Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: J Affect Disord. 2021 Mar 23;288:31–40. doi: 10.1016/j.jad.2021.03.045

Childhood maltreatment reports in adult seasonal affective disorder: Associations with sleep disturbances, maladaptive cognitions, and brooding

Yuqi S Wang 1, Abbey L Friedman 1, Karen P Jakubowski 2, Delainey L Wescott 1, Praise Iyiewuare 3, Julia S Feldman 1, Daniel S Shaw 1, Kathryn A Roecklein 1
PMCID: PMC8154709  NIHMSID: NIHMS1692740  PMID: 33839556

Abstract

Background:

Although childhood maltreatment has been studied in multiple psychopathologies, its role in Seasonal Affective Disorder (SAD) is unknown. The current study examined possible mediators of the relationship between retrospectively-reported childhood maltreatment and adult SAD symptom severity during a major depressive episode in winter.

Methods:

Participants (N = 113), ages 18 to 65, completed measures of childhood maltreatment, SAD severity, sleep disturbances, ruminative brooding, and maladaptive cognitions. Mediation analyses testing the relationship between childhood maltreatment and SAD symptom severity via sleep and cognitive factors were conducted using PROCESS (Hayes, 2012).

Results:

Mediation analyses suggested that insomnia, hypersomnia, brooding, and seasonal maladaptive beliefs may account for the association between childhood maltreatment and SAD symptom severity.

Limitations:

Analyses were cross-sectional and should be interpreted with caution. Participants completed self-report childhood trauma measure retrospectively as adults.

Conclusion:

The present study is the first to examine childhood maltreatment in SAD, a disorder commonly viewed with circadian etiology. Covariance between childhood maltreatment and SAD symptom severity is indirectly explained by sleep difficulties, cognitive factors, and brooding, which may suggest therapeutic targets if replicated in longitudinal or experimental manipulations of sleep and cognition.

Keywords: Seasonal Affective Disorder, Childhood Maltreatment, Hypersomnia, Insomnia, Brooding

Introduction

Unipolar seasonal affective disorder (SAD) is diagnosed as Major Depressive Disorder With Seasonal Pattern (MDD-SP; American Psychological Association, 2015), and most often recurs during the fall and winter months and remits during the spring and summer (Rosenthal et al., 1984). Approximately 10–20% of individuals with recurrent depression experience a seasonal pattern (Magnusson, 2000). Individuals with SAD spend more than 40% of the year suffering from depression, which negatively affects many aspects of daily life (Rosenthal et al., 1984). Thus, identifying improved treatments for SAD represents an important public health challenge (Rohan et al., 2009).

The majority of past studies on SAD etiology have concentrated on biological mechanisms (Rohan et al., 2009). Decreased levels of light during the winter are theorized to disrupt circadian rhythms and neurotransmitter levels (James et al., 1985). However, cognitive and behavioral mechanisms have also been postulated and tested as risk factors for SAD (Rohan et al., 2009), indicating SAD is not solely a biological disorder. For example, dysfunctional attitudes and ruminative response style may act as cognitive vulnerabilities for depressive onset (Rohan et al., 2003; Golden et al., 2006; Enggasser and Young, 2007; Young et al., 2008), suggesting similarities between SAD and non-seasonal depression. In non-seasonal depression, exposure to trauma, such as childhood maltreatment, combined with an underlying diathesis can lead to adult depression (Liu, 2017).

A meta-analysis of 261 studies found that exposure to childhood maltreatment was related to two-fold greater odds of depression in adulthood (Odds Ratio = 2.80; Mandelli et al., 2015). Despite the relationship of childhood maltreatment with adult depression in non-seasonal depression (e.g., Heim and Binder, 2012), childhood maltreatment has not been investigated in SAD. Of interest, childhood physical and sexual abuse are associated with reverse neuro-vegetative symptoms in non-seasonal depression (i.e., hyperphagia, weight gain, and hypersomnia; Levitan et al., 1998), which are common symptoms in SAD. Although the current study will focus on relations between childhood maltreatment and depression, it should be acknowledged that childhood trauma and maltreatment is related to many other types of psychopathology, including anxiety, alcohol abuse/dependence, antisocial behavior, posttraumatic stress disorder (PTSD), and obesity, as well as premature mortality (e.g., Anda et al., 2006; Bellis and Kuchibhatla, 2006; Jonson-Reid et al., 2007; Macmillan et al., 2001). With this important caveat in mind, an initial goal of the current study was to examine direct associations between adult reports of childhood maltreatment and depression symptom severity in SAD.

If in fact childhood maltreatment is associated with SAD symptom severity in adulthood, two factors that might explain this relationship include (1) increased vulnerability to stress, and (2) cognitive vulnerabilities learned in childhood. Childhood maltreatment is thought to affect brain development in childhood through altered neurotransmission, having lasting impacts into adulthood (Rutter, 2004). Accordingly, increased stress sensitivity from the experience of maltreatment could increase vulnerability to depression in response to stressors in adulthood (Kendler et al., 2004). Disruptions in sleep is another reliable index of heightened stress sensitivity (Harvey et al., 2014). Therefore, the effects of childhood trauma could generate lasting emotional reactivity to stress (Collip et al., 2008), which may interfere with sleep beginning in childhood and continuing through adulthood.

In addition, exposure to childhood maltreatment might lead to maladaptive schemas that could negatively affect information processing in adulthood in response to life events (Young et al., 2003; Dozois et al., 2009). Rumination is another cognitive process or mode of responding to life stress that increases risk of depression, and some evidence suggests that rumination begins to develop in childhood at higher rates because of maltreatment (Nolen-Hoeksema et al., 2008; Gold and Wegner, 1995).

As childhood abuse precedes 30% of adult-onset psychopathology (Green et al., 2010), the timing of the above proposed mediators is important. Thus, it is possible that stress sensitivity and early cognitive patterns develop in childhood and subsequently impact responses to life stress in adulthood, leading to adult depression.

Putative Mediators of the Association between Childhood Maltreatment and SAD

Sleep.

As mentioned above, sleep disruption is one reflection of stress sensitivity (Harvey et al., 2014). Theoretically, exposure to childhood maltreatment could also impair a sense of safety necessary to attain high-quality sleep (Bernier et al., 2013). Childhood maltreatment is associated with greater sleep difficulties (Glod et al., 1997) and sleep disruption (Cuddihy et al., 2013). In adulthood, childhood maltreatment is associated with insufficient sleep (Chapman et al., 2013; McWhorter et al., 2019), poorer sleep quality (Koskenvuo et al., 2010), daytime sleepiness, trouble sleeping (Agargun et al., 2003), later time going to bed, shorter sleep duration, and longer sleep latency (Noll et al., 2006). Insomnia symptoms are related to a two-fold increase in the odds of clinical depression and elevated depressive symptoms in adults (Odds Ratio = 2.10; Baglioni et al., 2011) and hypersomnia symptoms have been shown to be prevalent in 60% to 80% of adult SAD patients (Kaplan and Harvey, 2009; Roecklein et al., 2013). Sleep problems that began in childhood, in response to maltreatment could theoretically persist into adulthood, leading to an increased risk of depression in adulthood. In the case of hypersomnia, childhood trauma may lead to somnolent avoidance, in which individuals attempt to sleep to escape aversive experiences. Accordingly, sleep difficulties would be expected to persist into adulthood (Jacobson et al., 2001). As insomnia (18%) and co-occurring insomnia and hypersomnia (47%) are prevalent in SAD (Roecklein et al., 2013), the current study will examine current insomnia and hypersomnia in adulthood as possible explanations for the hypothesized covariance between childhood maltreatment and adult SAD.

Cognitive Factors.

In addition to sleep, rumination and other cognitive styles could also explain some of the hypothesized association between childhood maltreatment and SAD. Early experiences undergo cognitive and affective processing that influences an individual’s self-concepts and cognitive schematas of their environment (Rutter, 2006). Childhood maltreatment, especially emotional abuse, has been related to cognitive risk for depression in adulthood (Sachs-Ericsson et al., 2006; Liu, 2017). According to the cognitive model of depression (e.g., Rose and Abramson, 1992), cognitive schemas tend to be most malleable during early childhood and become more solidified from childhood into adolescence. Thus, exposure to childhood maltreatment may lead to the development of more negative cognitive styles, resulting in greater risk for adult depression after these depressogenic cognitive styles become entrenched. Importantly, adults with a history of childhood maltreatment history but no history of psychopathology, have been shown to remain at long-term risk for depression because of cognitive vulnerabilities (Wells et al., 2014).

Dysfunctional beliefs stem from maladaptive core beliefs, and contain negative content about the self, the world, and others (Miranda and Persons, 1988). Although evidence for childhood maltreatment and dysfunctional beliefs exists in non-seasonal depression (Bernet and Stein, 1999; Lewis et al., 2010), this relationship has not been tested in SAD. In SAD, dysfunctional beliefs are elevated compared to controls, and similar in magnitude to those with non-seasonal depression (Dalgleish et al., 2004; Golden et al., 2006). Cognitive risk factors in SAD include both typical depressogenic maladaptive thoughts and seasonal beliefs about the seasons, weather conditions and environmental light levels (i.e., dark or dreary days; Rohan et al., 2019). Seasonal beliefs have been proposed to constitute a unique cognitive vulnerability in SAD (Rohan et al., 2019).

Rumination, a form of repeated negative thought, has previously been identified as a mediator of the relationship between childhood maltreatment and non-seasonal clinical depression (Spasojevic and Alloy, 2002; Raes and Hermans, 2008), and also has been shown to predict the severity of depression during winter months (Young and Azam 2003; Rohan et al., 2003). Rumination may be uniquely related to SAD in response to neuro-vegetative symptoms of depression (e.g., fatigue, increased appetite, hypersomnia), predicting SAD episode onset better than rumination about typical depressive symptoms (Young, 2008). Brooding, defined as “a passive comparison of one’s current situation with some unachieved standard”, is a component of rumination that is thought to be highly maladaptive and also associated with SAD (Treynor et al., 2003; Enggasser and Young, 2007). The current study will examine brooding as a mediator, as other components of rumination, such as reflection, are less predictive of depression (Treynor et al., 2003) and possibly also less important in SAD.

We propose that dysfunctional beliefs, seasonal beliefs, and brooding initiated in childhood, in response to maltreatment, solidify in adolescence and persist into adulthood, leading to risk for depression. The current study will examine dysfunctional beliefs, seasonal beliefs, and brooding as mediators for the hypothesized relationship between childhood maltreatment and adult SAD symptom severity. The sample included individuals with a continuous range of SAD symptom severity, from clinically significant SAD to individuals with no history of depression and no self-reported seasonality.

Hypotheses

First, we hypothesized that childhood trauma would be positively correlated with higher SAD symptom severity during winter. Second, we hypothesized that childhood maltreatment would have an indirect effect on SAD symptom severity through the following hypothesized mediators: sleep disturbances (i.e., insomnia and hypersomnia), dysfunctional beliefs, seasonal beliefs, and brooding. Third, we hypothesized that some mediators would be correlated (e.g., dysfunctional beliefs and seasonal beliefs), thus, all five mediators were included in the same model to test which variables would remain signficiant outside of individual models.

Methods

Participants

Recruitment.

Individuals aged 18 to 65 were recruited from the greater Pittsburgh area through the Pitt+Me participant registry program of the University of Pittsburgh Clinical Translational Science Institute. Self-reported childhood maltreatment was added to the parent study on biomarkers in SAD. Participants volunteering for the study provided a genetic sample and underwent higher participant burden for biomarker assessments. The parent study recruited individuals along the full continuum of SAD severity consistent with the NIMH RDoC approach that psychopathology is the extreme end of the range of normal brain functioning and that disorders can be best understood in terms of dysregulation or impairment in these normal processes. Gender, race, and ethnicity were evenly distributed across this continuum by categorizing participants into contiguous groups (i.e., SAD, sub-syndromal SAD, and control) during sample ascertainment to avoid demographic under-representation at any point along the spectrum. This study was approved by the Institutional Review Board at the University of Pittsburgh and participants were compensated for time spent during participation.

Participant ascertainment procedures.

Inclusion and exclusion criteria were assessed briefly by phone and comprehensively in person, followed by assessment of retrospective self-report of childhood trauma. Depression symptom severity, sleep, rumination, and dysfunctional attitudes were assessed in all participants during winter months when individuals with SAD met criteria for a major depressive episode. Select modules of the Structured Clinical Interview for DSM-IV (SCID-IV; Spitzer et al., 2001) were administered to assess inclusion and exclusion diagnostic criteria. In addition, the following two SAD-specific measures were employed.

The Structured Interview Guide for the Hamilton Rating Scale for Depression—Seasonal Affective Disorder Version (SIGH-SAD; Williams et al., 1994) was administered to assess SAD symptom severity both for inclusion criteria and as a dependent variable. The SIGH-SAD includes the 21-item HAM-D (Hamilton, 1960) and 8 items assessing atypical symptoms of depression such as hyperphagia and hypersomnia. Total SIGH-SAD scores were computed without sleep items (insomnia and hypersomnia) because the SIGH-SAD is used as the outcome measure and sleep items are tested as mediators. The 29 items that comprise the SIGH-SAD total score had a Cronbach’s α = .89, which is consistent with other studies (α’s = .46–.92; Bagby et al., 2004).

As an additional assessment to characterize seasonality, participants completed the self-report Modified Seasonal Pattern Assessment Questionnaire (M-SPAQ; Blouin et al., 1992) that contains the global seasonality scale (GSS). Participants rate 6 SPAQ items (sleep, appetite, mood, energy level, weight, and social behavior) on a 5-point Likert scale for degree of change across the seasons that are summed to derive the GSS. Individual questions on the SPAQ address the above items comprising the GSS and the following; months of year in which participants endorse feeling best/worst, gaining or losing the most weight, socializing most/least, sleeping most/least, eating most/least; how much weight fluctuates during the course of the year (from 0 to over 20 lbs.); how many hours per day the individual sleeps in winter, spring, summer, and fall; changes in food preferences across seasons; and, an appraisal of whether or not any changes in mood and behavior across the seasons represent a problem (i.e., no problem, mild, moderate, severe, marked change).

Inclusion Criteria.

To meet criteria for the SAD group, participants (1) had a GSS of >10 with moderate severity and reported “feeling worst” during January and/or February, with or without other affected months, excluding July and August on the M-SPAQ (Kasper et al., 1989); (2) scored >19 total, >9 on the HAM-D, and >4 on the ATYP subscale of the SIGH-SAD thereby meeting Terman et al. (1990) criteria for a current SAD episode; and (3) met criteria for MDD-SP on the SCID. The criteria for the sub-syndromal SAD (S-SAD) group aimed to include individuals with intermediate SAD symptom severity. S-SAD participants had (1) a GSS of > 9 with no problems or mild severity or a GSS of 8 or 9 and at least mild severity on the M-SPAQ (Kasper et al., 1989), (2) SIGH-SAD scores below those for a current SAD episode, and (3) no current diagnosis for MDD-SP on the SCID-IV, although those with past Major Depressive Episodes were eligible. Individuals in the control group had (1) no lifetime history of MDD, (2) scored below 8 on the M-SPAQ GSS or scored 8 or 9 but report no problems across the seasons, and (3) scored below current SAD episode criteria on the SIGH-SAD.

Exclusion Criteria.

Participants with a history of Bipolar Disorder, Psychosis, Sleep or Circadian Disorders, Substance Abuse, Autism Spectrum Disorder (ASD), Anorexia Nervosa, or Post-Traumatic Stress Disorder (PTSD) were excluded. Participants who were receiving treatment for depression but still met criteria for a current major depressive episode were retained. Because participants were recruited for the biomarker study, individuals were free of self-reported retinal health disorders and disease, did not have Diabetes as determined using fasting blood glucose, and were not shift workers.

Enrollment.

A total of 289 participants were invited to the laboratory after being screened by phone. Sixty-eight participants did not meet inclusion and exclusion criteria after informed consent. Twenty-three participants withdrew after informed consent as they were no longer interested in participating in the study due to time constraints. Forty-four participants did not return to the laboratory for a winter assessment. Of the remaining 154 participants, 41 were missing one of the questionnaires in the analyses, resulting in a sample of 113 participants. Missing questionnaires were due to participants taking questionnaires home to complete and not mailing them back. Participation ranged from 2013 to 2018, concluding prior to the COVID-19 pandemic.

Assessment of Childhood Maltreatment

The Childhood Trauma Questionnaire (CTQ; Bernstein and Fink, 1998) assessed five types of maltreatment: emotional, physical and sexual abuse, and emotional and physical neglect. Participants were asked to rate statements on a five-point Likert scale ranging from “never true” to “very often true”. The 25-item CTQ has demonstrated high internal consistency, validity, and reliability (Bernstein et al., 1994). The total score ranges from 25 to 125 and the subscales range from 5 to 25. The total CTQ score was used rather than individual subscales as there are no specific hypotheses about certain types of abuse being more important in SAD. The CTQ total score had acceptable internal consistency (α = .86), consistent with previous reports (Bernstein et al., 1994).

Assessment of Sleep

The Insomnia Severity Index (ISI; Morin et al., 2011) assessed self-reported severity of insomnia symptoms in the past two weeks. Items in the ISI include “difficulty falling asleep”, “difficulty staying asleep” and “problems waking up too early”. Reliability and validity are high (Morin et al., 2011; Vegar and Hussain, 2017). The ISI had strong internal consistency (α = .91) in the current sample, comparable to other studies (α = .84 – .88; Vegar and Hussain, 2017).

Hypersomnia was assessed with one item (A6) from the SIGH-SAD (Williams et al., 1994; “Have you been sleeping more than usual in the past month? IF YES: How much more?”). Answers were scored from 0 to 4, indicating ‘no increase in sleep length’, ‘at least 1 hour increase in sleep length’, ‘2-hour increase’, ‘3-hour increase’ or ‘4-hour increase’ on average per day. Secondary probes are intended to help determine average daily sleep length and to account for naps and weekend sleep.

Assessment of Cognitions

The Dysfunctional Attitude Scale (DAS; Weissman and Beck, 1978) is a 40-item self-report measure that assesses endorsement of maladaptive beliefs about oneself and the broader world. Questions include, “I am nothing if a person I love doesn’t love me,” “If I fail at my work, then I am a failure as a person,” and “People will probably think less of me if I make a mistake.” The DAS is a well validated measure that is correlated moderately (r = .65) with the Beck Depression Inventory (Weissman and Beck, 1978; Richter et al., 1998). In the current sample, the DAS had acceptable internal consistency (α = .83).

The Seasonal Beliefs Questionnaire (SBQ; Rohan et al., 2019) is a 26-item scale that measures self-reported dysfunctional beliefs about the seasons and light availability. Questions include, “It’s difficult to feel good on dark, dreary days,” “There is something wrong with me in the winter,” and “Everything is easier in the summertime.” The SBQ has demonstrated good internal consistency, validity, and test-retest reliability (Rohan et al., 2019), including in the present sample (α = .84).

The Ruminative Response Scale (RSS; Nolen-Hoeksema and Morrow, 1991) measures the tendency to ruminate in response to negative events. The brooding subscale, made up of five Likert scale items, assesses the tendency to dwell passively and self-critically on one’s symptoms or personal shortcomings (Treynor et al., 2003). The RSS has demonstrated good test-retest validity and acceptable convergent and predictive validity (Treynor et al., 2003). In the current sample, the RSS had good internal consistency (α = .96).

Statistical Analysis

Descriptive and correlational analyses were conducted in SPSS Version 25 (Armonk, NY) to assess sample characteristics and correlations among variables. One-way ANCOVA tests and Chi-square tests were conducted to examine between-group differences in demographics, childhood trauma, and SAD symptom severity between the control, S-SAD, and SAD groups. Age and gender were included as covariates in all analyses based on their associations with sleep and SAD symptom severity (Magnusson, 2000; Reyner and Horne, 1995). The SAD, S-SAD, and control groups fall along a continuous dimension in terms of SAD symptom severity, so participants were combined for the subsequent mediation analyses. To test hypothesis one, a linear regression was computed to analyze the relationship between childhood maltreatment and SAD symptom severity. To test hypothesis two, mediation analyses tested indirect pathways between childhood maltreatment and SAD symptom severity via five potential variables (i.e., insomnia, hypersomnia, dysfunctional beliefs, seasonal beliefs, and brooding). Each mediation model and an omnibus model (i.e., hypothesis three) was tested using the PROCESS MACRO 3.3 (Hayes, 2012) with Ordinary Least Squares (OLS) regression and bootstrapping with 5000 resamples. Statistical significance was tested using 95% confidence intervals. Bootstrapped standard errors are robust to nonnormal distribution (Hayes, 2013), thus, bootstrapping was used instead of the Sobel test (Sobel, 1982) to increase statistical power. Individual and omnibus mediation analyses were conducted to explore collinearity. The distribution of total childhood maltreatment scores and hypersomnia item scores were zero inflated, meaning that zero was a common score and the distribution was not normal. However, the bootstrap methods used do not rely on assumptions of normality (Preacher and Hayes, 2008; Hayes, 2013).

In mediation analyses, the total effect represents the relationship between childhood maltreatment and SAD symptom severity without the inclusion of mediators. Therefore, the total effect will remain the same for all models. The direct effect represents the same relationship, but including the mediator(s). The significance of the indirect effect indicates whether or not a variable is a mediator of the relationship. As a measure of effect size for each mediator, proportion mediated (PM), with covariates adjusted, was calculated by dividing the indirect effect by the total effect (MacKinnon, 2012).

Results

Sample Characteristics and Bivariate Correlations

Participants (N = 113) were on average 38 years old (SD = 14.3) and predominantly female (82%), which approximates the gender ratio of SAD and is consistent with the targeted gender ratio for sample ascertainment (see Table 1). Results demonstrated a significant main effect of group on age but not on any other demographic variables. Individuals were older in the SAD and S-SAD groups compared to the control group. As expected, the SAD group had the highest SAD symptom severity. Bivariate correlations between study variables are reported in Table 2. All hypothesized mediator (M) variables were significantly correlated in expected directions with childhood maltreatment and SAD symptom severity, except for dysfunctional beliefs, which was not significantly associated with childhood maltreatment.

Table 1.

Participant Demographics

Sample SAD S-SAD Control Statistic2
N 113 48 23 42
Measure N(%) N(%) N(%) N(%)
Race X2 = 8.37, ns
 Caucasian 92(82) 40(83) 18(78.3) 34(81)
 African-American 12(10) 7(15) 2(8.7) 3(7)
 Asian 7(6) 0(0) 2(8.7) 5(12)
 More than one race 2(2) 1(2) 1(4.3) 0(0)
Ethnicity X2 = 3.68, ns
 Hispanic/Latino 5(4) 4(8) 1(4) 0(0)
 Non-Hispanic/Latino 108(96) 44(92) 22(96) 42(100)
Gender X2 = 3.48, ns
 Female 93(82) 42(88) 16(70) 35(83)
 Male 20(18) 6(12) 7(30) 7(17)
Measure M (SD) M (SD) M (SD) M (SD)
Age in years 39 (14.3) 40 (13.8) 43 (15.5) 34 (13.5) F(2,112) = 3.09, p < 0.05+
Years of school 17 (2.6) 16.3 (2.5) 16.8 (3.6) 16.6 (1.9) F(2,112) = 0.27, ns
Occupational rank1 8 (7.9) 6.6 (6.8) 7.7 (7.8) 10.2 (8.8) F(2,112) = 218, ns
Childhood Maltreatment 47.5 (14) 53.2 (13.3) 46.9 (13.5) 41.4 (12.7) F(2,112) = 9.18, p < 0.001◇
SAD Symptom Severity 15.1 (12) 26 (7.7) 13.9 (7.3) 3.4 (4.1) F(2,112) = 135.5, p < 0.001◇×+
Open in a new tab
1

A higher score means lower occupational rank. The lowest occupational rank is scored as 21 which means the participant in unemployed or retired. The highest occupational rank is scored as 1 which means the participant holds an executive, administrative or managerial position.

2

Main effect of group (SAD, S-SAD, Control) on a given variable.

◇

Significant difference between SAD and Control groups.

×

Significant difference between S-SAD and Control groups.

+

Significant difference between S-SAD and SAD groups.

Table 2.

Bivariate Correlations between main variables (R2 values). N = 113

Variables 1 2 3 4 5 6
1. SAD severity -
2. Childhood Maltreatment .376*** -
3. Insomnia .622*** .249** -
4. Hypersomnia .470*** .216** .278** -
5. Dysfunctional Beliefs .124 −.191* .131 .147 -
6. Seasonal Beliefs .705*** .197* .509*** .422*** .252** -
7. Brooding 604*** .331*** .610** .343*** .307** .457***
Open in a new tab
*

p < .05

**

p < .01

***

p < .001

Hypothesis One: Childhood Maltreatment and SAD Symptom Severity

In a hierarchical regression, the results of step 1 revealed that the variance accounted for by age and gender was not significant (R2 = .028, F(2, 110) = 1.61, p = 0.2). In step 2, childhood maltreatment was a significant predictor of SAD symptom severity above and beyond age and gender (ΔR2 = .13, β = .31, F(1, 109) = 16.30, p < 0.001).

Hypothesis Two: Individual Mediation Models

Hypothesis two tested the following variables as cross-sectional mediators in the relationship between childhood maltreatment and SAD symptom severity: insomnia, hypersomnia, dysfunctional beliefs, seasonal beliefs, and brooding. For the analyses including mediators of insomnia, hypersomnia, and seasonal beliefs, childhood maltreatment was associated with each mediator, each mediator was associated with SAD symptoms severity, and the direct effect was significant for all three models. Further, in each case, the indirect effect was significant, consistent with cross-sectional mediation. The proportion mediated (PM) effect size of the indirect pathways was 53% for insomnia, 24% for hypersomnia, and 35% for seasonal beliefs. Table 3 presents the effect coefficients, total effects, direct effects, indirect effects, and PM for each pathway in each of the proposed mediation models.

Table 3.

Results of Individual Mediation Analyses (N = 113)

Effect Path β [95% CI]a B [95% CI]b P-value
Insomnia F(4, 108) = 21. 88, p < 0.001
 a CTQ → ISI 0.326 0.140 [0.064, 0.216] <0.001
 b ISI → SIGH-SAD 0.586 1.168 [0.862, 1.474] <0.001
 TE (c) CTQ → SIGH-SAD 0.360 0.308 [0.157, 0.459] <0.001
 DE (c’) CTQ → SIGH-SAD 0.169 0.145 [0.015, 0.275] <0.05
 IE 0.191 [0.091, 0.297] 0.163 [0.083, 0.259]
 PM 0.531
Hypersomnia F(4,108) = 12.30, p < 0.001
 a CTQ → A6 0.209 0.015 [0.005, 0.031] <0.05
 b A6 → SIGH-SAD 0.407 4.752 [2.863, 6.640] <0.001
 TE (c) CTQ → SIGH-SAD 0.360 0.308 [0.157, 0.459] <0.001
 DE (c’) CTQ → SIGH-SAD 0.275 0.235 [0.095, 0.376] <0.01
 IE 0.085 [0.019, 0.162] 0.073 [0.015, 0.148]
 PM 0.236
Dysfunctional Beliefs F(4,108) = 6.87, p < 0.001
 a CTQ → DAS −0.110 −.0364 [−0.965, 0.237] ns
 b DAS → SIGH-SAD 0.232 0.060 [0.014, 0.107] <0.05
 TE (c) CTQ → SIGH-SAD 0.360 0.308 [0.157, 0.459] <0.001
 DE (c’) CTQ → SIGH-SAD 0.386 0.330 [0.181, 0.479] <0.001
 IE −0.026 [−0.087, 0.017] 0.022 [−0.078, 0.017]
 PM 0.072
Seasonal Beliefs F(4,108) = 33.34, p < 0.001
 a CTQ → SBQ 0.187 0.443 [0.0128, 0.873] <0.05
 b SBQ → SIIGH-SAD 0.667 0.241 [0.192, 0.290] <0.001
 TE (c) CTQ → SIGH-SAD 0.360 0.308 [0.157, 0.459] <0.001
 DE (c’) CTQ → SIGH-SAD 0.235 0.201 [0.089, 0.314] <0.001
 IE 0.125 [0.011, 0.235] 0.107 [0.008, 0.206]
 PM 0.347
Brooding F(4,108) = 20.19, p < 0.001
 a CTQ → RSS-B 0.415 0.099 [0.057, 0.141] <0.001
 b RSS-B → SIGH-SAD 0.576 2.061 [1.492, 2.630] <0.001
 TE (c) CTQ → SIGH-SAD 0.360 0.308 [0.157, 0.459] <0.001
 DE (c’) CTQ → SIGH-SAD 0.121 0.104 [−0.033, 0.241] 0.137
 IE 0.239 [0.135, 0.347] 0.204 [0.120, 0.294]
 PM 0.664
Open in a new tab
a

β [95% CI] are completely standardized measures. The bootstrapped confidence intervals are provided for the indirect effect while the p-value is calculated for the other effects.

b

B [95% CI] are unstandardized measures.

TE, total effect; DE, direct effect; IE, indirect effect; PM, proportion mediated.

Abbreviations: CTQ – Child Trauma Questionnaire; DAS – Dysfunctional Attitudes Scale; SIGH-SAD – Structured Interview Guide for the Hamilton Depression Rating Scale Seasonal Affective Disorder Version; ISI – Insomnia Severity Index; RRS-B – Rumination Response Scale – Brooding Subscale; SBQ – Seasonal Beliefs Questionnaire; A6 – self-reported Hypersomnia item from the SIGH-SAD.

In the model testing dysfunctional beliefs as a mediator, childhood maltreatment was not significantly associated with dysfunctional beliefs (see Table 3), although dysfunctional beliefs were significantly associated with SAD symptom severity and the direct effect was significant. The association between childhood maltreatment and SAD symptom severity was not mediated by dysfunctional beliefs. Although not statistically significant, the proportion mediated (PM) was 7% for dysfunctional beliefs.

In the model testing brooding as a mediator, childhood maltreatment was significantly associated with brooding (see Table 3) and brooding was significantly associated with SAD symptom severity, but unlike the above analyses, the direct effect was no longer significant once brooding was included in the model. As the indirect effect was significant, the association between childhood maltreatment and SAD symptom severity was mediated by ruminative brooding symptoms. The indirect pathway via brooding yielded an effect size of 66% proportion mediated (PM).

Hypothesis Three: Omnibus Mediation Model

An omnibus mediation model tested the four variables (insomnia, hypersomnia, seasonal beliefs, and brooding) that emerged as significant mediators of the relationship between childhood maltreatment and SAD symptom severity in individual models; results are reported in Figure 1 and Table 4. Childhood maltreatment was significantly associated with insomnia, hypersomnia, seasonal beliefs, and brooding. Insomnia, hypersomnia, seasonal beliefs, and brooding remained significantly associated with SAD symptom severity. In this mediation analysis, the association between childhood maltreatment and SAD symptom severity was mediated by all four mediators (insomnia, hypersomnia, seasonal beliefs and brooding); see indirect effects in Table 4. Completely standardized measures of the indirect effect were calculated with bootstrapped confidence intervals and are reported in Table 4. Comparing these standardized effects indicate that insomnia, seasonal beliefs, and brooding are all of similar magnitude (β = 0.076–0.080), while hypersomnia was lower (β = 0.031). Similarly, proportion mediated of the indirect pathways was 26% for insomnia, 25% for seasonal beliefs, and 24% for brooding, while the proportion mediated by hypersomnia was 10%.

Figure 1.

Figure 1.

Open in a new tab

Omnibus Mediation Model

β is the standardized effect size. The direct association between childhood maltreatment and SAD symptom severity is shown as the c path. Paths for a and b for each of the four cross-sectional mediators are shown on the left and right. The c’ data indicate that including the mediators in this multiple mediaton model leads the association between childhood maltreatment and SAD symptom severity to be no longer statistically significant, supporting mediation.

Table 4.

Indirect Effects of Omnibus Mediation Analysis (N = 113)

Indirect Effect β [95% CI]a PM B [95% CI]b
Insomnia 0.080 [.02, .16] 0.258 0.069 [.02, .14]
Hypersomnia 0.031 [.001, .07] 0.100 0.026 [.001, .06]
Seasonal Beliefs 0.076 [.01, .16] 0.245 0.065 [.006, 14]
Brooding 0.076 [.002, .16] 0.235 0.065 [.001, .14]
Open in a new tab
a

β [95% CI] are completely standardized measures.

b

B [95% CI] are unstandardized measures.

Discussion

This is the first study to report that individuals who retrospectively reported more childhood maltreatment had higher depression severity. Cross-sectional mediation analyses found that insomnia, hypersomnia, seasonal beliefs, and brooding, but not dysfunctional beliefs, mediated the association between childhood maltreatment and SAD symptom severity. In the omnibus model, insomnia, hypersomnia, seasonal beliefs, and brooding remained significant mediators of the association between childhood maltreatment and SAD symptom severity. The effect size for each of these associations was medium for insomnia (26%), seasonal beliefs (25%), and brooding (24%) while it was small for hypersomnia (10%; Cohen, 1988). These findings may indicate that insomnia, hypersomnia, seasonal beliefs, and brooding independently contribute to the association between childhood maltreatment and depression severity in SAD. However, hypersomnia may contribute to this association to a lesser degree than insomnia, seasonal beliefs, and brooding. Because of the cross-sectional nature of our study, these relationships would need to be tested prospectively or in experimental manipulations of sleep or cognitions in order to draw causal inferences.

The findings from this study extend existing literature on the relationship between childhood maltreatment and non-seasonal depression to the population of individuals with a seasonal pattern of unipolar depression, although a limitation is that only current sleep was assessed, as discussed below. Insomnia as a mediator between childhood maltreatment and SAD symptom severity is consistent with heightened stress sensitivity theory which posits that individuals who have experienced childhood maltreatment experience elevated stress sensitivity shortly after incidents of maltreatment, precluding a sense of safety necessary for sleep (Barlow, 2003; Bernier et al., 2013), although sleep during childhood was not assessed here. Our findings also support previous literature suggesting that hypersomnia in SAD may be a reflection of somnolent avoidance (Jacobson et al., 2001) and/or increased time in bed (Kaplan and Harvey, 2009). Due to the co-occurance of insomnia and hypersomnia in SAD (Roecklein et al., 2013) and nonseasonal depression (Soehner et al., 2014) it is possible that insomnia and hypersomnia, rather than distinct symptom profiles, are related processes that may reinforce one another. For example, early-onset insomnia could extend into morning hypersomnia (e.g., staying in bed longer), which could then precipate difficulty initiating sleep the following night.

In this study, seasonal beliefs had higher predictive power than dysfunctional beliefs and remained a mediator in the omnibus model which may indicate that seasonal beliefs are better predictors of the association between childhood maltreatment and depression severity in SAD than typical depressogenic maladaptive beliefs reflected in the Dysfunctional Attitudes Scale (Weissman and Beck, 1978). Although empirical studies are few, one possible explanation for this finding is that childhood maltreatment might have occurred during the fall and winter months, making the time of year a trigger for traumatic memories (Beratis et al., 1994; Cohen, 2007). Another explanation is mood state congruence, in which one’s negative beliefs are more activated when one is depressed and the memories that are more negatively valanced include childhood maltreatment (Matt et al., 1992). Accordingly, it would be expected that seasonal beliefs, but not general depressogenic beliefs, might mediate the link between childhood maltreatment and SAD symptom severity; further, seasonal beliefs are a better predictor of SAD (Rohan et al., 2019).

Additionally, our study found that the brooding component of rumination mediated the relationship between childhood maltreatment and depression severity in SAD, extending previous research on non-seasonal depression (Spasojevic and Alloy, 2002; Raes and Hermans, 2008) to SAD. The findings of seasonal beliefs and brooding as mediators are consistent with cognitive models of depression, in which individuals who have experienced maltreatment in childhood (when their cognitive schema tend to be malleable) develop depressogenic cognitive styles, increasing risk for depression in adulthood (Rose and Ambramson, 1992; Liu, 2017).

Strengths

The present study is the first to our knowledge to test an association between childhood maltreatment and SAD symptom severity. It is also the first to examine self-reported sleep disturbances, maladaptive cognitions, and brooding as mediators in the relationship between childhood maltreatment and SAD symptom severity. The findings from this study contribute to our understanding of sleep disturbances and cognitive style as putative mechanisms in SAD. All variables were assessed using well validated and internally consistent measures. The study also included participants with a full range of seasonality and depression severity.

Limitations

The current study used a cross-sectional design, limiting our ability to infer directionality, much less causality, with respect to direct or mediated models and the timeframe in which risk factors and depression develop. Future work should consider using prospective longitudinal designs, which would also allow us to determine when depression develops in response to childhood maltreatment and specifically identify the timing of when risk factors develop. The present study only measured current adult depression symptoms severity, while earlier episodes of depression commonly occur. Because childhood maltreatment, insomnia, dysfunctional beliefs, seasonal beliefs, and brooding were measured using self-report questionnaires, our results may be influenced by response bias, memory bias, and individual differences in the participants. While self-report sleep measures capture the perceptions of one’s sleep, behavioral or objective measures capture actual sleep behaviors such as periods of physical rest or inactivity (Girschik et al., 2012; Lauderdale et al., 2008). Additionally, hypersomnia was measured with a single question response which limits content validity and reliability. Future studies could also employ more objective and behavioral measures of sleep such as actigraphy, and measure sleep latency, start time, duration, and end time. There are also limitations associated with the use of retrospective reporting of childhood maltreatment. Following youth prospectively from childhood or using government documentation and/or medical records of child abuse incidents could improve assessment of childhood maltreatment (McKinney et al., 2009). Although we did not analyze the minimization scale of the CTQ, minimization, if present, would have the effect of attenuating the impact and prevalence of childhood trauma. Mean values of each subscale indicate that emotional neglect had the highest mean value and sexual abuse had the lowest mean value. Separate sub-scale analyses would allow for a comparison of the relative effect of different abuse and neglect categories in future work with larger samples. Exclusion criteria for the study included comorbid mental and physical health conditions related to the biomarkers assessed as part of separate aims. Employing these criteria may mean that individuals with fewer comorbid health conditions were more likely to be excluded. Thus, as rates of childhood trauma are associated with conditions including PTSD and diabetes (e.g., Nasca et al., 2019), our exclusion criteria may have resulted in lower rates of self-reported childhood trauma in our sample. Participants who reported receiving treatment for SAD (i.e. antidepressant medication, bright light therapy, psychotherapy) may have lower scores for all constructs, as the treatment may have attenuated their symptoms, including sleep problems, maladaptive cognitions, and depressed mood. However, among the 11 participants reported receiving treatment, one person reported no history of treatment, and the remaining 101 participants were not asked about treatment history because assessment of treatment history was added later in the study. Therefore, treatment effects could not be analyzed in the present sample, but should be considered in future work. Finally, future studies should also recruit a greater number of participants who are located in places of different latitudes in order to increase generalizability of results.

In addition, future studies may benefit from the inclusion of potential moderators of the association between childhood maltreatment and adult SAD including genetic factors (i.e., 5-HTTLPR) the specific developmental stage at which maltreatment may have occurred (Liu, 2017), as well as the maltreatment’s chronicity (Warmingham et al., 2019) and perpetrator (e.g., stranger vs. family member – betrayal level; Fergusson et al., 2008; Steel et al., 2004). In the case of SAD specifically, it may be possible to determine whether maltreatment occurred in fall or winter through interviews, or if the experience of chronic maltreatment was exacerbated by season. It may also be the case that behavioral inactivation in winter among children with a predisposition to SAD could interact with the experience of childhood maltreatment to lead to a seasonally recurring pattern of depressive episodes in adulthood. Although the present study focused on sleep, cognitive, and behavioral aspects of SAD, a biopsychosocial approach as described in the integrative model of SAD that incorporates biological, psychological, and environmental risks in the development of SAD would be more comprehensive (Rohan et al., 2009).

Although SAD has been historically understood as a biological or chronological disorder, the elevated rates of childhood maltreatment suggest important implications for both research and treatment. For example, trauma-informed interventions for depression have been found to significantly decrease depression symptomatology in incarcerated women (Kubiak et al., 2012; Lynch et al., 2012). However, there are still mixed reviews in the field, with several studies showing positive outcomes but with many limitations (e.g., small sample size). Therefore, researchers have suggested that the evidence is not strong enough to make conclusive recommendations about trauma-informed interventions (Warshaw et al., 2013). Future studies should test the efficacy of these new treatments and expand them to populations such as SAD. This research could be critical as only about 50% of individuals with SAD recover with light therapy (Terman et al., 1988), and up to about 60% of individuals treated with CBT-SAD improve (Rohan et al., 2007). Sleep may also be an important target for treatment in SAD. Troxel et al. (2012) discovered that prolonged sleep latency, insomnia, and short sleep duration predicted an increased risk of non-remission in depression. Furthermore, an experimental pilot study of individuals with comorbid depression and insomnia found that all participants no longer met criteria for insomnia and the majority reported non-clinical post-treatment depression scores after receiving cognitive-behavioral therapy for insomnia (CBT-I; Taylor et al., 2007). Sleep may also be a modifiable risk factor for seasonal and non-seasonal depression (Roecklein et al., 2013; Wescott et al., 2020).

In conclusion, this is the first report to our knowledge to demonstrate that retrospectively reported childhood maltreatment is associated with depression severity in SAD. The majority of past research on SAD etiology has focused on biological mechanisms related to environmental light levels, and only more recently, psychological theories (Rohan et al., 2009). Including a role for childhood maltreatment in the etiology of SAD may open new avenues of research and treatment.

Highlights.

  • Childhood maltreatment is associated with the severity of depression symptoms in Seasonal Affective Disorder (SAD).

  • Insomnia and hypersomnia are cross-sectional mediators of the association between childhood trauma and SAD.

  • Seasonal beliefs are a stronger cross-sectional mediator of the association between childhood trauma and SAD than typical depressogenic beliefs.

  • Rumination, specifically brooding, is a cross-sectional mediator of the association between childhood trauma and SAD.

ACKNOWLEDGEMENTS

We would like to thank Aidan G.C. Wright, Ph.D. for advice on statistical analyses.

Funding:

This work was supported by the National Institutes of Mental Health [grant numbers MH-103313, MH-096119]. The NIMH had no involvement in the collection, analysis and interpretation of data. Participant recruitment was supported by the National Institutes of Health through Grant Number UL1TR001857

Footnotes

CONFLICT OF INTERESTS

All authors declare that there are no conflict of interests.

DATA ACCESSIBILITY

Data analyzed in this study are available from the corresponding author.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Agargun MY, Kara H, Ozer OA, Selvi Y, Kiran U, Kiran S, 2003. Nightmares and dissociative experiences: The key role of childhood traumatic events. Psychiat Clin Neuros. 57, 139–145. 10.1046/j.1440-1819.2003.01093.x. [DOI] [PubMed] [Google Scholar]
  2. Anda RF, Felitti VJ, Bremner JD, Walker JD, Whitfield C, Perry BD, Dube SR, Giles WH, 2006. The enduring effects of abuse and related adverse experiences in childhood: convergence of evidence from neurobiological and epidemiology. Euro Arch Psy Clin N, 256, 174–186. 10.1007/s00406-005-0624-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bagby MR, Ryder AG, Schuller DR, Marshall MB, 2004. The Hamilton Depression Rating Scale: Has the Gold Standard Become a Lead Weight? Am J Psychiatry. 161, 2163–2177. 10.1176/appi.ajp.161.12.2163. [DOI] [PubMed] [Google Scholar]
  4. Baglioni C, Battagliese G, Bernd F, Spiegelhalder K, Nissen C, Voderholzer U, Lombardo C, Riemann D, 2011. Insomnia as a predictor of depression: A meta-analytic evaluation of longitudinal epidemiological studies. J Affect Disord. 135, 10–19. 10.1016/j.jad.2011.01.011. [DOI] [PubMed] [Google Scholar]
  5. Bellis MD, Kuchibhatla M, 2006. Cerebellar Volumes in Pediatric Maltreatment-Related Posttraumatic Stress Disorder. Biol Psychiatry. 60, 697–703. 10.1016/j.biopsych.2006.04.035. [DOI] [PubMed] [Google Scholar]
  6. Beratis S, Gourzis P, Gabriel J, 1994. Anniversary Reaction as Seasonal Mood Disorder. Psychopathology. 27, 14–18. 10.1159/000284843. [DOI] [PubMed] [Google Scholar]
  7. Bernet CZ, Stein MB, 1999. Relationship of childhood maltreatment to the onset and course of major depression in adulthood. Depress Anxiety. 9, 169–174. 10.1002/(SICI)1520-6394(1999)9:4&lt;169::AID-DA&gt;3.0.CO;2-2. [DOI] [PubMed] [Google Scholar]
  8. Bernier A, Belanger M, Bordeleau S, Carrier J, 2013. Mothers, fathers, and toddlers: Parental psychosocial functioning as a context for young children’s sleep. Dev Psychol. 49, 1375–1384. 10.1037/a0030024. [DOI] [PubMed] [Google Scholar]
  9. Bernstein DP, Fink L, Handelsman L, Foote J, Lovejoy M, Wenzel K, Sapareto E, Ruggiero J, 1994. Initial reliability and validity of a new retrospective measure of child abuse and neglect. Am J Psychiatry. 151, 1132–1136. 10.1176/ajp.151.8.1132. [DOI] [PubMed] [Google Scholar]
  10. Bernstein D, Fink L, 1998. Childhood Trauma Questionnaire: A retrospective self-report. San Antonio, TX: The Psychological Corporation. [Google Scholar]
  11. Blouin A, Blouin J, Aubin P, Carter J, Goldstein C, Boyer H, & Perez E, 1992. Seasonal patterns of bulimia nervosa. Am J Psychiatry. 149, 73–81. 10.1176/ajp.149.1.73 [DOI] [PubMed] [Google Scholar]
  12. Chapman DP, Liu Y, Presley-Cantrell L, Edwards VJ, Wheaton AG, Perry GS, Croft JB, 2013. Adverse childhood experiences and frequent insufficient sleep in 5 U.S. States, 2009: A retrospective cohort study. BMC Public Health. 13, 3. 10.1186/1471-2458-13-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cohen J, 1988. Statistical power analysis for the behavioral sciences, second ed. Academic Press, New York. [Google Scholar]
  14. Cohen PF, 2007. Anniversary Reactions in the Therapy Group. Int J Group Psychother. 57, 153–166. 10.1521/ijgp.2007.57.2.153. [DOI] [PubMed] [Google Scholar]
  15. Cuddihy C, Dorris L, Minnis H, Kocovska E, 2013. Sleep disturbance in adopted children with a history of maltreatment. Adopt Foster. 37, 404–411. 10.1177/0308575913508715. [DOI] [Google Scholar]
  16. Dalgleish T, Spinks H, Golden AM, du Toit P, 2004. Processing of emotional information in seasonal depression across different cognitive measures. J Abnormal Psychol. 113, 116–126. 10.1037/0021-843X.113.1.116. [DOI] [PubMed] [Google Scholar]
  17. Dozois DJA, Martin RA, Bieling PJ, 2009. Early maladaptive schemas and adaptive/maladaptive styles of humor. Cogn Ther Res. 33, 585–596. 10.1007/s10608-008-9223-9. [DOI] [Google Scholar]
  18. Enggasser JL, Young MA, 2007. Cognitive vulnerability to depression in seasonal affective disorder: Predicting mood and cognitive symptoms in individuals with seasonal vegetative changes. Cogn Ther Res. 31, 3–21. 10.1007/s10608-006-9076-z. [DOI] [Google Scholar]
  19. Fergusson DM, Boden JM, Horwood LJ, 2008. Exposure to childhood sexual and physical abuse and adjustment in early adulthood. Child Abuse Negl. 32, 607–619. 10.1016/j.chiabu.2006.12.018. [DOI] [PubMed] [Google Scholar]
  20. Girschik J, Fritschi L, Heyworth J, Waters F, 2012. Validation of Self-Reported Sleep Against Actigraphy. J Epidemiol. 22, 462–468. 10.2188/jea.JE20120012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Glod CA, Teicher MH, Hartman CR, Harakal T, 1997. Increased Nocturnal Activity and Impaired Sleep Maintenance in Abused Children. J Am Acad Child Psy. 36, 1236–1243. 10.1097/00004583-199709000-00016. [DOI] [PubMed] [Google Scholar]
  22. Gold DB, Wegner DM, 1995. Origins of ruminative thought: trauma, incompleteness, nondisclosure, and suppression. J Appl Soc Psychol. 25, 1245–1261. 10.111/j.1559-1816.1995.tb02617.x. [DOI] [Google Scholar]
  23. Golden AM, Dalgleish T, Spinks H, 2006. Dysfunctional attitudes in seasonal affective disorder. Behav Res Ther. 44, 1159–1164. 10.1016/j.brat.2005.09.004. [DOI] [PubMed] [Google Scholar]
  24. Harvey CJ, Gehrman P, Espie CA, 2014. Who is predisposed to insomnia: a review of familial aggregation, stress-reactivity, personality and coping style. Sleep Med Rev. 18, 237–247. 10.1016/j.smrv.2013.11.004. [DOI] [PubMed] [Google Scholar]
  25. Hayes AF, 2012. PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling. http://www.afhayes.com/public/process2012.pdf
  26. Hayes AF, 2013. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach, second ed. The Guilford Press, New York. [Google Scholar]
  27. Heim C, Binder EB, 2012. Current research trends in early life stress and depression: Review of human studies on sensitive periods, gene-environment interactions, and epigenetics. Exp Neurol. 233, 102–111. 10.1016/j.expneurol.2011.10.032. [DOI] [PubMed] [Google Scholar]
  28. Jacobson NS, Martell CR, Dimidjian S, 2001. Behavioral Activation Treatment for Depression: Returning to Contextual Roots. Clin Psychol Sci. 8, 255–270. 10.1093/clipsy.8.3.255. [DOI] [Google Scholar]
  29. James SP, Wehr TA, Sack DA, Parry BL, Rosenthal NE, 1985. Treatment of seasonal affective disorder with light in the evening. Brit J Psychiat. 147, 424–428. 10.1192/bjp.147.4.424. [DOI] [PubMed] [Google Scholar]
  30. Jonson-Reid M, Chance T, Drake B, 2007. Risk of Death Among Children Reported for Nonfatal Maltreatment. Child Maltreat. 12, 86–95. 10.1177/1077559506296722. [DOI] [PubMed] [Google Scholar]
  31. Kaplan KA, Harvey AG, 2009. Hypersomnia across mood disorders: A review and synthesis. Sleep Med Rev. 13, 275–285. 10.1016/j.smrv.2008.09.001. [DOI] [PubMed] [Google Scholar]
  32. Kasper S, Wehr TA, Bartko JJ, Gaist PA, Rosenthal NE, 1989. Epidemiological findings of seasonal changes in mood and behavior: A telephone survey of Montgomery County, Maryland. Arch Gen Psychiatry. 46, 823–833. 10.1001/archpsyc.1989.01810090065010. [DOI] [PubMed] [Google Scholar]
  33. Kendler KS, Kuhn JW, Prescott CA, 2004. Childhood sexual abuse, stressful life events and risk for major depression in women. Psychol Med. 34, 1475–1482. 10.1017/S003329170400265X. [DOI] [PubMed] [Google Scholar]
  34. Koskenvuo K, Hublin C, Partinen M, Paunio T, Koskenvuo M, 2010. Childhood adversities and quality of sleep in adulthood: A population-based study of 26,000 Finns. Sleep Med. 11, 17–22. 10.1016/j.sleep.2009.03.01. [DOI] [PubMed] [Google Scholar]
  35. Kubiak S, Kim WJ, Fedock G, Bybee D, 2012. Assessing short-term outcomes of an intervention for women convicted of violent crimes. J Soc Social Work Res. 3, 197–212. 10.5243/jsswr.2012.13. [DOI] [Google Scholar]
  36. Lam RW, Goldner EM, Grewal A, 1996. Seasonality of symptoms in anorexia and bulimia nervosa. Int J Eat Disord. 19, 35–44. 10.1002/(SICI)1098-108X(199601)19:1&lt;35::AID-EAT5&gt;3.0.CO;2-X. [DOI] [PubMed] [Google Scholar]
  37. Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ, 2008. Sleep duration: how well do self-reports reflect objective measures? The CARDIA Sleep Study. Epidemiology. 19, 838–845. 10.1097/EDE.0b013e318187a7b0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Levitan RD, Parikh SV, Lesage AD, Hegadoren KM, Adams M, Kennedy SH, Goering PN, 1998. Major Depression in Individuals With a History of Childhood Physical or Sexual Abuse: Relationship to Neurovegetative Features, Mania, and Gender. Am J Psychiatry. 155, 1746–1752. 10.1176/ajp.155.12.1746. [DOI] [PubMed] [Google Scholar]
  39. Lewis CC, Simons AD, Nguyen LJ, Murakami JL, Reid MW, Silva SG, March JS, 2010. Impact of childhood trauma on treatment outcome in the Treatment for Adolescents with Depression Study (TADS). J Am Acad Child Adolesc Psychiatry. 49, 132–140. 10.1016/j.jaac.2009.10.007. [DOI] [PubMed] [Google Scholar]
  40. Liu RT, 2017. Childhood Adversities and Depression in Adulthood: Current Findings and Future Directions. Clin Psychol Sci Pract. 24, 140–153. 10.1111/cpsp.12190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lynch SM, Heath NM, Mathews KC, Cepeda GJ, 2012. Seeking Safety: An Intervention for Trauma Exposed Incarcerated Women? J Trauma Dissociation. 13, 88–101. 10.1080/15299732.2011.608780. [DOI] [PubMed] [Google Scholar]
  42. MacKinnon D Introduction to statistical medication analysis. London, UK: Routledge, 2012. [Google Scholar]
  43. MacMillan HL, Fleming JE, Streiner DL, Lin E, Boyle MH, Jamieson E, Duku EK, Walsh CA, Wong MYY, Beardslee WR, 2001. Childhood Abuse and Lifetime Psychopathology in a Community Sample. Am J Psychiatry. 158, 1878–1883. 10.1176/appi.ajp.158.11.1878. [DOI] [PubMed] [Google Scholar]
  44. Magnusson A, 2000. An overview of epidemiological studies on seasonal affective disorder. Acta Psychiatr Scand. 101, 176–184. 10.1034/j.1600-0447.2000.101003176.x. [DOI] [PubMed] [Google Scholar]
  45. Mandelli L, Petrelli C, Serretti A, 2015. The role of specific early trauma in adult depression: A meta-analysis of published literature. Childhood trauma and adult depression. Euro Psychiatry. 30, 665–680. 10.1016/j.eurpsy.2015.04.007. [DOI] [PubMed] [Google Scholar]
  46. Matt GE, Vazquez C, Campbell EK, 1992. Mood-congruent recall of affectively toned stimuli: A meta-analytic review. Clin Psychol Rev. 12, 227–255. 10.1016/0272-7358(92)90116-P. [DOI] [Google Scholar]
  47. McKinney CM, Harris TR, Caetano R, 2009. Reliability of self-reported childhood physical abuse by adults and factors predictive of inconsistent reporting. Violence Vict. 24, 653–668. 10.1891/0886-6708.24.5.653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. McWhorter KL, Parks CG, D’Aloisio AA, Rojo-Wissar DM, Sandler DP, Jackson CL, 2019. Traumatic Childhood Experiences and Multiple Dimensions of Poor Sleep among Adult Women. Sleep. 42, 1–11. 10.1093/sleep/zsz108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Miranda J, Persons JB, 1988. Dysfunctional attitudes are mood-state dependent. J Abnorm Psychol. 97, 76–79. 10.1037/0021-843X.97.1.76. [DOI] [PubMed] [Google Scholar]
  50. Morin CM, Belleville G, Belanger L, Ivers H, 2011. The Insomnia Severity Index: Psychometric Indicators to Detect Insomnia Cases and Evaluate Treatment Response. Sleep. 34, 601–608. 10.1093/sleep/34.5.601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Nasca C, Watson-Lin K, Bigio B, Robakis TK, Myoraku A, Wroolie TE, McEwen BS, Rasgon N, 2019. Childhood trauma and insulin resistance in patients suffering from depressive disorders. Exp Neurol. 315, 15–20. 10.1016/j.expneurol.2019.01.005 [DOI] [PubMed] [Google Scholar]
  52. Nolen-Hoeksema S, Morrow J, 1991. A prospective study of depression and posttraumatic stress symptoms after a natural disaster: the 1989 Loma Prieta Earthquake. J Pers Soc Psychol. 61, 115–121. 10.1037/0022-3514.61.1.115. [DOI] [PubMed] [Google Scholar]
  53. Nolen-Hoeksema S, Wisco BE, Lyubomirsky S, 2008. Rethinking Rumination. Perspect Psychol Sci. 3, 400–424. 10.111/j.1745-6924.2008.00088.x. [DOI] [PubMed] [Google Scholar]
  54. Noll JG, Trickett PK, Susman EJ, Putnam FW, 2006. Sleep Disturbances and Childhood Sexual Abuse. J Pediatr Psychol. 31, 469–480. 10.1093/jpepsy/jsj040. [DOI] [PubMed] [Google Scholar]
  55. Pincus HA, Davis WW, McQueen LE, 1999. ‘Subthreshold’ mental disorders: A review and synthesis of studies on minor depression and other ‘brand names’. Br J Psychiatry. 174, 288–296. 10.1192/bjp.174.4.288. [DOI] [PubMed] [Google Scholar]
  56. Preacher KJ, Hayes AF, 2008. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods. 40, 879–891. 10.3758/BRM.40.3.879. [DOI] [PubMed] [Google Scholar]
  57. Raes F, Hermans D, 2008. On the mediating role of subtypes of rumination in the relationship between childhood emotional abuse and depressed mood: Brooding versus reflection. Depress Anxiety. 25, 1067–1070. 10.1002/da.20447. [DOI] [PubMed] [Google Scholar]
  58. Reyner A, Horne JA, 1995. Gender- and Age-Related Differences in Sleep Determined by Home-Recorded Sleep Logs and Actimetry From 400 Adults. Sleep. 18, 127–134. 10.1093/sleep/18.2.127. [DOI] [PubMed] [Google Scholar]
  59. Richter P, Werner J, Heerlein A, Kraus A, Sauer H, 1998. On the validity of the Beck Depression Inventory. Psychopathology. 31, 160–168. 10.1159/000066239. [DOI] [PubMed] [Google Scholar]
  60. Roecklein KA, Carney CE, Wong PM, Steiner JL, Hasler BP, Franzen PL, 2013. The role of beliefs and attitudes about sleep in seasonal and nonseasonal mood disorder, and nondepressed controls. J Affect Disord. 150, 466–473. 10.1016/j.jad.2013.04.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Rohan KJ, Sigmon ST, Dorhofer DM, 2003. Cognitive-behavioral factors in seasonal affective disorder. J Consult Clin Psychol. 71, 22–30. 10.1037/0022-006X.71.1.22. [DOI] [PubMed] [Google Scholar]
  62. Rohan KJ, Roecklein KA, Tierney Lindsey K, Johnson LG, Lippy RD, Lacy TJ, Barton FB, 2007. A randomized controlled trial of cognitive-behavioral therapy, light therapy, and their combination for seasonal affective disorder. J Consult Clin Psychol. 75, 489–500. 10.1037/0022-006X.75.3.489. [DOI] [PubMed] [Google Scholar]
  63. Rohan K, Roecklein K, Haaga D, 2009. Biological and Psychological Mechanisms of Seasonal Affective Disorder: A Review and Integration. Curr Psychiatr Rev. 5, 37–47. 10.2174/157340009787315299. [DOI] [Google Scholar]
  64. Rohan KJ, Meyerhoff J, Ho SY, Roecklein KA, Nillni YI, Hillhouse JJ, DeSarno MJ, Vacek PM, 2019. A measure of cognitions specific to seasonal depression: Development and validation of the Seasonal Beliefs Questionnaire. Psychol Assess. 31, 925–938. 10.1037/pas0000715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Rose DT, Abramson L, 1992. Developmental Predictors of Depressive Cognitive Style: Research and Theory, in: Cicchetti D, Toth SL (Eds.), Developmental Perspectives on Depression. University of Rochester Press, New York, pp. 323–348. [Google Scholar]
  66. Rosenthal NE, Sack DA, Gillin JC, Lewis AJ, Goodwin FK, Davenport Y, 1984. Seasonal affective disorder: A description of the syndrome and preliminary findings with light therapy. Arch Gen Psychiatry. 41, 72–80. 10.1001/archpsyc.1984.01790120076010 [DOI] [PubMed] [Google Scholar]
  67. Rutter M, 2006. The Psychological Effects of Early Institutional Rearing, in: Marshall PJ, Fox NA (Eds.), Series in affective science. The development of social engagement: Neurobiological perspectives. Oxford University Press, Oxford, pp. 355–391. [Google Scholar]
  68. Sachs-Ericsson N, Verona E, Joiner T, Preacher KJ, 2006. Parental verbal abuse and the mediating role of self-criticism in adult internalizing disorders. J Affect Disord. 93, 71–78. 10.1016/j.jad.2006.02.014. [DOI] [PubMed] [Google Scholar]
  69. Sobel ME, 1982. Asymptotic Confidence Intervals for Indirect Effects in Structural Equation Models. Sociol Methodol. 13, 290–312. 10.2307/270723. [DOI] [Google Scholar]
  70. Soehner AM, Kaplan KA, Harvey AG, 2014. Prevalence and clinical correlates of co-occurring insomnia and hypersomnia symptoms in depression. J Affect Disord. 167, 93–97. 10.1016/j.jad.2014.05.060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Spasojevic J, Alloy LB, 2002. Who becomes a depressive ruminator? Developmental antecedents of ruminative response style. J Cogn Psychother. 16, 405–419. 10.1891/jcop.16.4.405.52529. [DOI] [Google Scholar]
  72. Spitzer RL, Gibbon M, Williams JBW, 2001. Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Non-Patient Edition (SCID-I/NP). Biometrics Research, New York State Psychiatric Institute, New York. [Google Scholar]
  73. Steel J, Sanna L, Hammond B, Whipple J, Cross H, 2004. Psychological sequelae of childhood sexual abuse: abuse-related characteristics, coping strategies, and attributional style. Child Abuse Negl. 28, 785–801. 10.1016/j.chiabu.2003.12.004. [DOI] [PubMed] [Google Scholar]
  74. Taylor DJ, Lichstein KL, Weinstock J, Sanford S, Temple JR, 2007. A Pilot Study of Cognitive-Behavioral Therapy of Insomnia in People with Mild Depression. Behav Ther. 38, 49–57. 10.1016/j.beth.2006.04.002. [DOI] [PubMed] [Google Scholar]
  75. Terman M, Terman JS, Quitkin FM, Cooper TB, Lo ES, German JM, Stewart JW, McGrath PJ, 1988. Response of the melatonin cycle to phototherapy for seasonal affective disorder. J Neural Transm. 72, 147–165. 10.1007/BF01250238. [DOI] [PubMed] [Google Scholar]
  76. Treynor W, Gonzalez R, Nolen-Hoeksema S, 2003. Rumination reconsidered: A psychometric analysis. Cogn Ther Res. 27, 247–259. 10.1023/A:1023910315561. [DOI] [Google Scholar]
  77. Troxel WM, Kupfer DJ, Reynolds CF, Frank E, Thase M, Miewald J, Buysse DJ, 2012. Insomnia and Objectively Measured Sleep Disturbances Predict Treatment Outcome in Depressed Patients Treated with Psychotherapy or Psychotherapy-Pharmacotherapy Combinations. J Clin Psychiat. 73, 478–485. 10.4088/JCP.11m07184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Vegar Z, Hussain ME, 2017. Validity and reliability of insomnia severity index and its correlation with Pittsburgh sleep quality index in poor sleepers among Indian university students. Int J Adolesc Med Health. 10.1515/ijamh-2016-0090. [DOI] [PubMed] [Google Scholar]
  79. Warmingham JM, Handley ED, Rogosch FA, Manly JT, Cicchetti D, 2019. Identifying Maltreatment Subgroups with Patterns of Maltreatment Subtype and Chronicity: A Latent Class Analysis Approach. Child Abuse Negl. 87, 28–39. 10.1016/j.chiabu.2018.08.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  80. Warshaw C, Sullivan C, Rivera E, 2013. A systematic review of trauma-focused interventions for domestic violence survivors. National Center on Domestic Violence, Trauma and Mental Health. https://doi.org/20.500.11990/536. [Google Scholar]
  81. Wells TT, Vanderlin WM, Selby EA, Beevers CG, 2014. Childhood abuse and vulnerability to depression: Cognitive scars in otherwise healthy young adults. Cogn Emot. 28, 821–833. 10.1080/02699931.2013.864258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  82. Weissman AN, Beck AT, 1978. Development and validation of the Dysfunctional Attitude Scale: A preliminary investigation. Paper presented at the meeting of the Association of the Advancement of Behavior Therapy, Chicago. [Google Scholar]
  83. Wescott DL, Soehner AM, Roecklein KA, 2020. Sleep in Seasonal Affective Disorder. Curr Opin Psychol 34, 7–11. 10.1016/j.copsyc.2019.08.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Williams JB, Link MJ, Rosenthal NE, Amira L, Terman M, 1994. Structured interview guide for the Hamilton Rating Scale of depression—seasonal affective disorder version (SIGH-SAD). New York State Psychiatric Institute, New York [Google Scholar]
  85. Young JE, Klosko JS, Weishaar ME, 2003. Schema Therapy: A Practitioner’s Guide. The Guilford Press, New York. [Google Scholar]
  86. Young MA, Reardon A, Azam O, 2008. Rumination and vegetative symptoms: A test of the dual vulnerability model of seasonal depression. Cogn Ther Res. 32, 567–576. 10.1007/s10608-008-9184-z. [DOI] [Google Scholar]

RESOURCES