Sleep and chronotype in adolescents

As it is commonly known and confirmed by several studies,
adolescence is the time in life that goes along with being the
latest chronotype in community. This implies that physiologically
adolescents tend to go to sleep later and get up later than other
age groups. At the same time adolescents are the age group who
spend the highest amount of time studying (at school), in order to
prepare for their later (working) life. Since the usually requested
school start times in Germany around 8:00 a.m. rather meet the
needs of earlier chronotypes than those of normal- and later ones,
adolescents and shift workers belong to the groups with the largest
sleep deficit. There is a recognition that healthy sleep (adjacent
to healthy nutrition and physical exercise) states one pillar of
health and wellbeing, as well as playing a role in the
consolidation of memory. For this reason it appears worthwhile to
aim at optimising sleep behaviour and –circumstances in
adolescents. Consequently the relations between chronotype and
sleep should be understood, in order to gain new insights for the
conductance of health prevention programs; especially in schools.
The aim was to create one building block for such research. Thus
the present study aspired to finding a method of examining the
sleep of adolescents “in real life” via a field study with a mainly
explorative approach. In order to do so, a simple and
cost-effective method was sought, to obtain hypnograms of students
in a mobile sleep lab at their school. The mobile, automated and
easy to use EEG “Zeo” was elected since it appeared to be an ideal
tool for meeting the requests of the present study. This device
consists of a headband with three frontal electrodes and a base
station that records, inter alia, the following sleep parameters:
total sleep, sleep latency, time awake after falling asleep, light
sleep (stage 1 and 2), deep sleep (stage 3 and 4) and REM sleep.
These are interpreted automatically so that no more manual
evaluation of raw EEG-data has to be performed. After hypnograms
were obtained, their data were assessed in relation with the
students’ chronotypes. To do so, the total duration of the
respective named sleep parameters were correlated with the
chronotypes. Total durations of the respective sleep phases were
also correlated with each other. Prior to deducing EEGs on two
consecutive nights per student, the chronotype of each participant
was determined via the Munich Chronotype Questionnaire. In order to
validate the obtained data and gain further insights into the
individual sleeping-behaviour of participants, these were asked to
fill in sleep logs for two weeks during the test-phase. The main
question of this thesis was weather common chronobiological
expectations about sleep timing and –phases could be replicated in
the sleep-mobile-setting, using Zeo. In opposition to the usual
observance in sleep-labs, no first night effect was seen between
the first- and second nights in repeated measures ANOVA. For this
reason both nights were used for further analysis in this study.
The first hypothesis was that later chronotypes would be observed
to fall asleep later in the sleep mobile, and wake up later. This
hypothesis could not be confirmed. Similarly the second hypothesis,
which expected later chronotypes to be observed to spend more time
overall sleeping in the sleep mobile than earlier types, because
they would have to catch up on their accumulated sleep deficit
throughout the week, could not be approved. Both outcomes may be
influenced by the study’s set-up in which respectively four
students slept in the sleep mobile at the same time. Thus there
hardly was a possibility for one student to get up or go to sleep
without waking up the others. No correlation was seen between
chronotype and the total duration of the above named sleep
parameters. Sleep onset and sleep end were compared to an MCTQ- and
sleep-log-deduced 24-h-sleep window. While sleep onset, as measured
by Zeo was correlated with the calculated value, no such
correlation could be shown between calculated- and measured values
for sleep end. An unexpected finding in half of the hypnograms was
that students were observed to have fallen asleep via a REM-phase
rather than via a light sleep phase, as usual. Testing for
correlations between the total durations of sleep phases, the
following observations were made: • Total sleep showed a positive
correlation with light sleep. • Total sleep showed a positive
correlation with REM sleep. • Time awake after falling asleep
showed a negative correlation with REM sleep The latter discovery
was unexpected, since there is no explanation as to why wake-up
phases throughout the night might lead to a decline in REM-sleep.
Due to the named unexpected findings regarding REM sleep, a
post-hoc hypothesis was generated. This hypothesis assumes that Zeo
tends to confound wakefulness with states of REM. Further
literature research showed a high probability of this hypothesis
being correct, since 1) the frontal EEG-deductions during REM-sleep
are rather similar to those deducted during wakefulness, whereas
alpha-waves that can be observed in relaxed, awake test-persons
with closed eyes are ideally deducted in the dorsal regions of the
head. 2) all studies in which Zeo’s output is claimed to have a
high correlation with classical polysomnography exist only as
abstracts and have not yet been published completely. Furthermore,
towards the end of the present study, oral communication with Zeo
approved that the device’s hardware could not facilitate a perfect
evaluation of REM-phases. In retrospection the setting of measuring
student’s sleep profiles in school on weekends within the sleep
mobile was accepted well by students and can be recommended for
further research. A limitation in this regard is the analysis of
total sleep, sleep onset and sleep end that are being
de-individualised by the collective residence of students in the
sleep mobile. A further use of Zeo for scientific purposes cannot
be advised, while a repetition of the present study with classical
EEGs is regarded to be commendable. Although such proceeding would
include a higher workload in applying and manually evaluating EEGs,
the reliability of data would be considerably higher. Moreover, the
raw, unevaluated data that would be obtained could be used for a
refined evaluation. In the meantime Zeo may well serve for use in
health care programs, where it could be applied for individuals to
gain insights into their own sleep, its structure and importance.