Although direct comparisons are difficult due to differences in study design and exposure, our novel findings are consistent with the emerging literature indicating a prospective association between shortened sleep duration and depression in similarly aged populations [
19‐
22]. A previous longitudinal study with three years of sleep and depression data used cross-domain latent growth models to show that students with more total sleep time reported fewer depressive symptoms (β = − 0.20,
p < 0.001) [
19]. Another five-year study used three waves of actigraph data and cross-lagged panel models to demonstrate that increased sleep duration predicted reduced depression (β = − 0.29,
p < 0.01) and other mental health outcomes [
20]. Both longitudinal studies measured depression using the Children’s Depressive Inventory in children under 14 years old [
19,
20], and sample sizes ranged from 100 [
20] to about 2300 [
19]. Two other one-year longitudinal studies assessed sleep deprivation (defined as ≤6 h / night) and risk of depression measured by NIMH Diagnostic Interview Schedule for Children Version IV [
21,
22]. Using the Teen Health 2000 cohort of 3134 adolescents aged 11–17 years at follow-up, researchers controlled for multiple confounders and reported that sleep deprivation on weeknights and weekends was linked to a 41% higher odds of depressed mood a year later; and, that sleep deprivation on weeknights alone showed a 23% higher odds of depressed mood [
22]. In a similar study that adjusted for baseline depression, sleep deprivation on weeknights was more strongly associated with major depression (adjusted OR 3.76 [1.65–8.58]) than with depressive symptoms (1.38 [1.02–1.85]) at one-year follow-up [
21]. Roberts and colleagues also found that the direction of association differed for each depression outcome: reduced quantity of sleep increased the risk for major depression which in turn increases the risk for decreased sleep, but sleep deprivation predicted symptoms of depression and not the reverse [
21]. Recently, a very small Australian prospective study, comprising mostly young men (
n = 138, 64% male), found that total weeknight sleep time at baseline was not associated with CESD depression scores one year later, [
34] which concurs with our null finding for young men and highlights the importance of incorporating a gender perspective in this area of research.
There are several interacting factors that might mediate the relationship between chronic sleep deprivation and greater subsequent depressive symptomatology in young women, ranging from biological development to the environmental context. The transition into adolescence has been identified as a period of increased risk for the development of both depression and altered sleeping patterns [
13]. Both neurological and hormonal changes associated with puberty not only decrease the total duration of sleep, but also shift the timing and stage of sleep from day to evening chronotype and from deep to lighter stages [
13]. Sleep deprivation alone causes the physiological stress pathway (i.e., hypothalamic-pituitary-adrenal axis) to be hyperactive [
35], and females are known to be more reactive to stress than males due to the differential effect of testosterone on the hypothalamic-pituitary-adrenal axis [
36]. Simultaneously, puberty is associated with cognitive and emotional changes whereby adolescents become more responsive to socioemotional contexts at a time when neural mechanisms for executive functioning and impulse inhibition have yet to fully develop [
37]. This differential brain development could influence young peoples’ experiences of depression symptomology [
35‐
37], with young women typically reporting earlier onset and higher levels of depression than young men [
7‐
9]. Overlaying these developmental changes, there are a number of environmental factors as well as other health behaviours that contribute to the sleep-depression paradigm such as early school start times, extensive use of electronic devices, lack of physical activity (which is higher among young women), caffeine and/or alcohol consumption, among others [
23].