Target population was healthy adults aged more than 19 years. We recruited participants through advertisements posted at Severance Hospital and Yonsei University College of Medicine. Exclusion criteria were as below; current or previous major psychiatric disorder (e.g., schizophrenia or related psychotic disorder, bipolar disorder, major depressive disorder, and anxiety disorder), substance use disorder (except for nicotine), uncontrolled medical condition, and regular use of sleeping medication. All of the participants had given explanation of the research purposes and procedure beforehand, and provided written consent voluntarily. The study was approved by the Institutional Review Board of Severance Hospital, Yonsei University Health System (no: 4-2017-0893 and 4-2016-0570).

Baseline characteristics including sex, age, job status (regular worker, part-time job, housemaker, student, or inoccupation), education level (years), presence of current psychiatric disorder, and other medical condition were obtained through face-to-face interview. Presence of psychiatric disorder was evaluated using the Structured clinical interview for DSM-5 (SCID-5) [17].

For measuring baseline sleep status, Pittsburgh Sleep Quality Index (PSQI) and Insomnia Severity Index (ISI) were used. PSQI is a self-rated questionnaire that assesses sleep quality and disturbances during past month [18]. It consists of 7 components including subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction based on experience within 1 month of patients’ experience. Each component scored from 0 point to 3 point, and the sum of scores for these seven components yields one global score for PSQI. Higher PSQI score implies worse sleep status. Cut-off score of PSQI is generally considered as 5, which implies poor sleep quality [18,19], while others suggest a cut-off score of 8 is more suitable depending on the sample characteristics [20,21]. ISI is a 7-item self-report questionnaire evaluating below items; severity of sleep-onset, sleep maintenance, early morning awakening problems, satisfaction with current sleep pattern, interference with daily functioning, noticeability of impairment attributed to the sleep problem, and level of distress caused by sleep problem [22]. Each of these items is rated on a five-point Likert scale (‘0’ not at all, ‘4’=extremely) based on experience within two weeks. High scores indicating greater insomnia severity. Cut-off score of ISI is 8, which suggesting sub-threshold insomnia [23]. In this study, if the participants’ PSQI score was 8 or higher regardless of ISI score, or ISI score was 8 or higher regardless of PSQI score, all were defined as poor sleepers. If participants have both PSQI score lower than 8 and ISI score lower than 8, they were defined as good sleepers.

After their initial meeting with researchers, participants were asked to continuously wear the actigraphy device (ActiGraph GT3X+^®, ActiGraph, Pensacola, FL, USA) on their non-dominant wrist for 14 days, and to write sleep diary during same periods. If they did not wear actigraphy or write sleep diary for more than 7 days, they were excluded from the study. After the experiment, the participants were asked to conduct a survey on self-consciousness through e-mail.

Actigraphy was used to gather information of objective sleep parameters including objective sleep onset latency (SOL), objective total sleep time (TST), objective time in bed (TIB), and objective wake after sleep onset (WASO). Objective sleep efficiency was defined as ratio of objective TST per objective TIB. Actigraphy detects participants’ movement as activity counts in every 60 second (1 epoch). All epochs were scored as either sleep or wake status by comparing the calculated total activity counts to awake threshold value based on Cole-Kripke algorithm [24]. If the total activity counts were less than or equal to the wake threshold value, the epoch was scored as sleep. ActiLife software (version 6.13.3, ActiGraph) was used to assess and analyze data from actigraphy.

International consensus format of sleep diary was used to measure subjective sleep parameters to assess subjective SOL, subjective TST, subjective TIB, subjective WASO, subjective sleep quality, consumption of alcohol, and number and time of naps [25]. Subjective sleep efficiency was defined as ratio of subjective TST per subjective TIB. We encouraged participants to complete the diary immediately upon awakening each morning to minimize errors.

Self-consciousness scale (SCS) was used to evaluate self-consciousness of participants [26]. It is self-rated questionnaire consisted with 23 questions assessing three domains of self-consciousness; private self-consciousness (nine items, e.g., “I’m always trying to figure myself out.”), public self-consciousness (seven items, e.g., “I’m concerned about the way I present myself.”), and social anxiety (six items, e.g., “I have trouble working when someone is watching me.”) [13]. The private self-consciousness means one’s concern with attending to their inner thoughts and feelings. The public self-consciousness is defined by a general awareness of the self as a social object that has an effect on others. The third domain, social anxiety, is defined by a discomfort in the presence of others. Higher score indicates higher self-consciousness.

A total of 106 individuals were enrolled to the study. Data from 16 participants were excluded from analysis for the following reasons; consent withdrawal (n=1), without use of actigraphy for more than 7 days (n=1), without use of sleep diary for more than 7 days (n=2), incompletion of self-questionnaire (n=12). Therefore, data from a total of 90 people were used for the analysis.

To evaluate the effects of self-consciousness on insomnia symptoms, we performed independent sample t-test, binary regression analysis, and multiple logistic regression analysis. Independent sample t-tests were used to compare differences between poor sleepers (PSQI≥8 or ISI≥8) and good sleepers (PSQI<8 and ISI<8) including demographic factors, subjective and objective sleep parameters, and self-consciousness. Binary regression analysis was conducted to evaluate the effects of factors, including self-consciousness, on poor sleepers. Multiple logistic regression analysis was used to evaluate the effects of factors on each insomnia scale. The presence of multicollinearity among variables in regression model was examined by variance inflation factor, and no factors were found to have multicollinearity problem. Statistical significance was set at p<0.05. All analyses were performed by using statistical package for social science 23 for windows version (IBM Corp., Armonk, NY, USA) [27].

A total of 90 participant including poor sleepers (n=30) and good sleepers (n=60) were included in the study. Demographic and clinical characteristics of the 90 study participants (aged 20 to 61 years) are shown in Table 1. Subjective sleep quality was worse in poor sleepers than in good sleepers (p<0.001). There was a tendency of less subjective TST in poor sleepers than in good sleepers (p=0.050), whereas other sleep parameters including objective sleep parameters measured by actigraphy were not significantly different between two groups. Among the three domains of self-consciousness, only social anxiety was significantly higher in poor sleepers than in good sleepers (p=0.010).

Higher social anxiety was significantly associated with poor sleepers in binary regression analysis [odds ratio (OR)=1.091, p= 0.046] (Table 2). As shown in Table 3, we found that only social anxiety had significant effect on the poor sleepers with PSQI score of 8 or higher in multiple regression analysis (B=0.091, p=0.002). On the other hand, both private self-consciousness and social anxiety showed significant effects on the poor sleepers with the ISI score of 8 or higher (B=-0.202, p=0.043, B=0.087, p=0.047, respectively).