Depression in Children and Adolescents: Current Thoughts by Mike Connor

Depression in Children and Adolescents: Current Thoughts

The first part of these notes explores whether there has been sharp rise in the prevalence of depression among young people in recent years; with the conclusion that, while prevalence rates are high, they may reflect changing diagnostic procedures or heightened awareness of a disorder previously under-diagnosed.

There follows a description of risk factors for depression in the pre-school years with reference to the impact of early stressful events.

Gender differences in the proneness to depression or in the aetiological routes are discussed.

The final section concerns the links between depression in adolescents and both attention difficulties and a proneness to ruminate on negative experiences.

M.J.Connor January 2007

The Prevalence of Depression in Young People

Costello et al begin their epidemiological survey by citing an apparent belief, within some recent research papers (eg Kessler et al 2001) and in the popular media or forums for debate, that recently recorded rates of depression among children and adolescents represent an “epidemic”.

The authors describe four sources of this belief:

The evidence for observed increases in anti-depressant prescriptions for children and young people.
Retrospective recall of by successive birth-cohorts of adults of life-time experience of depression.
Rising suicide rates among adolescents (according to data from the USA).
Epidemiological data indicating an ongoing increase in emotional problems (anxiety and depression) among 15-16 year olds cited in 1974, 1986, and 1999.

However, other available research evidence does not indicate this kind of cohort effect; and a counter argument has arisen that there has been no significant increase the actual prevalence rates, or that any apparent increase may be explicable in terms of some artefact associated with methodology.

Patten (2003) represents this school of thought in suggesting that it is dangerous to rely upon retrospective reports of childhood depression, and even a small degree of recall bias could give the impression of increasing rates of depressive disorders.

Accordingly, the current authors argue that one needs to look at current psychiatric data provided by participants of similar ages and representing successive birth cohorts.

They use prevalence estimates from the information collected since the 1970s from representative samples of children and adolescents by means of structured interviews and standard statistical procedures. The data relating to individuals born between 1965 and 1996 are used to test the belief that children and adolescents now are more likely to experience depression disorders than earlier generations of children and adolescents.

From the various data bases available, the authors selected their samples on the basis of age up to 18; formal psychiatric diagnoses of depressive disorders based upon valid diagnostic criteria and assessment/interview processes; access to information about dates of birth, gender, and age at the time of diagnosis; and access to information about the time frame of the interview used.

To avoid the confounding factors related to age and gender (given consistent evidence that prevalence of depression is low among boys and girls before puberty, but that the rates increase after that time, especially among girls), the authors completed separate analyses for participants below 13 years of age and those of 13+; and, for the older age group, there were separate analyses for boys and girls.

In their summary of the general outcomes, Costello et al refer to a meta-analysis of studies which involved more than 60,000 observations of children born over the past 30 years. They provided no evidence of increasing rates of depression among the later-born cohorts.

The evidence supported the prediction that the rates of depression would be higher among adolescents than among children, and higher among adolescent girls than among adolescent boys.

Also, interviews conducted across a longer time frame produced greater prevalence estimates than those using a shorter time frame …. in particular, recall for events was shown to become unreliable after about 6 months.

The data indicated a lower prevalence rate than the estimates generated by some current surveys and reviews of adult depression; and the sex difference for adolescents, while still significant, were smaller that some current reviews have suggested (perhaps reflecting the differences in sample sizes).

Nevertheless, the belief in the epidemic proportions of depression is of such a strength that the authors feel it necessary to seek some underlying reasons for it.

Initially, they cite the pattern over the years of increasingly earlier puberty and the association that has been made between pubertal factors and the enhanced risk of depression (in girls). However, closer analysis reveals that the most significant fall in the age of puberty occurred before the 1960s, and any further fall since that time is of little significance. (Mean age of menarche fell by only 2-3 months between the 1960s and the early 1990s.)

Secondly, they follow up the issue of the increase in suicide rates cited in several countries between 1970 and 1990 (with declining rates since that time). They argue that many young people who commit suicide will have experienced a depressive disorder (with estimates ranging from 49 to 64 %). However, although depression is more likely among suicidal girls than boys, nearly all the increase in suicides among young people between the 1960s and the 1990s is reported to have involved males.

(Further, the statistics concerning drug abuse are said to have followed a similar pattern to those of suicide, and a plausible hypothesis to challenge the link between depression and suicide is a link between drug abuse and suicide.)

Thirdly, the authors hold that one should be cautious about forming conclusions based upon clinical samples of children and young people since such individuals may differ in a number of ways from the majority of un-treated or un-referred cases of some psychiatric disorder. The probability is cited, for example, that the children who do receive treatment may well come from families with a history of depression with parents attuned to the signs and symptoms of depression.

Finally, the rates of childhood depression might appear to have increased because of changes in the diagnostic criteria which enable the inclusion of more people of younger ages … (although it is acknowledged that this circumstance could have occurred alongside some actual increase in prevalence.)

The authors also acknowledge limitations in their study in terms of an absence of data before 1970, so that there is always the possibility that there was a real increase in depression in childhood and adolescence before that time.

They also acknowledge that prevalence rates may be less accurate than incidence rates and more prone to errors, but the published incidence data are too few to allow for valid conclusions.

Nevertheless, they feel able to argue that, while there is good reason to be concerned about the rate of depression among these age groups, the data emerging from 30 years of research suggest that there has been little change in the proportions of children and adolescents who have been diagnosed with depressive disorders.

It may be the case that more children and young people with depression are being recognised and formally identified, but this is more likely to reflect a greater sensitivity to a long-standing situation rather than to an epidemic of depression over recent years.

Risk Factors for Depression in the Preschool Years

In the introduction to their survey concerning children up to the age of 5+, Luby et al are able to cite converging evidence that a family history of mood disorders and experience of stressful life events are key risk factors in the development of depressive disorders among older children and adults.

Further, there are a number of studies (such as Jaffee et al 2002) whose results have suggested that the strength of the association between risk factors and emotional disorders will vary according to the age of the children.

Depression is linked with genetic influences, with recent evidence (Caspi et al 2003) indicating that the genetic risk may be the more realised when there is exposure to maltreatment or to stressful circumstances in childhood.

The authors are also able to cite studies involving primates which have shown that exposure to stress early in development is linked to increased rates of depressive disorders observed later in development and in adulthood.

The study by Luby et al themselves used a form of regression (correlation) analysis to examine the relative significance of a family history of mood disorders or other forms of depressive disorder and of stressful life events as risk factors for early onset depression among children between the ages of 3 and 5+ years who were monitored over a 6-month period.

It was held that, if stressful events could be identified as a mediator of risk for early onset depression, there would be a means of pinpointing those children (and their families) who would be the prime targets for early intervention and preventive strategies.

The participants were 174 pairs of children and their mothers identified from referrals to community paediatric clinics or specialist mental health clinics, and by the use of a diagnostic checklist on which the preschool children were found to have 3 or more depressive symptoms or 3 or more disruptive albeit non-specific symptoms. Children with a known developmental or neurological disorder, or measured cognitive ability 2SD or more below the mean, were not included.

The sample on which there were complete data from base-line assessments involved 149 cases (80 girls and 69 boys); but 30 of these did not attend follow up assessments or had incomplete data, and were not included in the final analysis.

The children were grouped into two sub-samples … those displaying symptoms of depression, and those showing disruptive symptoms of an ADHD kind. A control group of children with no observable symptoms was also monitored.

Measures included the severity of depressive symptoms determined by structured parental interview about the children’s moods and behaviours (The Diagnostic Interview Schedule for Children), and by classifying symptoms against the DSM-IV criteria for depressive disorder.

Also assessed was the family history of mood disorder using a structured interview

(The Family Interview Schedule); and the history of stress in the lives of the children using the Coddington Life Events Schedule. Such events could cover a range of severities including the birth of a sibling, problems with social interaction among peers, and the loss of a parent.

The results provided clear support for the hypothesis that a higher family history of mood disorder (such as depression, bipolar disorder, and/or suicidal behaviour) was significantly associated with higher depression severity scores in the children at the time of the follow-up assessment.

Stressful life events were shown to be significant predictors of pre-schoolers’ depression severity at the time of follow up.

There was also found to be a significant interaction effect between the family history of mood disorders and stressful life events which occurred during the preceding year.

The latter “mediational” relationship was demonstrable by the fulfilment of certain conditions such as the preschoolers’ experience of stressful events following the mood disorders impacting upon other family members; and the experience of the stressful events occurring before the assessment of the severity of depression; and the reduction in the overall relationship between a family history of mood disorder and the children’s severity of depression when the effects of stressful life events are controlled.

In their general summary, the authors restate their finding that early experience of stressful life events mediates the relationship between a family history of mood disorder and the children’s level of depression observable over time.

(One might put this another way, presumably, by arguing that certain children have an enhanced [genetic] propensity for depression, and that the experience of certain stressful events will act upon any such propensity to bring about a significant level of depression in the children themselves.)

In any event, the authors emphasise the potentially powerful influence of psychosocial stressors in the risk for early onset depression. The effect is specific to depression and is not linked with more general psychopathology.

However, what matters is that these stressors may well be amenable to therapeutic intervention and prevention.

In conclusion, the authors highlight the role of stressful events in providing the mechanism by which the risk factor associated with a family history of mood disorder may be realised (and the level of any subsequent depression in the child determined).

The implication lies in the ability to determine those preschoolers who have an enhanced risk of depressive disorder and to work with the families both to reduce the probability of stressing events and to enhance the resilience of the children in coping with the stressful circumstances which do arise.

(There is also the reminder conveyed in these notes that even very young children can be subject to depression.)

Gender, Interaction, and (Mild) Depression

It has long been recognised that, in general, males and females tend to display their reaction to stressing or distressing or arousing circumstances in different ways. The male tendency is towards externalising behaviour while the female reaction is more likely to involve internalising behaviour.

(This may go a considerable way towards explaining gender differences in respect of the rates of referral for therapeutic intervention. Boys’ maladaptive behaviour could well involve acting out and some form of disruptiveness, in contrast to the tendency among the girls towards anxiety or some reaction that does not impinge upon anybody else.)

These thoughts provide an introduction to the work of van Beek et al (2006) who note that there is a steady increase in depressive symptoms from middle childhood into young adulthood, with mid-adolescence a time of particularly high prevalence levels of depression.

While there may be physiological and cognitive factors at work, these authors have focused upon the role of social skill deficits in underlying this circumstance; and they cite a number of studies which have linked depression to problems with social relationships and a lack of popularity.

(Relevant here is the content of the recent paper “ Attention Deficit Disorder : Research Update 46 ” [MJC January 2007] which includes a description of the way in which the behavioural symptoms of ADHD can interfere with social interactions and lead to some rejection on the part of peers. This in itself, or when combined with a recognition of the negative image held by others and with a growing negative self image, can lead to a reaction in the form of depression.)

The social deficit model, as set out by a number of authors (eg Segrin 2000), suggests that inadequate social skills may evoke less positive responses from significant others, which, in turn, increase the individual’s own feelings of loneliness and have a negative effect upon self image. Depression may result.

Meanwhile, one effect of depression itself is to reduce social skills thus reinforcing the negative reactions of other people.

The current authors appreciate the plausibility of this perceived pattern, but note the lack of empirical evidence concerning the nature of social skills in adolescence and how this relates to individual differences in the development of depression.

Accordingly, their own study involved samples of mildly depressed and non-depressed young people (pre-adolescent) identified from an initially large sample of Dutch students attending schools in several towns in the central part of the Netherlands. They were drawn from the top end of primary schooling (mean age 10+) or from the first or second year of secondary schooling (mean age 13+).

The particular interest was in the pattern of depressive disorders among girls and boys, with evidence noted that there seem to be more cases of boys with depression in middle childhood, but more girls than boys with depression from mid-adolescence … and this latter difference is maintained into adulthood.

The question was raised whether there are some gender or age specific factors contributing to the onset of depression.

Non-verbal behaviours were hypothesised to be particularly significant given the (general) male-female differences in non-behavioural style.

Social rules dictate that women ought to present a more attentive and empathic style than men, as reflected in the focusing of gaze while listening to indicate positive interest in the other person. Men are more likely to focus their gaze upon the other person while they, themselves, are speaking thus to convey their own confidence and their capacity to dominate an exchange.

Further, converging evidence indicates that women typically show more “back-channeling” behaviour while listening to another person, such as nodding the head, smiling, and using positive expressions to mask any negative reactions.

These skills, and the gender differentiation, become established during middle childhood with children becoming aware of these “display” rules and learning how behaviour is to be adjusted.

Even from infancy, gaze focusing is more common among females both when speaking and listening. Males become more visually dominant over time. From around 8 years of age, backchannel behaviours become evident and girls use these (empathic) actions more than boys do.

The authors go on, however, to cite existing evidence that depressed adults, females and males, do not show this orientation towards other people. Rather, they gaze less and display fewer smiles or nods.

One hypothesis has it that the higher incidence of depression among adolescent girls is a matter of the greater consequence for girls, whose friendships are usually characterised by higher levels of intimacy than are observed in boys’ friendships, of a failure to comply with these social rules.

Flouting of the rules by girls would be considered more of a deviance than the same behaviour observed in boys, with the girls in question likely to receive more negative reactions from their communication partners.

There is the further question concerning whether the behaviour of the communication partner will vary in its impact upon the young person’s feelings and development of depressive symptoms according to age. Would the link between social deficits and depression be more observable during peer interactions than during exchanges with adults ?

This further issue was built into the research of van Beek et al whose first study involved exchanges between their two samples of pre-adolescents with like aged peers, and whose second study linked their target samples with adult communication partners.

The methodology included the use of the Children’s Depression Inventory by which to establish the level of symptomatology among the participating children.

In the study involving adult partners, the primary school children (N = 122; 58 boys and 64 girls) were filmed in their conversation with a female college student. The children were encouraged to describe both positive and negative experiences with peers. Then the adult gave accounts about such experiences providing prolonged listening periods during which the listening and non-verbal behaviour of the target children could be examined. In particular, there was a record made of the onset and duration of gaze directed to the speaker’s face, the onset and duration of speech, backchannel behaviours, and the extent of smiling and laughing.

In the study involving same-sex peers, the secondary school participants (N = 154; 73 boys and 81 girls) were paired with partners of the same age and same sex, but not classmates or close friends. The partners were identified as having no signs of depression.

The pairs were asked to discuss some social dilemma concerning relationships with peers until they agreed on the best solution. If this was achieved very quickly, they were invited to discuss what characteristics apply to the most popular teachers or those that apply most importantly to friendships. Again, listening and non-verbal behaviours were recorded.

In both studies, the target participants were sub-divided into two groups … those with measurable symptoms of depression and those with no such symptoms … in order to compare their interactive styles.

The results were in accord with expectation in that gender-specific social rules were more visible in the displays of the older children, although some different age effects were observed for the various behaviours examined.

Females in both studies were more likely than males to focus their gaze upon partners while listening.

In the peer interactions, a gender difference was clearly evident for the gaze behaviour suggesting the increasing importance of this behavioural style for girls as they get older. Meanwhile, the older boys showed a higher visual dominance ratio than the girls during peer interactions.

Unlike the pattern observed with girls, boys’ gazing behaviour while listening decreased with age …. confirming the validity of the gender specific nature of the display “rules” which indicate that females should behave in a more “ other-oriented” manner.

Backchannel behaviours increased with age in both studies. Among 10 year olds, no significant gender differences were observed; but, by age 13, the females demonstrated more backchannel behaviour than boys.

Further, females in the older age group showed a higher frequency of smiling than their male counterparts, but no differences of this kind emerged in the younger group, suggesting again that this difference in style emerges during middle adolescence.

Turning to the issue of how the experience of depression might modify these behavioural patterns, the authors describe how the effects of (mild) depression differed between the two studies.

In particular, in comparing the oldest children with the youngest children, there appeared to be a partner effect. The suggestion was that the more “dominating” adult figure (compared to a like-aged peer) was more likely to elicit gazing and smiling even among the children with depression. In peer interactions, it was speculated that more anxiety would be evoked in mildly depressed children and adolescents because of negative experiences with peers and a heightened fear of rejection … as reflected in some inhibition of gazing and smiling.

In any event, there were clear differences observable in backchannel behaviours between the depressed and non-depressed participants during interactions with an adult (possibly, it was held, because these behaviours are less conscious and less directly controlled than gazing or smiling).

Among the older children in the first study, it was found that the adult responded to the lower frequency of backchannel behaviours in those with depression by showing more backchannel behaviour herself.

In study 2, mildly depressed adolescent girls gazed less during listening, and their partners showed a reduced frequency of smiling.

This was taken to support the hypothesis that the behaviour of mildly depressed adolescents (especially girls in their interaction with peers) showed signs of lesser ability or motivation to adopt behaviours conforming to gender-specific rules.

In addition, their partners showed less positive responses which was held further to support the social skill deficit theory of depression.

The gender linked prevalence was also underlined by the finding that some effects of mild depression were only present in girls. The social skill deficit model was, therefore, thought to be particularly relevant in respect of the gender linked differentiation in the risk for depression during adolescence.

The authors suggested that this could be because the level of “other-oriented” behaviours is usually higher in girls than in boys, so that an absence of such behaviours, or the presence of signs of social anxiety, are more obvious among girls.

On the other hand, not all the current findings were gender specific in that some of these negative responses observable in the partners of boys with depression. The speculation was that this could reflect the presence of other features among the depressed boys, not simply an absence of gaze or smiling, etc. Such features might include some disruptiveness or an aggressive style; or certain characteristics of speech including unusual intonation or volume, along with postural or gestural behaviours that mark out the depressed individual as somehow different.

In their conclusion, the authors repeat that non-verbal behaviours differ between mildly depressed and non-depressed adolescents, even in this non-referred sample of participants, and that these differences match the style of depressed adults.

The (social) deficits are perceived, accordingly, to be concomitants of mild depression and not limited to “full” clinical depression.

The findings also indicate the significance of gender-specific social development in adolescence in that gazing, smiling, and backchannel behaviours are shown more

(and expected more) among girls than boys; and reduced rates of such behaviour are related to depression, especially in girls.

It is acknowledged that these results do not indicate causal links between mild depression and non-verbal behaviours, and longer term research is required to determine if some deviance in the development of gender-specific (social and interactive) behaviours is validly predictive of an increase in the risk of increased depression and/or profound depression ……

…. (with the implication, one might presume, that certain behaviours observable at a pre- or peri-adolescent stage may prove to be markers of potential depression, with the opportunity afforded to initiate preventive or therapeutic actions).

(The present writer - MJC - would also presume that these findings have implications for those children and young people with high functioning ASD or Asperger Syndrome who attend mainstream schools. While, educationally, they will probably cope satisfactorily enough, it is noted that such individuals are more prone to emotional and psychiatric disorders like anxiety and depression. It is likely that this proneness can be linked to the social skill deficits … such as an absence of backchannel behaviours … which are among the defining characteristics of ASD. Such deficits and the associated reactions of significant others, notably peers, may well underlie at least some of the enhanced risk for these additional disorders.)

Attention Difficulties and Ruminative Style in Adolescents with Depression

The study by Wilkinson and Goodyer (2006) begins with their description of depressive disorders as complex and heterogeneous problems within which impaired attention is a significant symptom.

For example, they cite existing evidence that, among adolescent girls, impaired attention or concentration may be observed in less than 10% of a community sample, but in around 70% of those identified with major depression.

Meanwhile, attention is, in itself, a complex concept in comprising an inter-connected set of components including searching, selecting, sustaining, and switching; and it is not clear precisely where in these processes the deficit among individuals with depression is located.

They go on to cite the work of Manly et al (2001) who proposed three distinguishable components of attention ….. attention control and switching (the capacity to switch from one task, or type of task, to another), selective attention (the capacity to filter input to identify what is important and to ignore the rest), and sustained attention

(the capacity to maintain a focus upon a task which has little intrinsic interest or reward).

One of their findings was that boys with ADHD could be consistently differentiated from peers in respect of their impairment in sustained attention but not in respect of selective attention.

On their second theme, these current authors refer to “rumination” (a tendency to dwell upon negative experiences and to allow negative thoughts to persist).

The (excessive) focus of thoughts upon the symptoms of depression and the perceived causes is described as a ruminative mood-related response style; and evidence is cited that a highly ruminative style is predictive of persistent symptoms of depression , and may be seen as a significant cognitive vulnerability prior to the emergence of clinically-significant levels of depression.

The question is raised about how the ruminative thinking style is actually involved in the onset and maintenance of depression, and the authors speculate that it may be a matter of failing to switch attention to a less negative set of cognitions … reflecting a general and underlying difficulty with attentional switching.

Alternatively, it could be that individuals who are depressed focus sttentional resources upon rumination so that there is correspondingly less processing capacity for alternative processing tasks.

Accordingly, Wilkinson and Goodyer set out to determine whether major depression in adolescents is associated with an impairment in attention-switching (which is not linked to the effects of medication or to differences in overall cognitive ability). A related task involved determining whether or not impaired attention-switching is significantly linked to an increased mood-related ruminative style.

The participants were drawn from referrals to community health centres in two English regions who met the DSM-IV criteria for a diagnosis of depression.

The age range of the target sample was from 11 to 17 years.

A control group was recruited from two secondary schools, with these participants matched as far as possible with the target participants for age, gender, and scores in the SATs completed in Year 6 at age 10-11 years.

Among the children and young people with depression, there were 20 receiving anti-depressants (4 male and 16 female) and 19 receiving no medication (6 males and 13 females). The control group included 38 participants, 11 male and 27 female.

The measures included The Test of Everyday Attention for Children which assessed for the ability to switch from one mental set to another, selective attention, and sustained attention; The Responses to Depression Questionnaire which provided information about what individuals think or do or feel during periods of low moods

(with a subscale indicating the degree of rumination and of the ability to use distractors); plus schedules to produce an indication of the level of depression.

The findings confirmed that adolescents diagnosed with depression are significantly slower than controls at switching attentional resources from one task to another, and they showed a clear trend towards making more errors than controls when making the shifts in attention. In other words, the increased response latency was not simply a matter of their taking more care but seems, rather, to indicate that they found these demands more difficult (and the increased difficulty is marked by a deficit in basic processing speed across a range of tasks).

Further, when processing speed was controlled, the continuing difference between the depressed and non-depressed individuals suggested that the findings do not simply reflect reduced motivation.

There was also some indication of impairments among the target group in selective and sustained attention, although the authors recognised the possibility that this could be linked with the use of antidepressant medication which includes tiredness among the side-effects.

Compared to the control participants, the young people with depression described a significantly higher level of a dysphoric style of ruminative thinking … (ie an emphasis upon generalised anxiety and negative cognitions).

The authors suggested that deficits in attentional switching and increases in mood-related ruminative thinking may be independent processes, but that the risk of depression is much heightened when both characteristics are present rather than just one or the other.

(Meanwhile, the present writer - MJC - would return to the theme of autistic spectrum disorder and note the consistency with which an attentional switching deficit is regarded as a characteristic of children and young people so-diagnosed. Given the earlier noted issue about social skill deficits and their association with the possible onset of depressive disorders, there are the beginnings of some understanding of the mechanisms by which individuals with high functioning autism or Asperger Syndrome come to be at greater risk of emotional disorders of a depressive or dysphoric type.)

In any event, returning to the study of Wilkinson and Goodyer, there are some acknowledged limitations, such as the small sample size which makes for difficulty in ruling out the possibilities that the results reflect some co-morbidity or that significant differences exist between medicated and non-medicated individuals. It may also be the case that some impairment in working memory rather than in attentional processing underlies the poor performance in some tasks.

Nevertheless, the authors conclude that existing evidence does suggest that depression in adolescence is linked with dual deficits, one concerned with attention and the other with a maintenance of negative thinking (rumination).

Further studies would be valuable in determining whether a greater ruminative style is directly linked to attentional impairments, and whether both processes commonly precede the emergence of full depression (with implications for early identification of children and young people at risk, via observing individual differences in executive functioning and socio-emotional processing, and for the provision of therapeutic interventions at an equally early stage).

* * * * * *

M.J.Connor January 2007

REFERENCES

Van Beek Y., van Dolderen M., and Dubas J. 2006 Gender-specific development of nonverbal behaviours and mild depression in adolescence. Journal of Child Psychology and Psychiatry 47(12) 1272-1283

Caspi A., Sugden K., Moffitt T., et al 2003 Influence of life stress on depression. Science 18 291-293

Costello E., Erkanli A., and Angold A. 2006 Is there an epidemic of child or adolescent depression ? Journal of Child Psychiatry and Psychology 47(12) 1263-1271

Jaffee S., Moffitt T., Caspi A., Fombonne E., Poulton R., and Martin J. 2002 Differences in early childhood risk factors for juvenile-onset and adult-onset depression. Archives of General Psychiatry 59 215-222

Kessler R., Avenevoli S., and Merikangas K. 2001 Mood disorders in children and adolescents. Biological Psychiatry 49 1002-1014

Luby J., Belden A., and Spitznagel E. 2006 Risk factors for pre-school depression : the mediating role of early stressful life events. Journal of Child Psychology and Psychiatry 47(12) 1292-1298

Manly T., Anderson V., Nimmo-Smith I., Turner A., Watson P., and Robertson I. 2001 The differential assessment of children’s attention. Journal of Child Psychology and Psychiatry 42 1065-1081

Patten S. 2003 Recall bias and major depression lifetime prevalence. Social Psychiatry and Psychiatric Epidemiology 38 290-296

Segrin C. 2000 Social skill deficits associated with depression. Clinical Psychology Review 20 379-403

Wilkinson P. and Goodyer I. 2006 Attention difficulties and mood-related ruminative response style in adolescents with unipolar depression. Journal of Child Psychology and Psychiatry 47(12) 1284-1291