Moderating effects of depressive symptoms on the relationship between problematic use of the Internet and sleep problems in Korean adolescents

The prevalence of sleep problems in adolescents is very substantial; the prevalence of insomnia in adolescents is as high as 20~ 30% [1, 2], and approximately 16% of high school students suffer from excessive daytime sleepiness [3]. Sleep problems during adolescence may result from both endogenous and exogenous factors and some endogenous reasons for sleep problems are very common in adolescents, such as primary insomnia, insomnia related to depression and delayed-phase sleep patterns, which occur according to changes in the biological clock [4, 5]. Some sleep problems in adolescents, especially short sleep duration, are commonly induced by external factors, such as both increased academic pressure and electronic device use time including the Internet [6]. Among the various factors known for their relationship with adolescent sleep problems, depression and problematic Internet use (PIU) have received considerable attention.

It has been reported that the incidence of depression is increased in adolescents who sleep less than 6 h per day [7], and regardless of sleeping hours, impairment of sleep quality itself is related to depression, excessive daytime sleepiness and decline in academic performance [8]. Furthermore, it is known that inadequate sleep and impairment of sleep quality can affect not only the severity of depressive symptoms but also suicide ideation [9]. Not only can sleep problems cause depression, but it also true that depression can cause sleep problems [10, 11]. Approximately, 75% of individuals with depression suffer from a lack of sleep and poor sleep quality [12], and the typical sleep pattern of individuals with depression is characterized by decreased slow -wave sleep, short rapid eye movement (REM) sleep latency, long duration of first REM sleep, and increased density of REM sleep [13]. A total of 72% of children with major depressive disorder have sleep disturbances, 53.5% have insomnia alone, and 10.1% have both disturbances [14].

It is known that PIU is related to sleep problems such as reduce sleeping time [15, 16], the tendency to sleep later [17], sleep disturbance [18], insomnia [19], an increased level of tiredness [20], and excessive daytime sleepiness and poor school performance [21‐25]. Furthermore, exposure to the bright light of a screen until late at night may suppress melatonin secretion and sleep-wake phase-delay, and may increase alertness [26, 27], which can cause sleep induction problems [28, 29]. Many terms have been used to describe the phenomenon characterized by an individual’s inability to control their Internet use leading to marked distress and/or functional impairment [21] such as PIU, compulsive Internet use, Internet addiction, Internet dependence, pathological Internet use and so on. Among these concepts, we decided to use “PIU” in this paper, because we wanted to explore more general and common Internet- use problem rather than the minimum numbers of extreme pathologic Internet usage. It is also well known that compared with non-depressed adolescents, depressive adolescents show a high prevalence of PIU and vice versa [30] .

Interestingly, around the world South Korean adolescents are notorious for a prevalence of depression and PIU and the shortest sleep duration on weekdays [31]. Therefore, we aimed to examine whether depressive symptoms moderate the effect of PIU on sleep in South Korean adolescents, a group that is very characteristic and famous for a high prevalence of these three problems. Thus, we hypothesized that PIU acts differently on sleep between adolescents in the depressed group and non-depressed groups.

Among the total 766 participants who completed the questionnaires, 152 were categorized as PIUG and 614 as NIUG. Table 1 shows the differences in demographic and clinical characteristics between the NIUG and PIUG. Compared with the NIUG, the PIUG had significantly higher ISI, ESS, and SSS_ Sleep-Wake Behavior Problems Scale scores (p < 0.0001, p < 0.0001, and p = 0.001, respectively). Compared with the NIUG, the PIUG had a significantly lower Morningness-Eveningness scale total score (p = 0.021), and the ratio of morning types, as classified by their total Morningness-Eveningness scale scores, was significantly lower in the PIUG than in the NIUG (25.57% VS 15.13%, p = 0.007). The mean bedtime of the PIUG was significantly later than that of the NIUG, both on weekdays (p = 0.006) (Table 1). The reason for the later bedtime in the PIUG is that 3.3% of the NIUG and 11.3% of the PIUG went to bed late due to computer or smartphone use (p = 0.005) on weekdays and 39.1% of the NIUG and 45.3% of the PIUG went to bed late due to computer or smartphone use (p = 0.189) on weekends. Overall, 62.6% of the NIUG and 82.0% of the PIUG answered that they use smartphones or computers immediately before falling asleep (p < 0.0001). The mean weekend wake up time was significantly later in the PIUG than in the NIUG (p = 0.001). However, although the mean weekday wake up time tended to be later in the PIUG, no significant difference was detected between the two groups (p = 0.757).

Additionally, compared with the NIUG, the PIUG showed significantly higher CDI score (p = 0.001). The depressed participants (CDI score equal to greater than 16) had higher ISI scores (9.90 ± 4.44 vs 8.69 ± 3.92, p = 0.003) and lower Morningness-Eveningness scale scores (24.12 ± 4.25 vs 26.28 ± 3.36, p < 0.001) than the non-depressed participants. The ratio of evening to morning types was significantly higher in the depressed group (73.19%) than in the non-depressed group (38.32%, p < 0.0001).

Among the 152 individuals in the PIUG, 109 (71.74%) were classified as the depressed group in the PIUG. No differences were found in sleep-related scores (ISI, ESS, SSS_ Sleep-Wake Behavior Problems Scale, and Morningness-Eveningness scale scores) and sleep-related habits (mean wakeup time, mean bedtime and total sleep duration on both weekdays and weekends) between the depressed and non-depressed groups in the PIUG.

Additionally, we examined the moderating effect of depressive symptoms on sleep–related problems (insomnia, excessive daytime sleepiness, and sleep-wake behavior problems) with PIU using the Baron and Kenny method. In step 1, we conducted a regression analysis in which we analyzed the effects of the IAS scores and whether the individuals were included in the depressed group. Then, in step 2, we input the interaction variable between the IAS scores and whether the individuals were included in the depressed group back into the step 1 analysis. The step 1 analysis revealed that PIU did not have an effect on the SSS_Sleep-Wake Behavior Problems Scale (p > 0.05) or ISI (p > 0.05). However, in step 2, we found a moderating effect of depressive symptoms on the SSS_ Sleep-Wake Behavior Problem Scale and the ISI scores (Fig. 1). In the depressed group, the SSS_ Sleep-Wake Behavior Problems Scale and ISI scores were not increased even with increasing IAS scores, but in the non-depressed group, the SSS_Sleep-Wake Behavior Problem Scale and ISI scores increased with increasing IAS scores. The effect of IAS scores on ESS scores was different, depending on whether the individuals were depressed. In the depressed group, the ESS score was decreased with increasing IAS scores, but in the non-depressed group, the ESS score was increased with increasing IAS scores.

The results of this study show that compared with the NIUG, PIUG members were more prone to insomnia, excessive daytime sleepiness and sleep-wake behavior problems. The PIUG also tended to include more evening types than the NIUG. Interestingly, the effect of Internet use problems on sleep problems appeared to be different according to the presence or absence of depressive symptoms.

PIU is associated with poorer sleep in nondepressed adolescents but not in depressed adolescents. In detail, the effect of PIU on sleep-wake behavior problems, insomnia and excessive daytime sleepiness increased with increasing IAS scores in the non-depressed group. However, in the depressed group, the effects of PIU on sleep-wake behavior problems and insomnia did not change with increasing Internet use problems, and the effect of Internet use problems on excessive daytime sleepiness was relatively decreased with increasing Internet use problems in the depressed group. This result suggests that when a person has depressive symptoms and PIU at the same time, the two factors might act differently rather than exhibit additive, mutually reinforcing effects. However, we failed to identify the underlying causes of these patterns due to the cross-sectional nature of this study and the lack of observable differences in sleep problems between the depressed and non-depressed groups in the PIUG. A previous study that examined the effect of insomnia and Internet addiction on depression showed that both Internet addiction and insomnia affect depression, but exert distinct effects [19]; when insomnia is an explanatory factor, insomnia strengthens the negative effects of Internet addiction on depression, and when Internet addiction is an explanatory factor, Internet addiction strengthens the negative effects of insomnia on depression.

The Morningness-Eveningness type is well-recognized as a reflection of the endogenous biological clock. Morning types prefer performing intellectual and physical activities during the morning, and evening types prefer performing activities during the late afternoon or evening [43]. Compared to morning types, evening types tend to subjectively feel that their sleep time is not enough during weekdays and have increased sleep time during the weekend due to sleep debt accumulated through the weekdays [43]. Compared to morning types, evening types tend to report subjectively poor sleep quality and excessive daytime sleepiness [44]. Morning types tend to have healthier life-styles than evening types [46], and evening types tend to have higher depression rates than morning types [47]. In this study, the PIUG tended to include more evening types than the NIUG, and the depressed group also tended to include more evening types than the non-depressed group. However, interestingly, within the PIUG group, there was no difference in the Morningness-Eveningness type between the depressed and the non-depressed groups. Generally, it is expected that self-induced sleep restriction in adolescents would aggravate the changes in the sleep-wake cycle and therefore cause delayed sleep phases and lack of enough sleep [48]. However, this study result suggests that if an individual adolescent has clinical-range PIU and depressive symptoms, depressive symptoms and PIU act independently of sleep problems rather than mutually initiating a vicious cycle.

During adolescence, there is a dramatic change in sleep as a normal developmental process. As adolescents get older, they tend to have delayed sleep-wake phase, less sleep time and sleep architectural changes, such as declines in slow-wave sleep and slow-wave activity [48]. While the large part of these sleep changes across adolescence are shown to be a consequence of normal brain development in adolescence, more prevalent causes of sleep problems, especially short sleep duration, resulted from exogenous reasons in South Korean adolescents [49, 50]. South Korean adolescents are famous for their short sleep duration around the world [51, 52]. National statistics show that the average sleep duration of high school students is as short as 5 h 27 min in Korea [53]. Strong cultural emphasis on academic performance in South Korea and the high rate of PIU among Korean adolescents (12.5%) [54] are two strong exogenous reasons for this short sleep duration [6]. According to an international survey, Korean adolescents recognize themselves as the most unhappy adolescents in the world and also have a high prevalence of depression [55].

Sleep is essential for mental and physical development and daily function in adolescents [7, 56‐58]. Therefore, it is very important for adolescents to secure adequate sleep time and quality. Although understanding the underlying causes of adolescent sleep problems is critical for solving the problems, it is not easy to identify the cause of individual adolescent sleep problems because adolescent sleep problems may arise from the interactions of various factors, such as PIU, depression and normal physiological and hormonal changes. Therefore, the collaboration of experts in related areas is critical for exploring the causes in a more multi-dimensional way.

The limitations of this study include the following. 1. Due to the cross-sectional study design, a causal relationship cannot be examined. Therefore, caution is needed when interpreting or generalizing the current findings. 2. PIU and depressive symptoms are unlikely to be the only two explanatory factors of sleep disturbance. Other covariates may include the sleep environment, anxiety, dietary ingestion, etc., and those factors were not evaluated.

This study demonstrated that the effect of PIU on sleep presented differently between the depressed and non-depressed groups. PIU is associated with poorer sleep in non-depressed adolescents but not in depressed adolescents. This finding might be observed because PIU may be the biggest contributor to sleep problems in the problematic Internet user without depression, but in the problematic Internet user with depression, depression might be a more important contributor to sleep problems; thus, the influence of PIU on sleep effect might be diluted.