Logo Logo
Switch language to English
Kühnle, Tim (2006): Quantitative Analysis of Human Chronotypes. Dissertation, LMU München: Fakultät für Biologie



Like most other organisms, human behaviour is under control of a circadian (= about a day) clock which can adapt to the daily change of light and darkness. The capability of the circadian clock to synchronize to the light-dark cycle is highly systematic and under genetic control. Daily behaviour of humans can vary greatly, leading to different chronotypes and extreme chronotypes are commonly described as 'owls and larks'. While larks fall asleep in the early evening and wake up early in the morning, owls go to bed when larks wake up, sleeping until about noon. Recently, the Munich ChronoType Questionnaire (MCTQ) has been developed in order to quantitatively assess chronotype and to face problems arising from common work day schedules. Chronotypes are normally distributed in the general population. MSF-Sc, the middle of sleep phase on free days corrected for sleep dept accumulated on work days, is used as reference point for chronotype. The objectives of this study were both to validate and improve the MCTQ and to identify and estimate biological and social factors that potentially influence the distribution of chronotypes. Test-retest reliability was controlled at an interval of six months (n=101). Both dates show a highly significant correlation (p<0.001) and no significant difference (paired t-Test: p>0.027, Bonferoni corrected level of significance: p=0.0012). Test-retest reliability was controlled separately for females and males, different age (≤30 years and >30 years), and chronotype groups (MSF-Sc ≤4.28 and >4.28). A sample of 15 people was tested at an interval of three weeks. The results of both dates correlate with high significance (p<0.001) and paired t-Test showed no significant difference between both dates (p>0.05, Bonferoni corrected level of significance: p=0.002). The capability of the MCTQ to assess actual sleep times was tested with 6-week long sleep logs (n=628). Again, gender, age, and chronotype groups were evaluated separately and chronotypes were separated into five different groups (MSF-Sc <2.17; 2.17≤7.25; 2.17<4.28; 4.28≤7.25; >7.25). Assessed (MCTQ) and actual (sleep log) sleep times correlate with high significance (p<0.0001, Bonferoni corrected level of significance: p=0.0007). There is a tendency of overestimation of chronotype towards extreme chronotypes. While average chrono-types (MSF-Sc 4-5) can assess their actual sleep times very precisely, earlier chronotypes (MSF-Sc <2) assess themselves earlier and late chronotypes (MSF-Sc >7) assess themselves later as they actually are. Due to this overestimation, paired t-Test showed assessed and actual sleep times to be significantly different in most of the sub-groups (p<0.0001, Bonferoni corrected level of significance: p=0.0007). However, the tendency of overestimation is highly systematic and a linear correction could be applied for further investigations. Principal Component Analysis (PCA), containing all questions of the MCTQ that could represent a phase reference point of daily behaviour, resulted in a four factor solution and revealed several redundancies in the MCTQ. Besides phase of sleep, as represented by MSF-Sc, duration of sleep, as represented by average sleep duration, was identified as independent property of the sleep-wake cycle. Phase of entrainment explaines about 50% of total variance and sleep duration about 20% of total variance within the set of MCTQ questions. Most questions, except mid-day dip, are highly associated with the same factor as MSFSc, thus not representing statistically independent dimensions of daily behaviour. Another factor is represented only by mid-day dip and explains about 7% of total variance. However, the independency of this factor might rather result from social influences. The last factor (about 12% of total variance) is dominated by sleep end on work days which limits sleep for most people on work days. This factor does not represent an independent structure of the sleep-wake cycle but the influence of the social clock on daily life. Using Multiple regression analysis, the influence of biological and social influences on chronotype has been quantified in relation to each other. Phase and duration of sleep are differently associated with age, gender/sex, daily outside light exposure, photoperiod (season), BMI, and place of residence. Phase of entrainment is substantially associated with age, gender/sex, daily outside light exposure, photoperiod (season), and Place of residence, while sleep duration is associated only with age, gender/sex, daily outside light exposure, and body mass index (BMI). Analysis of covariance (ANCOVA) showed interacting effects between gender/sex & age and gender/sex & photoperiod (season) only for sleep phase. As a results of this work, the Munich ChronoType Questionnaire (MCTQ) has been proven to be a reliable, quantitative tool for the assessment of chronotype. Its precision, shortness, and the ability to correct for factors influencing chronotype make it the most preferable questionnaire in human chronoscience. In depth statistical analysis with extreme high numbers (up to n=29,000) hypothesizes that sleep duration is an independent property of daily behaviour besides chronotype, and showed both, chronotype and sleep duration, to be differently influenced by biological and social factors. This offers a good basis for new hypotheses and experiments.