor sleep is a neutral symptom: it seems likely that a patient will acknowledge that their sleep is impaired long before they are ready to accept the potentially more stigmatizing judgement that their mood has become low and they need a psychiatric intervention. In addition, adherence to treatment is likely to be much better before depression has taken its toll on an individual's cognitive reserve and motivation.

The application of this principle to young people has the most obvious appeal. It offers a way in which depression might be prevented, if insomnia is treatable to some degree. It turns out that behavioural interventions are quite effective at improving sleep duration and quality. This has been shown in a reasonably sized randomized clinical trial (RCT) which compared a cognitive behavioural therapy (CBT) with a dummy therapy which made no attempt to modify sleep habits (and a waiting list control) [18]. The results favoured the CBT approach and argue for some specificity in its action. Crucially, it was trialled and has subsequently been made available through a digital platform and app (Sleepio). Users of the app receive a series of instructions and suggestions from an animated cartoon presentation. The focus is on beliefs about sleep, the daily schedule, lifestyle and how the bed and bedroom are used. Essentially it takes the form of a course of education about sleep and a set of behavioural instructions designed to facilitate good sleep hygiene. The experience of the user can be shared with others and the app is interactive in other ways that allow self-monitoring of sleep outcomes. Access to such treatment can therefore be made both easy and inexpensive.

We recently completed a study of students with insomnia who were able to access this Internet-based treatment [19]. It is currently the largest RCT conducted in the mental health field. We randomly assigned 3755 participants to receive digital CBT for insomnia (n = 1891) or usual practice (n = 1864). Compared with usual practice, the sleep intervention at 10 weeks reduced insomnia, paranoia and hallucinations. The reduction in insomnia was a mediator of change in paranoia and hallucinations. Although these were the primary outcomes, and of particular interest as measures of psychotic experience, the effects on other symptoms were also striking. Measures of both anxiety and depression were reduced in the treated group. The findings fit with the idea developed above that sleep itself is a key mediator of psychiatric symptoms and, if it can be improved, so also will these symptoms. How far a relatively brief intervention like this can influence long-term outcomes remains a key question. The short-term benefits are certainly encouraging.

3. Conclusion

Depression is common in young people. It may be self-limiting and require rather little in the way of intervention. For others, depression in adolescence may be the harbinger of more severe mental illness and even of premature death. Our understanding of depression across this spectrum of severity is informed by a growing confidence that genetics and neuroscience can help explain why individuals are vulnerable to depression in the first place and what mechanisms precipitate the onset and maintenance of low mood. For the moment, our understanding rests on generalizations based on provisional measurements of symptoms. This has allowed the identification of sleep disturbance as a potential mediator of the transition from well to depressed and raises the important possibility that its treatment may play a part in preventing depression in the first place. The role of sleep is poorly understood in the general population. As Russell Foster also argues in this issue, better sleep hygiene, as it is sometimes called, does not yet receive the attention from public health policy makers that it probably should. Parents need to ensure that home environments are conducive to good sleep habits. What is required when insomnia is diagnosed is improved access to sleep interventions for young people and careful measurement of long-term outcomes. Other things being equal, better sleep should mean better mental health.

Data accessibility

This article has no additional data.

Competing interests

We declare we have no competing interests.

Funding

We received no funding for this study.