Analysis of the Psychological States of the Scientic Expeditioners in Extreme Environment

Background: To investigate the relationship among individuals’ mood, sleep and personality traits in a long-term polar environment. Methods: A total of 85 Arctic expedition team members were recruited in this study. In addition to personality assessment, they also completed assessments of anxiety, depression, and sleep at baseline, the second and eighth week. The rank sum test was used to compare the differences between mood and sleep at different time periods, and the generalized linear mixed model was used to explore the relations between personality and mood, as well as sleep. Results: The PHQ-9 scores showed a statistically signicant difference between baseline and week 8. The PSQI scores showed a signicant difference between baseline and week 2. At the same time, there was a signicant difference between those who are married and those being single. Single individuals scored signicantly lower on GAD-7 than married ones. The relationship between personality traits and PHQ-9 scores was statistically signicant and individuals’ coping styles have an impact on their subjective sleep quality, sleep disturbance and daytime dysfunction. Conclusion: In the polar environment, over time, people's mood and sleep will be affected, and that is related to individual characteristics, personality and coping styles. Cochran's Q test was used to explore the similarities and differences among different symptomatic groups of PHQ-9, GAD-7 and PSQI at different time and latitudes. The results showed that based on PHQ-9 scores, a total of 62 participants were included. The percentages of the symptomatic group at baseline, the 2 nd and 8 th week were 25.8%, 30.6%, and 37.1%, respectively, χ2=5.286, p=0.071>0.05. There were no signicant differences at these three time points for the symptomatic group. Based on scores on GAD-7, a total of 61 participants were included. The proportions of the symptomatic group at baseline, the 2 nd and 8 th week were 14.8%, 21.3%, and 19.7%, respectively, χ2=1.733, p=0.42>0.05. There were no signicant differences at these three time points for the symptomatic group. Based on PSQI scores, 43 people were included. The proportions of the symptomatic group at baseline, the 2 nd and 8 th week were 34.9%, 37.2%, and 44.2%, respectively, χ2=2, p=0.368 >0.05. There were no signicant differences at these three time points for the symptomatic group (Figure whether A total 29 completed relevant This study also explored the relationships between personality traits and the PSQI subscale. It was found that different personality traits have an impact on multiple subscales, including sleep quality, sleep latency, sleep duration, sleep eciency, and sleep disturbances Factor Q4 Tension (coecient 0.21, p<0.05) score is positively correlated with sleep quality score. Factor L Vigilance (coecient -0.15, p<0.05), Factor N Privateness (coecient -0.12, p<0.05) score and sleep latency score is negatively correlated. Factor M Abstractedness (coecient 0.06, p(cid:0) 0.05), Factor Q3 Perfectionism (coecient 0.19, p(cid:0)0.05), Factor Q4 Tension (coecient 0.22, p(cid:0)0.05) score are positively correlated with sleep duration score. However, the scores of the following factors are negatively correlated with sleep duration, including Factor B Reasoning (coecient -0.09, p<0.05), Factor C Emotional Stability (coecient -0.07, p<0.05), Factor F Liveliness (coecient -0.08, p<0.05), Factor I Sensitivity (coecient -0.09, p<0.05), Factor O Apprehension (coecient -0.12, p<0.05), and Factor Q1 Openness to Change (coecient -0.07, p<0.05). The scores of the following factors are positively related to sleep eciency scores, Factor B Reasoning (coecient 0.18, p<0.05), Factor F Liveliness (coecient 0.20, p<0.05), Factor M Abstractedness (coecient 0.16, p<0.05), Factor N Privateness


Introduction
Human beings are curious about the unknown. With the development of science and technology, human beings have begun to challenge the polar regions. The polar environment is harsh, the climate is cold, the scenery of long-term maritime navigation is monotonous, faced with challenges such as di culty in replenishment and many uncertainties. This is a huge challenge to the psychological state of the ship's crew members. Studies have shown that the polar environment can in uence the cognition of researchers. For example, the time of decoding complex memories is prolonged, short-term memory and simple spatial cognition are improved in cold environments [1].
Polysomnography, there are no signi cant differences between the two, and PSQI is better than the other in assessing the duration and e ciency of sleep in the past month [13].

Personality test
The Sixteen Personality Factor Questionnaire (16PF) was used as the personality test in this study. 16PF was developed by Cattel based on his personality theory. It has the following features: 1. Objectivity. The structure of 16PF is clear. There are neutral options so that participants are unlikely to be forced to choose an answer biased towards certain way. 2. Standardized. With good reliability, validity, and an average Cronbach's α coe cient of 0.63. 3.
Generalizability. 16PF can be administered individually, as well as in group-settings. Fourth, profundity. The assessment results can be explained in conjunction with psychoanalytic theory.
Coping style rating scale The Coping Style Questionnaire (CSQ) was developed by a group of Chinese researchers, led by Xiao, based on research from abroad, taking into consideration of the language characteristics and behavioral habits of people in China [14]. CSQ is composed of 6 factors, and its validity is 0.35 or higher for the factor loadings of each factor in the adolescent student group and neurosis-control group. At the same time, it has good reliability in both groups.
According to the correlation, arranged in order in a diagram: avoidance, fantasy, self-blame, seek for help, rationalize, and solve the problem.

Procedures
The scienti c expedition team members who undertake scienti c research tasks completed 16PF and CSQ assessments before departure. All the participants completed PHQ-9, GAD-7 and PSQI before departure, at the 2 nd and 8 th week.

Statistical analysis
All data in this study was analyzed using SPSS 25. For the continuous variables obtained from PHQ-9, GAD-7 and PSQI, if the data was normally distributed, the paired t test was used. Otherwise, the Wilcoxon Signed-Rank Test was used to explore the effects of the extreme environment on team members' emotions and sleep quality at different time periods. At the same time, participants were categorized into different groups based on their scores on the questionnaires mentioned above. In terms of scores on PHQ-9, those with a score of no more than 4 were in the asymptomatic group, those with 5 points and above were in the symptomatic group. For GAD-7, the cutoff was a score of 4. The obtained a score no more than 4, and the symptomatic group got a score of 5 and above. Those with a PSQI score no more than 5 were in the asymptomatic group. For those who obtained a PSQI score of 6 and above, they were in the symptomatic group. Comparisons were made between different time periods and latitude, to see whether there are statistical differences among individuals in the symptomatic groups. In addition to investigating the changes of PHQ-9, GAD-7, and PSQI scores over time, this study conducted an open exploration of the relations among individual characteristics and coping styles, emotions, as well as sleep. A hybrid model was adopted to explore the relationships among various factors. We discard the missing data.

Demographic information
A total of 79 people participated in this study, and some of the participants had missing data, missing data were discarded in the statistical analysis. Among them, there were 71 males and 3 females, and gender information of the other ve persons was missing. Regarding marital status, the number of single persons is 23, two individuals were divorced, 48 people were married. One person indicated "Other", and data was missing from ve participants. A   number of 15 participants lived alone, 16 participants reported to live with their parents, and seven individuals lived  with their partners, two people live with children only. There were 31 participants living with partners and children, 8 people had missing data. In terms of levels of education, two people have middle school education and below. A number of eight people have high school education. There are 29 people who nished undergraduate. 35 individuals completed graduate studies and above, and data from ve participants is missing.
At baseline, 73 participants experienced mild depression or none (PHQ-9 scores ≤9), one person had moderate level of depression or above (PHQ-9 score ≥10). Participants' GAD-7 scores indicated that 72 of them self-reported as experiencing mild anxiety or below (GAD scores ≤9). Two people experienced moderate level of anxiety and above (GAD scores ≥ 10). There were 62 people indicating acceptable sleep quality (PSQI scores ≤ 10), and three people had poor sleep quality (PSQI scoreS ≥ 11).
At the second week, 71 people experienced mild depression or none (PHQ-9 scores ≤9). One person had moderate level or above (PHQ-9 scores ≥10). A number of 70 participants had GAD-7 scores indicating mild anxiety or below (GAD-7 scores ≤9) , and two people were moderate or above (GAD scores ≥ 10). When it comes to sleep quality, 52 participants reported an acceptable level of sleep quality (PSQI scores ≤ 10), and 7 people had poor sleep quality (PSQI scores ≥ 11).
At the 8th week, 65 people reported to have mild depression or none (PHQ-9 scores ≤9), 2 had moderate level or above (PHQ-9 scores ≥10). GAD-7 scores showed that 64 people experienced mild anxiety or below (GAD-7 scores ≤9) , two people had moderate anxiety or above (GAD scores ≥ 10). Regarding sleep quality, a total of 55 people reported to have acceptable sleep quality (score ≤ 10), and ve people had poor sleep quality (score ≥ 11).
Relationship between time, the latitude and GAD-7, PHQ-9, PSQI In this scienti c expedition itinerary, the latitude and longitude were recorded once every 6 hours. The average latitude during the rst week was 41.03°±6.59°, the average latitude during the second week was 60.33°±6.72°, and the average latitude during the eighth week was 76.42°±1.21°. Over time, the latitude at which the scienti c expedition team arrived continued to increase. The distribution of GAD-7, PHQ-9, and PSQI scores is not normally distributed.
Therefore, Wilcoxon signed-rank test was used. There were no statistically signi cant differences in GAD-7 scores between different time period (baseline, week 2 and week 8). The There was a statistically signi cant difference in PHQ-9 scores between baseline and the 8 th week. At baseline, the median of PHQ-9 scores was 2 and that at week 8 was 3. Wilcoxon signed-rank test showed that Z=-2.768, P<0.05. However, there were no statistically signi cant differences between baseline and the 2 nd week, the 2 nd and the 8 th week in PHQ-9 scores. The PSQI scores showed a statistical difference between baseline and the 2 nd week. The median score of PSQI at baseline was 5 and the median at the 2 nd week was 5. Wilcoxon signed-rank test showed that Z=-2.104, P <0.05. However, there were no statistically signi cant differences in PSQI scores between baseline and the 8 th week, the 2 nd week and the 8 th week.
This study also explored the PSQI subscales, and found that statistical differences existed in sleep latency between baseline and the 2 nd week, the 2 nd and the 8 th week, Z=-3.748, Z=-2.436 respectively. There were statistical differences in sleep duration between baseline and the 2 nd week, baseline and the 8 th week, and between the 2 nd and 8 th week, Z=-2.179, Z=-2.745, Z=-3.947 respectively. Statistical differences in daytime dysfunction existed only between baseline and the 8 th week, Z=-2.408 ( Figure 1).

Relationship between time, latitude and different symptomatic groups of PHQ-9, GAD-7, PSQI
Cochran's Q test was used to explore the similarities and differences among different symptomatic groups of PHQ-9, GAD-7 and PSQI at different time and latitudes. The results showed that based on PHQ-9 scores, a total of 62 participants were included. The percentages of the symptomatic group at baseline, the 2 nd and 8 th week were 25.8%, 30.6%, and 37.1%, respectively, χ2=5.286, p=0.071>0.05. There were no signi cant differences at these three time points for the symptomatic group. Based on scores on GAD-7, a total of 61 participants were included. The proportions of the symptomatic group at baseline, the 2 nd and 8 th week were 14.8%, 21.3%, and 19.7%, respectively, χ2=1.733, p=0.42>0.05. There were no signi cant differences at these three time points for the symptomatic group.
Based on PSQI scores, 43 people were included. The proportions of the symptomatic group at baseline, the 2 nd and 8 th week were 34.9%, 37.2%, and 44.2%, respectively, χ2=2, p=0.368 >0.05. There were no signi cant differences at these three time points for the symptomatic group ( Figure 2).

Personal characteristics and emotional changes in the polar environment
Relationship between personal background information and PHQ-9, GAD-7, PSQI and its subscales This study also explored the in uence of personality traits on mood and sleep conditions, and investigated whether different individual characteristics play a role in changes in mood and sleep. A total of 29 people completed the relevant assessments, and the missing data were discarded. This study applied a generalized linear mixed model to analyze data from repeated measures. The following factors were taken into consideration: age, marital status, level of education, current living conditions, personality traits, and coping styles.
The results indicated that the model exploring relations between PHQ-9 scores and personal background information such as age, marital status, education level, personal income and current living conditions etc., was not statistically signi cant. The relation between marital status and GAD-7 scores was statistically signi cant (F=5.065, p<0.05).
Speci cally, singleness was associated with lower scores on GAD-7. Personal conditions and the PSQI total scores were not related in a statistically signi cant way, instead were related to the subscales of PSQI signi cantly, including sleep quality, sleep latency, sleep duration, and sleep e ciency ( Table 1, Table 2). Pairwise comparison indicated that compared with individuals with income> 10,000 RMB per capita, those with an income of 5,000-10,000 RMB per capita had worse sleep quality. (coe cient 0.52, p<0.05). In terms of sleep latency, compared with individuals having an income> 10,000 RMB and a graduate degree or above, those with an income of 5,000-10,000 RMB per capita (coe cient 1.002, p<0.05), and a bachelor degree (coe cient 0.75, p<0.05) had longer sleep latency while those with a high school education (coe cient -0.87, p<0.05) had shorter sleep latency. With the increase of age (coe cient 0.04, p<0.01), the total sleep duration decreases gradually. Those with an undergraduate degree (coe cient 0.38, p<0.03) had shorter duration of sleep time compared with participants with high school education (coe cient -0.54, p<0.05).
In terms of sleep e ciency, with the increase in age (coe cient -0.07, p<0.05), sleep e ciency improved. People living alone (coe cient -1.71, p<0.05), living with parents (coe cient -1.84, p<0.05) experienced better sleep e ciency than those living with partners and children. People who had an income <5000 RMB (coe cient -0.94, p<0.05) and those with an income between 5000-10000 RMB (coe cient -0.77, p<0.05) had better sleep e ciency than individuals having an income greater than 10,000 RMB.  (Table 3).  Relationships between coping styles and PHQ-9, GAD-7, PSQI and its subscales The relations between coping styles and PHQ-9, GAD-7 and PSQI scores have statistical signi cance (Table 5).
Further analysis showed that coping styles have an impact on sleep quality, sleep disturbances and daytime dysfunction. The help-seeking factor (coe cient 0.91, p<0.05) is positively correlated with the sleep quality score; the problem-solving factor (coe cient -1.46, p<0.05) is negatively correlated with the sleep disturbance score, and the help-seeking factor (coe cient 0.77, p<0.05) is positively correlated with the score of sleep disturbance. The selfblaming factor (coe cient 3.08, p<0.05) is positively correlated with the daytime dysfunction score (Table 6). Discussion Changes in levels of anxiety, depression and sleep condition in polar environments It was found in the study that with the extension of polar time, the expedition members' level of depression increased, and the anxiety symptoms did not change signi cantly at every time of evaluation. This is similar to previous studies.
Moraes and colleagues conducted a 50-day study regarding Antarctic expedition and found that the emotional states of the team members changed signi cantly as the time on board increased, manifested by decreased vitality, increases in depression and the average score of negative dimensions. There were no obvious changes in other dimensions [15]. Mairesse monitored the sleep condition of Antarctic expeditioners using Polysomnography and selfrating scales, starting from 2-3 months after they arrived at the station. The Polysomnography showed that the overall sleep quality and various dimensions of the team members remained stable, while the sleep latency was extended. The self-rating scale showed the same results [16]. This is consistent with the conclusions of this study, there were no signi cant differences in sleep assessment results between the 2nd and 8th week. Furthermore, it was found that there is a difference in sleep condition between the baseline and the second week, which may be caused by the obvious changes in the sleep environment initially. With the adaptation to the environment, the sleep condition gradually went back to normal. Sleep latency at the second week was longer than that at the baseline.
Many studies have shown that the prevalence of depression is higher at higher latitudes, and migration from low latitudes to high latitudes can increase the prevalence of depression [17][18][19]. Meanwhile, years of residence may be negatively correlated with the prevalence of depression [20]. These studies suggest that higher latitudes population is more likely to suffer from depression than lower latitudes population. However, as years of residence becomes longer, their depression situation decrease. The prevalence of depression in people newly moved to higher latitudes will increase, and subsequently decrease gradually. In these studies, the in uence of the latitude on people's mood is longterm, usually counted in years, and the short-term effect may not be obvious. Moreover, latitude may only be an intermediate variable, because high latitudes will bring risks such as reduced daylight hours, which are risk factors for depression originally. Many studies have found that lack of sunlight can increase the risk of depression [21,22], and a lack of sunlight can lead to vitamin D de ciency, which is also considered to be closely related to depression [23]. Vitamin D supplementation can relieve depressive symptoms [24]. A meta-analysis showed that light with an intensity of ≥5000lux for no less than 30 minutes is effective for improving condition of depression [25].
The in uence of personal background information, personality traits and coping styles on anxiety, depression and sleep Personality is a unique integration pattern that constitutes a person's thoughts, emotions, and behaviors. This unique pattern contains a stable and uni ed mental trait that distinguishes an individual from others [26]. Its formation has both innate factors and in uence from the environment. Everyone's personality is unique and stable, so everyone responds to the same environment differently but stably, which is predictable to a certain degree. Having a good, wholesome personality and mature coping styles are necessary for individuals to deal with harsh environments.
This study showed that team members' depressive symptoms became severe over time and had no apparent relationship with the team members' personal background information and coping styles. Personality tests indicate that people with high level of tension and neuroticism are more likely to experience depression. The results suggest that time spent in the polar environment may be an independent risk factor for depressed mood. It is necessary to pay attention to the mental health of expedition team members who stay in polar region in a long-term, to intervene in a timely manner.
Studies have found that being single can reduce overall level of anxiety. There are also inconsistencies in previous research, Adamczyk et al. [27] believe that it is incorrect for the society to associate being single with negative stereotypes such as feeling miserable, insecure, and dissatis ed with life [28]. Especially for people who are willing to be single, being single is an expression of individuality and personal attitude [29]. On the other hand, Vivian et al. [30] believe that married people can expand shared social networks, share and buffer stress with their partners, while single persons may participate in more social activities due to social pressures and other reasons. Meanwhile, single persons are prone to experience stress related to loneliness, which in turn are more likely to bring about anxiety. Previous studies on the relationship between singleness and anxiety were conducted in a normal social environment.
However, the current study was conducted in an enclosed and isolated environment, there are differences between these two. The authors of the study believe that compared with married team members, single team members do not have a commitment to their partners, and have less responsibility for the family. This instead become an advantage for them in the enclosed and isolated polar environment. Besides, for married people, the pattern of sharing stress with partners is not applicable in the polar regions. Instead, they adapted to the environment slowly and experienced increased level of anxiety. Research indicates that people who are enthusiastic, outgoing, and unorthodox tend to have lower levels of anxiety, while people who are easily excited, full of fantasy, and well-versed are more likely to be anxious. This has been supported by other studies, high a nity, frankness, and introverted persons are prone to experience more distress when faced with interpersonal stressors [31,32]. It is worth noting that a study concerning high altitude regions showed that factor M (Abstractedness) is negatively correlated with anxiety state [33]. This suggests that personality traits are still inconsistent in research of anxiety levels in extreme environments.
Generally speaking, middle-income (5000-10000 RMB), undergraduate degree, increase in age, and living with partners and children are prone to have adverse effects on sleep. A meta-analysis showed that about 6.8% of adults sleep less than 6 hours, and this proportion expands to 25% with the increase in age [34]. Ohayon et al. found that as people become older, their total sleep time, sleep e ciency, slow-wave sleep, rapid eye movement (REM) sleep and REM latency decrease [35]. Not only that, how is the family environment, whether it is crowded also in uence sleep [36]. The sleep quality of individuals who live with their partners and children is worse than that of people who live alone or with their parents. This may due to the fact that persons have the responsibility for taking care of their partners and children when living with them. There is no need to think about taking care of families when living alone. When living with the parents, individuals are often the ones being taken care of, so they sleep well. Some studies indicate that high income is bene cial for sleep, but considering other bene ts brought by high income (such as better education, medical care, and the environment), high income will affect sleep [36]. Therefore, high income does not bring about a good sleep directly. However, it can improve other factors that in uence sleep. We found that middleincome team members sleep worse than higher-income (>10,000 RMB) and low-income (<5000 RMB) team members. This was not completely in line with the nding that high-income can improve sleep quality. This may be related to the possibility that middle-income team members have more pursuits in life and it can be di cult for them to achieve. Furthermore, personality also has an impact on sleep. For example, people with high scores on factor Q4 (Tension) have worse sleep quality. These people have characteristics such as nervousness and neuroticism, which are consistent with other research ndings [37]. Suspicious and alert people fall asleep faster, have higher sleep e ciency, and sleep better, and so on. Different ways of coping with problems can also affect sleep. Those with high scores on factor of problem-solving are less likely to experience sleep disturbances, whereas the factors of help-seeking and self-blaming have a negative impact on sleep. These results indicate that if a person adopts a more mature way of dealing with incidents, and his personality is relatively stable, then his sleep quality may be better.

Conclusion
We found that depressive symptoms can change over time, and in the early two weeks, the quality of sleep decreased. Different personal background, personality and coping styles also in uenced mood and sleep to some extent.

Limitations
In this study, because there were no professional psychiatrists on board, the anxiety and depression scales were selfrated, the validity of the assessment was affected. At the same time, the ocean waves brought disturbances to some people during their sleep. Also, we were unable to monitor participants' sleep condition using Polysomnography, and the analysis of sleep condition cannot be combined with objective data. Although this study found that both personality traits and coping styles have an impact on mood and sleep, the results should be interpreted with caution. Continuous research concerning expedition team members on a large-scale is needed.

Declarations
Ethics approval The study was approved by the Ethics committee of Zhongshan Hospital Fudan University, China. (IRB No. B2020-204R).

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions TW, WL and JJ participated in the design of the study. JW carried out data collection. HL and YL performed the statistical analysis and participated in the coordination of the study. WJ, HL, LQ, TW and DG drafted the manuscript and revised the manuscript. WL and JJ provided critical revision of the manuscript for important intellectual content.

Funding
Not applicable.

Consent for Publication
All the authors agree to publish. the different symptomatic groups of PHQ-9, GAD-7 and PSQI