Circadian Levels of Serum Melatonin and Cortisol in relation to Changes in Mood, Sleep, and Neurocognitive Performance, Spanning a Year of Residence in Antarctica

Background. Altered circadian cortisol and melatonin rhythms in healthy subjects exposed to an extreme polar photoperiod results in changes in mood and sleep, which can influence cognitive performance. Materials and Methods. We assessed the circadian rhythm of 20 subjects who wintered over at Maitri (70°S, 11°E), India's permanent Antarctic station, from November 2010 to December 2011. Serum cortisol and melatonin levels were measured by radioimmunoassay at 8 am, 3 pm, 8 pm, and 2 am in a single day, once each during the polar summer and winter photoperiods. Conventional psychological tests, Depression, Anxiety, and Stress Scale (DASS-42), Epworth Sleepiness Scale (ESS), and a computerized neurocognitive test battery were used to measure mood, sleep, and cognitive performance. Results. The mean scores for DASS42 were higher during midwinter suggesting the presence of “overwintering.” Mean diurnal cortisol levels during summer and winter were comparable, but the levels of melatonin were markedly higher during winter. Higher 8 am melatonin levels were associated with better sleep quality, lower depression scores, and better performance in tasks like attention, visual memory, and arithmetic. Conclusion. Timing of artificial light exposure and usage of melatonin supplements in improving sleep and cognitive performance in expedition teams are of future research interest.


Introduction
Residence in Antarctica exposes humans working in the research bases to extremes of weather and photoperiod, in addition to geographic and social isolation. e annual average solar radiation decreases with increasing latitude, and this difference is marked during winter than in midsummer in the circumpolar regions. is alters temporal homeostasis and affects sleep wake cycles, mood, cognition and work performance of base personnel. e psychological disturbances in Antarctica, referred to as the "over winter syndrome, " may also relate to change in the circadian rhythms of the hormones cortisol and melatonin [1][2][3]. Photic stimuli are the primary synchronizer of the endogenous circadian pacemaker in the suprachiasmatic nucleus of the anterior hypothalamus. erefore temporal modulation of the ambient light settings over a 24 hour period in Antarctica can promote better adjustment to the seasonal photoperiod. e endocrine control over the circadian pacemaker is re�ected by the changes in serum levels of cortisol and the pineal indoleamine melatonin. Melatonin, noradrenaline, and acetylcholine decrease with light activation, whereas cortisol, serotonin, GABA, and dopamine levels increase. In normal subjects, we expect a peak melatonin level at approximately 3 am, whereas cortisol peaks at approximately 9 am [4]. In healthy individuals, the transition from low light to bright light in the morning results in a surge in cortisol levels and a decline in melatonin levels [5]. Since the photoperiod is the principal factor regulating seasonal adaptation, melatonin responses may mediate changes in endocrine, cognitive, psychological, and immune function [6].
ere is mounting evidence to suggest that long-term disruption of circadian rhythm correlates with disease states such as ischemic stroke [7], heart disease [8], metabolic dysfunction, and glucose intolerance [9]. Additional studies have also found that chronic shi work is signi�cantly associated with an increased risk of colorectal, endometrial, and prostate cancers [10]. Disintegration of sleep wake cycles and poor sleep health is also associated with a range of neuronal diseases such as Alzheimer's or Huntington's disease [11,12]. Psychological research in Antarctica provides insight into how individuals (and groups) function in an extreme and isolated milieu. However, while interpreting and comparing published data, the racial, psychosocial, educational and cultural differences of expedition teams needs to be considered.
Sleep disturbances and psychological changes have been reported in overwintering teams, but only one study on 9 individuals has tested the effect of the changed diurnal rhythm in the Antarctic on human cognition and performance [13]. Previous studies have reported that circadian cortisol levels have an important role in cognition in an age independent manner. Studies demonstrated that frequent transmeridian �yers with a disrupted diurnal cycle had higher evening cortisol levels, which was associated with cognitive de�cits and right temporal lobe atrophy [14]. Objective assessment of cognitive performance is important for comparison and reproducibility. Computerized neurocognitive tests (CNT) allow accurate measurements of reaction times and test-retest reliability. However, they may be less reliable than conventional tests when traits like verbal skills and personality are being tested [15]. Hence a combined approach is suitable in situations where comprehensive evaluation of personality and serial testing of cognitive skills is needed. e objective of this study was to (i) evaluate the seasonal variation in serum cortisol and melatonin diurnal rhythms during the polar summer and winter photoperiods, (ii) study mood and depression, and (iii) assess the change in objective neurocognitive performance of the base personnel over a period of one year in Antarctica.

Subjects and Study
Design. Twenty male members of the 30th Indian Scienti�c Expedition to Antarctica (ISEA), certi�ed to be in good mental and physical health and free of intrinsic sleep disorders following medical and psychiatric examination and acclimatization training, were recruited for this prospective study with written informed consent. e study was approved by an Institutional Ethics Board of the National Centre for Antarctic and Ocean Research (NCAOR), India's nodal agency for Antarctic research and was conducted in accordance with the standards of the Declaration of Helsinki. e study was started in November 2010 soon aer arrival of team members in Antarctica and the group was followed up prospectively till the conclusion of the expedition in December 2011.

Study Protocol. e Indian Antarctic research base
Maitri is located at 70 ∘ 45 � 57 �� S, 11 ∘ 44 � 09 �� E. e overwintering team consisted of 6 scientists and 14 technical support staff who are required to keep the base functional and the scienti�c experiments running. All the personnel are required to maintain the same sleep rhythm (wake at 0700 hours UTC) and work schedule except for one person who is on night watch, which is a periodic duty once every ten days by rotation. One meteorologist also maintained a night shi duty for the greater part of the year.
e winter period at 70 ∘ S starts in mid April and lasts till October. Serum levels of melatonin and cortisol were measured in the subjects at 8 am, 3 pm, 8 pm, and 2 am in a single day using an indwelling antecubital catheter during 2 different photoperiods; once during summer and the other during the polar winter. e samples re�ecting the summer photoperiod were collected in January 2011, a few weeks aer arrival in Antarctica, and processed to obtain serum which was stored at minus 25 ∘ Celsius until assay was done. Repeat serum samples were collected in June 2011 during the polar winter. e serum was frozen immediately and stored at minus 25 ∘ C. e samples were transported back to India on dry ice using monitored cold chain transport at −25 ∘ C to −70 ∘ C.
Melatonin was measured by radioimmunoassay (RIA), using I 125 -melatonin (Labor Diagnostika Nord GmbH & Co. KG, Nordhorn, Germany). e reference range, as stated by the manufacturer, was 30-150 pg/mL. e sensitivity of the method was 2.3 pg/mL and the intra-assay variability less than 9.8% and inter-assay variability 8%. e antimelatonin antibody used had a cross reactivity of >0.1% with melatonin related metabolites or precursors. Serum levels of cortisol were assayed by GammaCoat RIA (DiaSorin, Stillwater, MN) with a sensitivity of 0.2 g/dL and intra-assay coefficient of variation (CV) of 6.5% and inter-assay variability of 9%. e reference range was 7-25 g/dL for morning values and 2-9 g/dL for evening values.

Sleep Quality Evaluation.
All subjects were required to maintain a sleep log over a 2 month period in summer and again in winter. e sleep log documented the sleep/wake times, caffeine and alcohol intake, night time awakenings, use of sleep aids such as sleeping pills, music, reading, total duration of sleep, and a subjective evaluation of restorative sleep. Epworth Sleepiness Scale (ESS) [16] and Morningness Eveningness Questionnaire (MEQ) [17] were administered on arrival, during summer and during winter. e ESS is a validated scale for measuring daytime sleepiness and the MEQ scores reveal individual adjustment to morning and evening photoperiods and the synchronization of the internal clock to work schedules.

Neurocognitive Evaluation.
A combination of conventional psychological tests and a computerized neurocognitive test battery (CNT) were used to serially measure relatively mild degrees of neurocognitive impairment in the base personnel. CNTs were performed using Psychology Experiment Building Language (PEBL) Version 0.11. e tests included visual and verbal memory, attention, visual motor speed, arithmetic and reaction times, a modi�cation of the Halstead Reitan Battery, to test domains likely to be relevant to wintering and working in a challenging environment like Antarctica (see Table 1).
Neurocognition Tests: the battery of tests employed in this study included both conventional and computerized tests [15].

Attention
(1) Card Sort Test (CST). e card sort test measures shiing attention, that is, the ability of the subject to focus on one set of instructions and then to shi to another set in response to the change in test response. e test involves 4 packs of cards which have different geometric shapes (circle, star, triangle, and cross), different number of shapes (1, 2, 3, and 4) and different colours (red, yellow, green, and blue). en a card with a pattern having a speci�c colour and number of shapes is presented and the subject is instructed to match the card with one of the 4 packs by a rule which needs to be deduced from the test response. e rule is changed randomly (matching of colour, number, or shape) throughout the test and the subject is scored for the number of correct matches made in a �xed time. is is a modi�cation of the Wisconsin Card Sort Test and measures attention and cognitive �exibility [18].
(2) Digit Span Forward. Random digits (0 through 9) are presented to the subject on the screen and simultaneously by computer generated voice in increasing sequence as long as correct responses are obtained. If the response is incorrect, the subject is tested with a shorter sequence which is then increased by one digit successively until the longest sequence of digits is correctly repeated. e test is continued for 20 trials and the mean number and longest sequence of correct responses is calculated.
(3) Continuous Performance Test (CPT). is test measures sustained attention or task related vigilance. e subject is instructed to press the space bar when any alphabet is �ashed on the screen except the alphabet "X. " e reaction time for each target stimulus is measured. e test is performed with 300 trials of 16 blocks. e entire test takes about 14 minutes. e number of "X" foil stimuli is 25 and target alphabet stimuli are 300, respectively. e test gives the accuracy rates for the targets and foil non targets and also the mean reaction times for correct and erroneous responses. is is a sensitive test for central nervous system dysfunction [15].

Executive Function
Stroop Test. e subject is presented with four colors/color words (red, green, blue, yellow) and is instructed to select either the colour or the word. At �rst 5 practice trials are allowed. en 20 random trials for congruent and incongruent colour words are presented. In total 5 words are presented in a congruent color and word (e.g., word RED in red font) which generates a congruent reaction time and 15 test words are presented in an incongruent fashion (e.g., word RED in green font), which generates a second reaction time [15].

Visual Motor Ability
Finger Tapping Test. e subject is asked to press the key "X" on the computer keyboard as many times as possible in 90 seconds. is is recorded separately for the right and le index �nger. is simple test provides reliable and reproducible data about �ne motor control, motor speed, and visual motor function [19].

Visual Memory
Test. e visual memory test involves the presentation of a sequence of everyday object images with their names (apple, pear, bus, bird, etc.) on the screen. Each image is �ashed for 2 seconds in increasing number and the subject has to memorize the order of presentation. en a screen of nine �gures arrayed in a grid is presented and the subject is instructed to click on each image in the order of presentation. e test is repeated with 3 trial blocks of up to 9 image sequences. It generates the average number of images remembered correctly in any sequence. is is based on the Rey Visual Design learning test [20].

2.4.5.
Arithmetic. e subject is presented with simple addition, subtraction, multiplication, and division equations (up to 2 integers) at �rst slowly and then with increasing frequency for a period of 3 minutes. e score is based on the accuracy and number of equations solved per minute.

Speed of Processing
(1) Simple Reaction Time. e subjects were instructed to click the le mouse button each time a triangle was �ashed on the screen for ten trials. is generates the longest, shortest, and mean reaction times. A similar test for auditory reaction time was used by presenting a tone stimulus by headphones. e test generates the longest, shortest, and mean reaction times.
(2) Choice Reaction Time. e subjects were presented an image of a geometric design (circle, square, triangle, oval) and another image at random immediately aer. If the new image was identical to the one immediately before it, they were asked to click the right shi button and if not, the le shi button. e reaction times and accuracy was recorded over 1 minute for a sequence of random images [21].
(3) Psychological Tests. e instruments for assessing physical and mental health included demographic information on age, occupation, education, and subjective health complaints, and standardized scales such as depression, anxiety, and stress scale (DASS-42) [22] and Mini Mental Status Examination (MMSE) [23]. DASS-42 measures the behavioral manifestations of depression, anxiety, and stress in a self report interview to construct a score (see Table 2). Period differences in the circadian rhythm were tested by measuring serum cortisol and melatonin levels during summer and winter. e sampling periods were correlated with the ambient solar radiation over the station during the polar summer and winter photoperiods. is was measured by means of a sunlight photometer during the subjects' sojourn in Antarctica, and data was obtained with permission of the Indian Meteorological Department (IMD).

Statistical
Analyses. Conventional methods, mean and standard deviation (SD), Student's t-test, Fisher's T test, and Pearson's correlation coefficient were used to analyze the diurnal variation in cortisol and melatonin levels with the level of signi�cance <0.05. All data were reported as means ± SD. e soware SPSS for Windows ver.13 (SPSS Inc; Chicago, IL, USA; http://www.spss.com/) was used for statistical analysis.

Results
e mean age of the 20 male subjects was 39 ± 10.24 years. All the support staff had a high school education and an additional technical quali�cation. Eighteen (90%) of the 20 subjects were right handed. e entire team was subjected to psychological testing prior to selection for the expedition and none of them suffered from depression or sleep disorders. Table 3 shows the basic clinical characteristics of the subjects. ere was a trend for an increased weight, BMI, and systolic blood pressure in older individuals. Four (20%) were hypertensive, requiring antihypertensive drugs in addition to lifestyle and diet modi�cation. Two of these subjects were newly diagnosed with the disease in Antarctica.

Psychological
Changes. e maximum score of the DASS-42 is 42 in each of the depression, anxiety, and stress subscales. erefore lower scores are better. Table 4 showing the DASS42 subscale scores is revealing. e mean scores for depression and anxiety during midwinter were signi�cantly higher than summer scores ( < 5) suggesting the presence of "over wintering. " Daytime sleepiness was excessive (score 16) in just one subject as per the ESS, but 4 subjects had borderline scores (mean 11.4 ± 1.1) during summer and 2 of them continued to report daytime sleepiness during midwinter.

Sleep Quality and Subjective Polar
Acclimatization. e work schedule during summer for the technical and maintenance staff involved reporting for duty at 0900 hrs and mild to moderate physical work for a period of 3-4 hours. Post lunch (1230-1400 hrs) there was another work period from 1500 to 1800 hrs. With the onset of the polar night, the evening shi was stopped due to adverse weather conditions and poor visibility. However during blizzards, special rotation duties for snow clearance and station maintenance resulted in a peak in physical activity. ese periods of 2-4 days were also associated with increases in perceived stress. e scientists maintained the same work schedule in summer and winter, but participated in external work and blizzard duties. Additionally, all team members participated in "galley" duties by rotation involving night watch every 8-10 days. e sleep logs showed a marked difference between summer and winter periods in 5 (25%) subjects, all scientists. ey reported a decreased latency, increased duration of sleep, fewer early morning awakenings, and restorative sleep as compared to summer. e sleep patterns of the technical support team varied only in the total duration of sleep during winter (mean 8.42 ± 1.18 hours) as compared with summer (mean 8.64 ± 1.21 hours). In contrast, the total sleep duration in scientists was 7.33 ± 0.98 hours in summer and 7.66 ± 1.03 hours in winter (see Table 5). e total duration of sleep was increased due to an increased frequency of aernoon napping which was related more to no work shi in the post lunch period rather than due to increased fatigue. Sleep aids used by the subjects included listening to music, reading, and watching television. Pharmacological therapy, including use of benzodiazepines and zolpidem, was needed in 7 (35%) subjects at one or more times for sleep induction.

Cortisol and Melatonin Diurnal
Rhythm. e results of the diurnal and seasonal variation of cortisol and melatonin provide interesting biochemical correlates for our clinical data. As expected, the mean cortisol levels were highest at 8 am, with a progressive decline during the day, the nadir being reached at 2 am. e mean levels in winter were higher at 8 am and 3 pm but the difference was statistically equivalent (see Figure 1). During midwinter, we noted a marked 6 Neuroscience Journal increase in the mean serum melatonin levels. Again, the maximum levels were noted at 2 am which is physiological. But melatonin levels were persistently high throughout the day. Figure 2 demonstrates the marked contrast in melatonin diurnal rhythm during the two seasonal photoperiods. While the mean levels of these hormones showed a physiological and predictable diurnal rhythm, we noted marked individual variation. is was more so for scienti�c staff involved in time bound data collection. e scores on MEQ provided clues to the degree of subjective acclimatization in relation to polar work and photoperiod. Soon aer arrival and during midsummer, 5 (25%) subjects preferred starting work late between 10 and 11 am and up to 4 (20%) persons responded that they were mentally and physically more active in the evening. Twelve (60%) had a "de�nitely morning, " chronotype, 2 (10%) corresponded to "moderately morning" and 4 (20%) were neither type. In contrast, during midwinter, 14 (70%) subjects favored a "de�nitely morning" chronotype as opposed to a "moderately morning" 4 (20%) or "neither morning nor evening" chronotype 2 (10%). Most reported a preference to start work earlier, between 8 and 9 am.
We assessed whether cortisol and melatonin diurnal rhythms correlated with psychological tests and cognitive  performance (see Tables 6 and 7). In most cases, no clear trend was obtained for hormone levels and cognitive function. We did �nd that a higher evening cortisol negatively correlated with performance on tests for attention like digit span and card sort during summer. Higher evening cortisol levels also correlated with increased DASS 42 scores. During winter, elevated 8 am melatonin levels correlated signi�cantly with better performance on tests for attention, digit span, and arithmetic during winter. Although this may seem paradoxical, an increased 8 am melatonin level was also associated with better sleep quality, which may have improved performance status.

Neurocognitive
Assessment. e subjects were tested by both conventional methods and by computerized tests in the morning. ose with a "non morning" chronotype were retested in the evening, but there was no signi�cant change in the scores other than the continuous performance test (CPT) for attention. e results of the mean scores and reaction times are shown in Table 8. Overall, the mean scores and reaction times of many tests were statistically equivalent in the summer and winter photoperiods. ere was a signi�cant increase in arithmetic computational accuracy and speed in winter. Also subjects performed better on choice reaction time with lower mean RT and improved accuracy suggesting improvement in speed of processing. ere was a marginal improvement in scores of the card sort test for shiing attention and digit span in winter. is was more so for individuals with higher 8 am melatonin levels ( ). However CNT results need to be interpreted in the appropriate clinical context. Subjects with higher DASS-42 scores, scored lower on digit span ( ), shiing attention ( ), and CPT ( ). Mean scores for simple reaction time, digit span, and visual memory were better in individuals aged 30 years. Higher ESS scores did not correlate with poor performance on CPT, and no "sleep attacks" were noted during testing.

Discussion
Circadian rhythms are intrinsic to all living systems. Adaptation to the polar environment involves synchronization to temporally relevant stimuli such as light, temperature, work, meal schedules and thus the internal clock gets aligned with the external solar time. Rhythmic modulation of behavior and physiologic processes such as sleep, cardiovascular, cognitive performance results from the circadian control over diverse pathways [25].

Circadian Rhythm in Antarctica.
Few studies have been done on circadian rhythm in Antarctica and no study has ever been done on the impact of prolonged Antarctic residence on circadian rhythm in Indian expedition team members [1-3, 26, 27]. Sleep schedules, activity, plasma melatonin, and core body temperature were documented in 9 Japanese subjects at Dome station (77 ∘ S). ey concluded that the sleep rhythm was largely governed by the work schedule while the circadian rhythm in plasma and core body temperature is in�uenced by photoperiod [1]. Adaptation of the melatonin rhythm in British Antarctic personnel at Halley station (75 ∘ S) has also been reported in night shi workers over short periods [2]. Another study also evaluated the usage of light intensity regulation at the stations to in�uence sleep behavior and mood disturbance issues. ey concluded that further investigations of phase shiing techniques such as timed bright light and administration of melatonin are indicated to improve performance at work [3,13].
We measured the serum levels of cortisol and melatonin during summer about a month aer arrival in Antarctica. e period of transition from one expedition team to the next involves adjustment to new routines, constant daylight and cold dry windy weather. is high activity period results in an elevated stress hormone pro�le. erefore the effect of adaptation stress must be kept in mind while interpreting results. But we found that the mean 8 am cortisol levels were higher in winter when compared with summer. An increased 8 am summer cortisol level correlated with a reduced reaction time ( ). Subjects showing a higher 8 am melatonin level during winter also performed better in the neurocognition tests for immediate memory ( ), and arithmetic ( ), suggesting better adaptation. is may seem counterintuitive at �rst glance, as melatonin is a hypnotic hormone. e result may be best interpreted in the context of restorative sleep improving cognitive performance in winter. An elevated evening cortisol did not correlate with de�cits in visual recognition memory, in contrast with previous studies. High evening cortisol levels may indicate chronic stress and poor sleep quality, which in turn impairs performance [28,29]. In our study, no clear association could be obtained for individual cortisol or melatonin values and the corresponding neurocognitive test results of the subjects. is is largely due to the smaller sample size, which is restricted by the number of team members in an over wintering expedition. So results need to be interpreted carefully, as multiple factors may affect an individual's day to day performance.

Sleep
Quality. ere was an improvement in sleep quality, duration, and decreased latency in winter. e maximum sleep disturbances occurred in the �rst one month aer arrival, in the phase of adjustment to a 24 hour light photoperiod. e scienti�c staff had fewer total sleep hours than the logistic staff and also reported increased sleep latency. e sleep wake cycles of scientists were irregular, changing with night watch, galley duties, and timed data collection. Subjects also reported better sleep quality, fewer night awakenings, and decreased latency during winter which correlated with higher melatonin levels. Studies from other stations reported that poor sleep was associated with depressive symptoms. In our study, patients with higher DASS-42 scores reported night interruptions and early awakening. Winter was also associated with an increased volume and frequency of alcohol intake, but no changes in sleep parameters were associated with drinking. Sleep aids are useful and pharmacological therapy is oen needed in the early adjustment phase to avoid sleep disruption.

Circadian Rhythm and Neurocognitive Performance in
Antarctica. e subtle changes in behavior, sleep, memory, and cognition during long duration expeditions may have long term health consequences on aging, cardiovascular, and cerebrovascular disease. In this study, we performed tests for domains like memory, attention, reaction time, and speed of processing which can be affected by extrinsic factors like drugs, disease states, and time of day testing. Also, motivation, interest, mood, and vigilance affect individual scores. ough the cognitive performance for the whole team averaged the same during summer and winter, we noted other outcomes. Subjects with depression had lower scores for memory and shiing attention in midwinter. is may re�ect subtle changes in the level of alertness as a function of diurnal rhythm. Despite advances of technology and upgraded facilities at scienti�c bases, it remains a challenge to winter over in the Antarctic. Subjects have to perform energy consuming tasks including but not limited to snow clearance, prolonged driving, and maintenance of vehicles, heating systems, electrical generators, and so forth. ese activities are prone to human error. In addition interpersonal relationships and closed group psychology and the sense of physical isolation affect performance [30][31][32].

Conclusions
Circadian rhythms are modulated by central pacemakers like melatonin and cortisol which synchronize to stimuli such as photoperiod and activity, and thus indirectly the internal clock becomes predictive of solar time. Our study found that the diurnal rhythm for cortisol was comparable during the polar summer and winter. In contrast, circadian secretion of melatonin was affected during the polar winter and higher levels persisted throughout the day, peaking physiologically at 2 am in both seasons. is was associated with better subjective sleep quality during winter. Again, higher levels of 8 am melatonin correlated signi�cantly with better neurocognitive performances in tests for attention and immediate memory. We also report the effect of over wintering in our subjects re�ected as higher depression, anxiety, and stress scale scores during the polar winter. is impaired cognitive performance in few subjects. In light of the fact that the team size of wintering teams in the Antarctic is kept to a minimum, pooling of data will be useful in making recommendations for future expeditions. Timing of arti�cial light exposure and usage of melatonin supplements in improving sleep and cognitive performance in expedition teams are of future research interest.

Con�ict o� �nte�ests
e authors declare no con�ict of interests.