Relationship Between Circadian Preference and Lifetime Experiences of Hypomanic Symptoms

BACKGROUND: Circadian rhythms have been known to be associated with bipolar disorders. There are many cases in which hypomanic symptoms are not recognized as an illness. This study aimed to determine the correlation between the experience of lifetime hypomanic symptom and circadian type of university students. METHODS: A total of 8,562 university students participated in the study. The participants completed the Composite Scale of Morningness (CSM) for circadian typology and Mood Disorder Questionnaire (MDQ). Chi-square test and analysis of variance were performed, and the post-hoc result was computed using the FDR adjusted p-values. RESULTS: Overall, the MDQ score was high in the evening-type group. There was no signicant difference between the intermediate-type group and morning-type group for male students. In the evening-type group, the response rate was signicantly higher in 10 out of 13 items in MDQ. CONCLUSIONS: The evening-type group was more likely to experience hypomanic symptoms. This study shows that circadian and seasonal characteristics related to circadian types are associated with lifetime hypomanic symptoms. Hence, further investigation is needed to determine the eveningness, as it could be a trait marker of bipolar spectrum disorder.


Background
Circadian preference refers to the temporal characteristic of individuals' activity rhythm in relation to the sleep-wake cycle. It is a circadian factor that determines whether someone is a morning-type or an eveningtype person. Based on individuals' preferred time of the day for activity or sleep, those whose rhythm is set earlier in the day are classi ed as morning-type, and those whose rhythm is set later in the day are classi ed as evening-type (Hasler et al. 2010). In general, circadian preference is affected by social and environmental factors, such as the preference of family members (Horne et al. 1975). However, it is also determined by physiological factors such as body temperature, changes in melatonin secretion based on exposure to sunlight, and changes in cortisol levels (Horne andOstberg 1975, Tsujimoto et al. 1990, Kerkhof et al. 1996).
Generally speaking, morning-type people go to bed about two hours earlier than evening-type people, wake up earlier, and demonstrate higher work e ciency in the morning. They are also more active in the daytime than in the evening (Mecacci et al. 1983). Meanwhile, evening-type people have the opposite characteristic.
They have a lower quality of sleep compared to morning-type people and often experience drowsiness in the daytime, which makes them more reliant on caffeine (Barclay et al. 2010, Roepke et al. 2010).
According to previous research on circadian preference, eveningness is highly associated with mood disorders, especially depressive disorder (Chelminski et al. 1999, Johansson et al. 2003, Ahn et al. 2008, Hasler, Allen et al. 2010. A study that looked at the relationship between circadian preference and the level of hypomanic symptoms among university students revealed that those with a more de nite preference for eveningness tended to have a lower satisfaction of life and higher rates of bipolarity (Park et al. 2015).
Unlike clinically stable schizophrenia patients, who are known to show no difference with non-patients when it comes to circadian preference, clinically stable bipolar disorder patients have been observed to prefer eveningness and experience delayed sleep cycle more often than non-patients (Ahn, Chang et al. 2008). A study on depressive disorder also revealed that a stronger preference for eveningness is associated with depressive symptoms (Ahn, Chang et al. 2008, Hasler, Allen et al. 2010). Bipolar disorder, which is known to be highly associated with circadian preference, is classi ed into two types: bipolar I and bipolar II. Because the disorder encompasses a wide spectrum of symptoms ranging from psychotic mania to brief hypomania, it is highly likely that its association with circadian preference could be observed as a trait, not only in a clinical group but also in a non-clinical group (Judd et al. 2003).
This study aimed to identify the relationship between hypomanic symptoms experienced by a non-patient group-as opposed to a patient group-and circadian preference, in order to understand whether such characteristics could be considered a predisposition to bipolar disorder.

Participants
This study is based on part of the survey results conducted by an in rmary at Kongju National University for an educational capacity-building project. The participants were informed that the information they provide would be kept strictly con dential and used for research purposes only, and written consent was obtained. Out of the 8,760 participants, the responses of 8,561 participants were analyzed as 199 had left at least one item unanswered. The study was approved by the Research Ethics Committee of Kongju National University.

Measurements
Composite scale for morningness-eveningness (CSM) CSM is a self-report questionnaire developed by Smith et al., in order to determine individuals' circadian preference (Smith et al. 1989). The questionnaire consists of 13 items, and the total score ranges from 13 to 55. In this study, the Korean translation of CSM was used (Yoon et al. 1997), and respondents with scores in the top 20% were classi ed as morning-type, while those with scores in the bottom 20% were classi ed as evening-type (Prat et al. 2013).
The most common method used for categorizing circadian rhythm types is to focus on individuals' sleepwake cycle phases and classify them into the so-called morning-type or evening-type groups. This study used the Korean translation of CSM, a self-report questionnaire developed by Smith et al. The questionnaire consists of 13 items, and the total score ranges from 13 to 55 as three items are scaled from 1 to 5, and ten items are scaled from 1 to 4. In general, respondents with higher scores are classi ed as morning-type, those with lower scores are classi ed as evening-type, and those with scores in-between are classi ed as intermediate-type. However, in this study, respondents with a score in the bottom 20% were identi ed as evening-type, and those with a score in the top 20% were identi ed as morning-type (Prat and Adan 2013). Thus, those with a score of 26 or lower were classi ed into the evening-type group, those with a score of 41 or higher were classi ed into the morning-type group, and those with a score in-between were classi ed into the intermediate-type group. (Randler 2008) Mood Disorder Questionnaire (MDQ) MDQ is a 3-part self-report questionnaire used to screen for Bipolar Spectrum Disorder (BSD). Part 1 is composed of 13 items that look into whether the respondents experienced manic or hypomanic episodes in their lifetime , Miller et al. 2004), part 2 asks whether those manic or hypomanic episodes or behaviors were demonstrated during the same period, and part 3 assesses on a 4-point scale whether the respondents had any functional impairment due to the illness. This study used the total score from part 1 of the Korean version of the MDQ (K-MDQ) (Jon et al. 2009), whose reliability and validity have been previously con rmed.
Patient Health Questionnaire-9 (PHQ-9) PHQ-9 is a screening instrument for depression. The nine items in the questionnaire that evaluate depressive symptoms to diagnose major depressive disorder have been developed to match the nine diagnostic criteria for major depressive episode in the DSM-IV. Each item is scaled from 0 to 3, and the total score ranges from 0 to 27. This study used the Korean version of the PHQ-9 translated in 2010, whose reliability and validity have been established (Park et al. 2010).

Statistical Analysis
Based on the nature of each data, continuous variables were denoted as mean (± standard deviation), and categorical variables were denoted as N (%). The analysis was conducted to compare the gender difference in bipolarity among the participants who were classi ed into the morning-type, intermediate-type, and evening-type groups according to their CSM scores. X 2 test was employed to compare which participants from each group gave positive responses to the respective MDQ items, and to adjust for post-hoc analysis and repeated measurements, the False Discovery Rate (FDR) method was used for comparison. Moreover, in order to identify the relationship between circadian preference and bipolarity, multinominal logistic regression was conducted with circadian preference as a dependent variable, and the total MDQ score (excluding items 3, 8, and 9 where the three groups showed no signi cant difference) as an independent variable. SPSS 24.0 for Windows (IBM Inc., NY, USA) was used for data analysis, and the signi cance level for each analysis was set at 0.05 on both sides. In addition, differences in the groups by circadian preference were compared with Chi-square tests with adjusted P-values for multiple comparisons from the R package fer (Fife et al. 2017).

Results
Among the 8,562 participants, 4,318 (50.4%) were female students and 4,244 (49.6%) were male students. There were relatively more evening-types among female students (1,101 students, 25.5%), and relatively more morning-types among male students (1,181 students, 27.8%, Table 1). The total MDQ score was signi cantly higher among female students (t = 9.366, p < 0.001). For male students, evening-types had signi cantly higher MDQ scores (F = 16.786, p < 0.001), and among female students, the MDQ score was signi cantly higher in the order of evening-type, intermediate-type, and morning-type (F = 16.668, p < 0.001, Table 2). For the 10 out of 13 MDQ items (items 1, 2, 4, 5, 6, 7, 10, 11, 12, 13), the rate of positive response was higher in the order of evening-type, intermediate-type, and morning-type, and for the remaining three items (items 3,8,9), there was no signi cant difference between the three groups ( Table 3, Fig. 1). In the regression analysis, where the total MDQ score (calculated without the three items above) was designated as an independent variable, it was found that even if the variables are adjusted for gender and depression, the evening-type group demonstrates signi cantly higher rates of bipolarity (1.035. 95% CI = 1.011-1.059).
In contrast, while the morning-type group demonstrated relatively lesser degrees of bipolarity, the results were not signi cant when adjusted for depression (Table 4).

Discussion
This study looked at the relationship between circadian preference and a lifetime history of hypomanic symptoms in a non-patient demographic. Over 8,000 university students' responses to the 13 MDQ items on experiences of lifetime hypomanic symptoms, as well as their circadian preference according to the Composite Scale of Morningness (CSM), were evaluated, and the results revealed that the evening-type group experiences more signi cant hypomanic symptoms.
When the morning-type, intermediate-type, and evening-type groups' responses to the MDQ were compared, the rate of positive response to hypomanic symptoms was higher among the evening-type group for 10 out of 13 items. In contrast, there was no signi cant difference between the groups for three items-more selfcon dence (item 3), more energy (item 8), and more activity (item 9). Such result is attributable to one of the changes made to the DSM-5 from the DSM-IV, where increased activity and energy were added to criterion A for manic and hypomanic episodes on top of changes in mood. The revision re ects the view that changes in the level of activity and energy are the core symptoms of bipolar disorder and is based on previous ndings that patients with bipolar disorder recall changes in the level of activity and energy more readily and accurately than changes in mood (Geoffroy et al. 2013). Thus, it is natural for the three groups to demonstrate no signi cant difference in terms of those core symptoms, as the participants were part of a non-clinical group. This, in turn, signi es that the changes in the level of energy and activity could be considered a clinical trait that could be used to identify patients with clinical bipolar disorder more sensitively. Meanwhile, item 3 of the MDQ-more self-con dence-was already found to be of less use in a previous validation study when it comes to differentiating the study group from the control group (Jon et al. 2005).
A long-term follow-up study of university students with hypomanic personalities has shown that the subject group is more likely to develop bipolar disorder and major depressive disorder than the control group (Kwapil et al. 2000). However, in this study, only 25% of the participants with hypomanic personalities were identi ed with bipolar disorder in a follow-up observation (Kwapil, Miller et al. 2000), and it is known that a signi cant number of people with personalities associated with bipolarity do not experience mood disorder (Akiskal 1992). This implies that individuals with bipolarity may continue to maintain those tendencies, rather than developing bipolar disorder over time. Moreover, it suggests that bipolarity could be understood as part of a spectrum expressing individuals' behaviors or emotional characteristics, rather than a disorder that belongs to the far ends of the bipolar disorder spectrum.
Such clinical characteristics are being con rmed in other cohort studies using the MDQ. For those who screened positive on the MDQ, there was a higher frequency of migraine and drug abuse , and among the patients that visited primary care, those who screened positive on the MDQ complained more of physical symptoms and were taking a higher number of drugs (Gorski et al. 2003).
Also, 59% of the patients who have bromyalgia screened positive on the MDQ (Carta et al. 2006), and among those aged 35 and younger, people who screened positive on the MDQ were observed to experience more anxiety and depression (Gorski et al. 2006).
These ndings suggest that MDQ scores may have the potential to re ect non-speci c psychological vulnerabilities rather than representing the clinical category of bipolar disorder. One of such psychological vulnerabilities could be the circadian preference of eveningness.
This study showed that the tendency for individuals in a non-clinical group to experience lifetime hypomanic symptoms is associated with their circadian preference, while it may not necessarily be related to bipolar disorder. Among the different circadian preference groups, the evening-type group was observed to be experiencing hypomanic symptoms more signi cantly in relatively more aspects of their lives.
However, even for an evening-type person, the increase in the level of activity and energy, which is readily observed among patients with bipolar disorder, was not very signi cant. Such ndings could be used in future endeavors to develop a diagnostic tool for bipolar disorder.
Bipolar disorder begins with depressive episodes for most patients, and many of them are diagnosed with major depressive disorder or other types of depression in the early stages. However, some studies point out the fact that the patients themselves do not perceive hypomanic symptoms as part of their illness , Muzina 2007, Goldberg et al. 2009). Thus, if a patient with rst-onset depression is identi ed as evening-type, it may be clinically necessary to consider the possibility of him or her having bipolar spectrum disorder.
The limitations of this study are as follows: First, the results cannot be generalized to other age groups as the study was conducted on students from one university. Second, there could be limits to the objectivity of the analysis, as this was a cross-sectional study that only employed self-report questionnaires. Third, the study did not control for the factors that may affect people's circadian preference, namely the ones related to people's biorhythm. Fourth, as this was a cross-sectional study, it is di cult to identify the direction of the interaction between circadian preference and hypomanic symptoms, and it was not clear as to what kind of clinical results were produced as a result of such interaction. However, despite these limitations, the study is signi cant because it identi ed the relationship between circadian preference and hypomanic symptoms based on large sample size. The ndings will be helpful in extending the research to clinical groups in the future and conducting long-term longitudinal studies on the subject group to identify the occurrence and prognosis of bipolar disorder. In addition, a comprehensive follow-up study using objective assessment tools, such as measurement of melatonin levels and actigraphy, will be necessary to further identify the relationship between circadian preference and hypomanic symptoms among people in different age groups and environments.

Conclusions
This study found that in a non-patient demographic, evening-types experience relatively more hypomanic symptoms than others. However, even for evening-types, the increase in the level of activity and energy, which is typically observed among patients with bipolar disorder, was not very signi cant. Such ndings could be used to develop a new tool to distinguish bipolar disorder from other conditions more accurately. Furthermore, the study suggests the possibility that circadian preference could be used as a factor to evaluate whether a patient with depression is likely to develop bipolar disorder later on.

Declarations
Ethics approval and consent to participate Written consent was collected from all participants. The study was approved by the Research Ethics Committee of Kongju National University.

Consent for publication
Not applicable.

Availability of data and material
The data are not publicly available due to information that could compromise research participant privacy.

Competing interests
No. Values were presented as mean ± SD or n (%).