Introduction

Sunlight plays a fundamental role in the regulation of various physiological processes, particularly through its influence on the circadian rhythm. This rhythm is primarily regulated by the light-dark cycle, which in turn affects the secretion of melatonin, a hormone critical for sleep onset and quality [1]. Exposure to natural light, especially sunlight, can modulate sleep patterns by affecting the synchronization of the circadian clock with the external environment. Research has shown that light exposure during the day, particularly in the morning, is linked to improved sleep outcomes, including better sleep quality, faster sleep onset, and longer sleep duration [2].

Circadian rhythms are governed by the suprachiasmatic nucleus (SCN) located in the hypothalamus, which serves as the central clock of the human body. The SCN responds to light signals received by the retina, influencing the production of melatonin and other neurochemicals associated with alertness and sleep [3]. Exposure to light is an important clue for the circadian clock, aligning it with the solar day and promoting an earlier start to sleep. This may help individuals who experience delayed sleep phase syndrome or suffer from social jetlag, defined as the misalignment between circadian and social clocks that often results in chronic sleep loss [4] due to irregular sleep patterns [5, 6].

Beyond the characteristic variables relating to sleep quality, such as total sleep time, sleep latency, and efficiency, one of the sleep variables often little examined is the midpoint of sleep, which refers to the intermediate point between sleep onset and awakening. The midpoint of sleep is closely linked to circadian rhythms and can provide information about the timing of sleep, particularly in the context of social jetlag or delayed sleep phase syndrome [4]. Sun exposure, especially in the morning, can shift the midpoint of sleep earlier, promoting more synchronized and restorative sleep. On the contrary, insufficient exposure to natural light can delay this midpoint, leading to misalignment between biological and social time, often resulting in chronic sleep deficits [7,8,9].

Exposure to sunlight has decreased due to work activities and sedentary lifestyles in front of screens, reducing exposure to natural light during the day [10]. The COVID-19 pandemic has worsened this situation, drastically changing daily routines around the world, with lockdowns and social restrictions limiting opportunities for outdoor activities and exposure to natural sunlight [11]. Many individuals shifted to indoor environments for extended periods due to remote work or unemployment, significantly reducing their exposure to morning and afternoon sunlight [12]. This sudden lifestyle change likely disrupted sleep patterns for many, exacerbating issues like delayed sleep phases, increased sleep latency, and reduced sleep efficiency. Even after the pandemic, we believe that altered routines, driven by social isolation measures, remote work, and other containment strategies, may have led to long-term reductions in sunlight exposure [13, 14]. During the pandemic, many individuals adopted new habits that limited outdoor activities [13], and those who continued working from home [14] or maintained pandemic-related routines may have experienced ongoing sleep disruptions due to reduced natural light exposure. This diminished sunlight exposure could affect circadian regulation and overall sleep quality.

Therefore, understanding the relationship between the timing of sunlight exposure and its interfaces with sleep parameters is crucial, particularly in light of how the pandemic may have long-lasting effects on sun exposure habits. This study aims to evaluate how sun exposure at different times of the day affects various sleep parameters, including total sleep time, sleep latency, sleep efficiency, midpoint of sleep, and sleep quality. The hypothesis is that exposure to the sun at different times of the day affects sleep quality differently, with morning exposure having a greater impact.

Methods

Study design

This is a population-based, cross-sectional survey that was carried out between October and December 2020 in two medium-sized cities (Ouro Preto and Mariana) in the south-central Minas Gerais region, also known as the Iron Quadrangle, which is one of the largest iron ore producing regions in Brazil [15]. The survey was conducted using stratified, multistage probability cluster sampling.

The sample design was executed by the conglomerate in three phases: census sector, household and resident. The National Household Sample Survey (PNAD) [16], the Family Budget Survey (POF) (IBGE 2020), the “Saúde em Beagá” survey [17], and, more recently, the “EPICOVID19” study[18] served as the foundation for this design. Consequently, the census sectors were regarded as primary sampling units in the study design, chosen with a probability proportional to the number of households, and the number of households was used as a measure of size, derived from the summary of the 2010 census of population. To reduce the possibility of selecting a sample from non-representative sectors, prior stratification was carried out before selecting the primary units. This involved taking into account the average income based on information from the Brazilian Institute of Geography and Statistics (IBGE) 2010 demographic census. Consequently, the final sample was guaranteed to be representative of the three socioeconomic strata (wages less than one, wages between one and three, and wages greater than four).

The updated list of current household units in the primary sampling units (selected census sectors) was used to systematically select the households that made up the secondary sampling units. Private households with occupants make up the household units. Following the census sectors’ selection, the household selection interval (k) for the systematic sampling was computed using the following formula: k = Ni/(xi/ni), where Ni is the total number of households in the census sector, xi is the sample size, and ni is the number of households to be selected. In this manner, the entire geographic area was covered and a proportionate number of homes per sector was obtained. After choosing the first household in the census sector based on IBGE indications, the next household was systematically sampled by the household selection interval (k).

The individuals, chosen through a random sampling procedure, constituted the tertiary sampling units. A list of every adult resident in the chosen home was created, and one resident was chosen at random to take part in the study. The sample calculation indicated that 1,464 people would be the minimum sample size.

We assessed 1,762 people who represented adult residents of the two cities’ urban areas throughout the data collection process. The interviewer used an electronic form to conduct in-person interviews with the residents while adhering to national protocols to stop the coronavirus from spreading, like wearing personal protective equipment. The questionnaire was broken down into sections based on general health conditions, living habits, sleep quality, and sociodemographic and economic factors.

Outcome variables: Sleep parameters

The Pittsburgh Sleep Quality Index (PSQI), a validated questionnaire intended to measure sleep quality and disturbances over 1 month, was used to assess sleep parameters [19]. The 19 items that make up the PSQI are divided into seven categories: (1) subjective quality of sleep; (2) sleep latency; (3) duration of sleep; (4) habitual sleep efficiency; (5) disturbances of sleep; (6) use of sleeping medication; and (7) dysfunction during the day. Every element is assigned a number between 0 and 3, with higher numbers denoting lower-quality sleep. The seven components add up to the total PSQI score, which is a number between 0 and 21. Higher scores denote lower overall quality of sleep. Sleep quality was classified as “good” (PSQI score ≤< 5) or “poor” (PSQI score > 5) in this study, and used the PSQI total score as a continuous outcome variable [20].

In addition to the total PSQI score, the following specific sleep parameters were used as outcome variables: Total sleep time represents the amount of sleep, in minutes, reported by participants each night. It was measured through the PSQI’s sleep duration component. Sleep latency, in minutes, it takes for participants to fall asleep after going to bed. Sleep efficiency represents the percentage of time spent asleep relative to the total time spent in bed, which is calculated as: (time spent in bed/total sleep time)​×100. The midpoint of sleep reflects the halfway point between the time the participant falls asleep and the time they wake up, which is a crucial indicator of circadian phase alignment. It was calculated using the following formula: sleep onset time + (total sleep time/2​). Sleep onset time was determined based on self-reported bedtimes, and wake-up times were recorded in the PSQI. A delayed midpoint of sleep suggests a misalignment of the circadian rhythm.

The decision to focus on these specific parameters, among the seven evaluated by the PSQI, was made due to their strong association with physiological markers of sleep quality and their implications for metabolic and circadian health. Other components, such as subjective sleep quality or sleep disturbances, though informative, were not included in this analysis as they do not provide direct quantifiable data on sleep duration or circadian timing, which are central to the objectives of this study.

Exposure variable: Sunlight exposure

Sunlight exposure was evaluated based on participant self-report. Each participant was asked to report how often and for how long they were exposed to sunlight at different times of the day: before 10 a.m., between 10 a.m. and 3 p.m., and after 3 p.m. For each period, the following questions were asked: “From Monday to Sunday, how many days per week are you exposed to sunlight before 10 a.m.?” and “On those days, how long (in minutes per day) are you exposed to sunlight?“. Similar questions were asked for the time intervals between 10 a.m. and 3 p.m. and after 3 p.m. Based on these responses, the total amount of time exposed to sunlight for each period was calculated. The daily average sunlight exposure for each period was computed using the formula: [weekly frequency of sunlight (0 to 7 days) x daily time of sunlight (minutes)/7]. This provided the average daily sunlight exposure for the three distinct periods: morning (before 10 a.m.), midday (10 a.m. to 3 p.m.), and evening (after 3 p.m.). For the statistical models, sunlight exposure was treated as a continuous variable and analyzed in increments of both 10 and 30 min to assess its impact on sleep parameters. The use of different increments allowed for a more detailed understanding of the dose-response relationship between sunlight exposure and sleep outcomes.

Covariates

The variables for potential confounding controls in the analysis of the relationship between sunlight exposure and sleep parameters were included in the questionnaire. Sex (male or female), age group (18–34 years; 35–59 years; ≥ 60 years), marital status (single or married); living status (living alone or not living alone); education level ( 0 to 8 years; 9 to 11 years or > 12 years of study); employment status during the pandemic (employed or not employed); and work-from-home schedule (percentage of active workers who were working from home) were the sociodemographic and economic variables evaluated. The IBGE categories for self-reported race and skin color were used for evaluation [21]. The participants were categorized into white, black, brown, indigenous, and yellow.

The following health conditions and lifestyle variables were assessed. Chronic diseases were assessed by self-reporting the following diseases: hypertension, diabetes, asthma, lung disease, chronic kidney disease, cancer, heart disease, or thyroid disease, which were dichotomized into morbidity (reporting at least one disease) and without morbidity (no disease). In addition, the following lifestyle variables were evaluated: whether or not they currently smoke and whether or not they currently drink alcoholic beverages.

Movement behavior, physical activity, and sedentary behavior were also assessed. Sedentary behavior was measured by the total time spent sitting during weekdays and weekends. The measurement was carried out using the following question: ‘Currently, how much time on average do you spend sitting per day? (Include time used for cell phone, TV, computer, tablet, books, car, and bus)’ and classified considering a cut-off point of ≥ 9hours. Physical activity during leisure time was evaluated based on the VIGITEL questionnaire and classified according to the degree of activity (active when meeting the recommended weekly minimum of 150–300 min of moderate-intensity aerobic physical activity or 75–150 min of vigorous-intensity aerobic physical activity otherwise). Participants who mainly engaged in light physical activity, those who did not meet these thresholds, were classified as “inactive in leisure time”. Light physical activity was considered insufficient to meet the recommended guidelines, in line with current public health recommendations [22].

Self-reported weight (kg) and height (cm) were used to calculate the body mass index (BMI). The BMI classifications were as follows: underweight (BMI < 18.5 kg/m2 if < 60 years old; BMI < 23.0 kg/m2 if ≥ 60 years old), eutrophic (BMI 18.5–24.9 kg/m2 if < 60 years old; BMI 23.0–28.0 kg/m2 if ≥ 60 years old), and overweight (BMI ≥ 25.0 kg/m2 if < 60 years old; BMI ≥ 28.0 kg/m2 if ≥ 60 years old). For adults and the elderly, respectively, according to WHO [23] and PAHO [24]. The Generalized Anxiety Disorder Scale (GAD-7) and the Patient Health Questionnaire (PHQ-9) scales were used to assess the symptoms of depression and anxiety, respectively. Scores of 10 or higher on both scales were taken into consideration when determining the presence of symptoms of depression and anxiety [25, 26].

Statistical analysis

The sample weight was calculated to adjust the natural weights of the design and/or rectify issues arising from the absence or refusal to respond. This was accomplished by assigning distinct weights to the sample elements, which corresponded to the inverse of the product from the probabilities included in the various selection stages [27]. When calculating the sample weights, we considered the probabilities of inclusion of the sample elements in the three stages: (1) probability of the census sector being randomized; (2) probability of the household being randomized; and (3) probability of the individual over 18 years old being randomized. The adjustment was implemented to compensate for the non-response loss of interviews and to calibrate the sample weight for the population totals by sex and age group to align with the population projections for 2019. Therefore, all statistical analyses were conducted using the “svy” package in Stata® version 15.0, taking into account the complex survey design and sample weighting factors. A significance level of 0.05 was set for all analyses.

Moreover, using the online program Dagitty, version 3.2, a theoretical causality model based on a directed acyclic graph (DAG) was created by the exposure variable (sunlight), outcome (sleep parameters), and covariates. There were established causal relationships between the variables, shown by arrows (Fig. 1). Every variable in the DAG was represented by a rectangle, and each color had a distinct meaning: the response variable was blue, circled by black, and the exposure variable was green. Additionally, variables thought to be potential confounders were included: the outcome variable’s antecedents are in blue, and the exposure and outcome variables’ antecedents are in red. A minimum set of confounding variables was chosen to fit the analyses using the backdoor criterion to prevent needless adjustment, erroneous associations, and estimation errors [28]. The model was adjusted by the following minimum and sufficient set of variables: age (continuous variable), sex (male or female), education (0 to 8 years; 9 to 11 years or > 12 years), employment status (not workers or active workers), presence of chronic diseases (no or yes), and movement behaviors (physical activity, minutes day of moderate to vigorous physical activity; and sedentary behavior, hours/day of total sitting time).

Fig. 1
Fig. 1
Full size image

Direct acyclic graph (DAG) of association with sunlight exposure and sleep quality, with covariates. COVID-Inconfidentes Study, 2020 Note: The variables in green with the symbol “►” inside the rectangle are explanatory; the variable in blue with the letter “I” inside the rectangle is the outcome variable; the variables in red without the symbol “►” inside the rectangle are the antecedents of the outcome and exposure variables. Variables in white are the collider variables

For the variables defined by the DAG, both unadjusted and adjusted multivariable linear regression models were employed to assess the relationship between sunlight exposure and sleep parameters. The outcome variables included were total sleep time (minutes), sleep latency (minutes), sleep efficiency (percentage), midpoint of sleep (hours), and the PSQI total score (range: 0–21). Sunlight exposure was analyzed in increments of 10 and 30 min for three different periods: before 10 a.m., between 10 a.m. and 3 p.m., and after 3 p.m. To generate these exposure increments, a formula was applied where the average daily exposure time was divided by 10 or 30, allowing for the assessment of the impact of incremental increases in sunlight exposure on sleep outcomes.

Before running the regression models, assumptions of linear regression were tested. Linearity between the independent and dependent variables was checked, and residual plots were examined to confirm homoscedasticity. The normality of residuals was assessed using the Shapiro-Wilk test and Q-Q plots, while the Durbin-Watson test was applied to ensure independence of residuals. Multicollinearity was evaluated through the variance inflation factor (VIF), with all values indicating no significant collinearity between explanatory variables. Finally, Cook’s distance and leverage statistics were used to identify any potential outliers or influential data points, which were found to be within acceptable limits. All assumptions were satisfied, ensuring the validity of the regression models used in the analysis.

In regression analyses, beta coefficients (β) for sleep outcomes measured in hours were converted to minutes to facilitate interpretation. This conversion was calculated by multiplying the beta coefficient by 60 (minutes/hour).

Results

Characteristics of study participants

The sociodemographic, health, and behavioral characteristics of the study population are presented in Table 1, which also displays the midpoint of sleep and sun exposure across different periods of the day. The sample comprised 1,762 study participants, the majority (74.4%) of whom self-reported their skin color as brown, black, yellow, or indigenous; more than half of them (53.2%) were married; and had more than eight years of education (70.8%). The participants’ ages ranged from 35 to 59 years (45.6%), 51.9% were women, and 39.7% had nine to eleven years of education. The participants’ monthly family income was less than two minimum wages (41.1%) and they had no chronic diseases (60.2%). As far as their occupation went, 52.5% of them were employed. Regarding their habits, 58.2% of them consumed alcohol, and 17% of them smoked at the moment. In addition, 15.3% engaged in sedentary behavior and 69.2% were physically inactive in leisure time. Furthermore, 52.5% of participants had trouble sleeping, and 23.4% and 15.8% of participants, respectively, reported having symptoms of anxiety and depression (Table 1).

Table 1 Sociodemographic, health, and behavioral characteristics according to midpoint of sleep and sunlight exposure, during the COVID-19 pandemic.COVID-Inconfidentes study, 2020
Full size table

No significant differences were observed for the midpoint of sleep across sexes, with both men and women averaging 4:48 h. However, men had significantly more sunlight exposure before 10 a.m. and after 3 p.m. than women. Age was also a factor, with older individuals (≥ 60 years) having an earlier midpoint of sleep and greater sunlight exposure in the morning than younger participants (18–34 years). Married participants had earlier midpoints and more morning sunlight exposure than unmarried individuals. Other factors, such as education and health behaviors, showed distinct patterns. Participants with higher education had more midday sun exposure, while those who were physically active also reported higher exposure, particularly in the early morning. Mental health indicators, such as anxiety and depression, were associated with delayed midpoints of sleep and lower sunlight exposure, particularly in the morning and late afternoon. Sleep quality also played a role, with those reporting poor sleep quality having a later midpoint and lower sun exposure (Table 1).

Exposure to sunlight and sleep parameters

The most commonly observed bedtime was between 9:00 p.m. and 11:00 p.m. (61.6%), while the most frequent wake-up time was between 5:00 a.m. and 7:00 a.m. (62.5%). The midpoint of sleep was 5:18 a.m. (95% CI: 4:54 a.m. – 5:83 a.m.), with the highest frequencies occurring between 1:00 a.m. and 3:00 a.m. (49.4%) (Fig. 2). The mean total sleep time was 7.01h per day (95%CI: 6.81h – 7.21h), with an average sleep latency of 47.00 min (95%CI: 41.42–58.59), and a sleep efficiency of 84.28% (95%CI: 82.83–85.74). The mean PSQI total score was 6.33 (95%CI: 6.03–6.62). Regarding sunlight exposure, participants were exposed to an average of 53.60 min before 10 a.m. (95%CI: 40.58–66.63), 65.09 min between 10 and 15 a.m. (95%CI: 44.36–85.83), and 31.27 min after 3 p.m. (95%CI: 21.70-40.84) (Table 2).

Fig. 2
Fig. 2
Full size image

Circular histogram of sleep onset time, awake time, and midpoint of sleep frequency in adults. COVID-Inconfidentes, 2020 Legend: (A) Circular histogram of sleep start times. (B) Circular histogram of wake up times. (C) Circular histogram of sleep midpoint. Sleep midpoint reflects the halfway point between the time the participant falls asleep and the time they wake up: sleep onset time + (total sleep time/2​)

Table 2 Descriptive statistics of sleep parameters and sunlight exposure during the COVID-19 pandemic. COVID-Inconfidentes study, 2020
Full size table

In the linear regression analysis, the midpoint of sleep showed the strongest association with sunlight exposure, being significantly affected by exposure before 10 a.m. and after 3 p.m. However, no significant association was observed for sunlight exposure between 10 a.m. and 3 p.m. In the multivariate model, for every 30-minute increase in sunlight exposure before 10 a.m., there was a 23-minute reduction in the midpoint of sleep (-0:23 hh:mm; 95%CI: -0:39, -0:14; beta: -0.387; 95%CI: -0.644, -0.233). Sunlight exposure after 3 p.m. also exhibited an association, though less pronounced (-0:19 hh:mm; 95%CI: -0:36, -0:03; beta: -0.325; 95%CI: -0.600, -0.051). Furthermore, the total PSQI score was significantly associated with sunlight exposure before 10 a.m., with a 0.184-point reduction in the score for every 30-minute increase (beta: -0.184; 95%CI: -0.362, -0.006). Other sleep parameters, such as total sleep time, latency, and efficiency, did not show significant associations with the various sunlight exposure timeframes (Table 3).

Table 3 Linear regression results for association of sunlight exposure with sleep parameters during the COVID-19 pandemic. COVID-Inconfidentes study, 2020
Full size table

Discussion

The results of this study highlight the importance of morning sun exposure for regulating sleep midpoint and overall sleep quality. The significant reduction in sleep midpoint with morning exposure suggests that sunlight may play a crucial role in adjusting circadian rhythms. Our findings indicate that early morning light exposure can help align the internal circadian clock, contributing to healthier sleep patterns.

Adequate sleep is essential for overall health and well-being, influencing both mental and physical health. Among the various sleep parameters, the midpoint of sleep is a key indicator of circadian rhythm alignment. The midpoint, which marks the halfway point between sleep onset and waking, reflects an individual’s internal biological clock. Research suggests that deviations in the sleep midpoint, particularly delayed sleep, can lead to adverse health outcomes, such as metabolic disorders, cardiovascular disease, and psychological disorders [29,30,31,32]. Furthermore, alterations in the timing of sleep midpoint have been strongly associated with metabolic dysfunction. Recent studies indicate that a later sleep midpoint is linked to changes in several metabolic pathways, particularly those involved in inflammation and oxidative stress [33]. For instance, research by [33] found that delayed midpoints are associated with metabolites like erythrulose, a product of advanced glycation, which may contribute to higher risks of diabetes and cardiovascular diseases. Additionally, delayed midpoints were inversely related to metabolites in the γ-glutamyl pathway, which plays a critical role in amino acid metabolism and antioxidant defense mechanisms. This suggests that disruptions in sleep timing could exacerbate metabolic stress and oxidative damage. Therefore, maintaining an early average sleep time can act as a protective factor, promoting healthier metabolic function and reducing the risk of related diseases [33].

The midpoint of sleep is particularly sensitive to disruptions in light exposure, with delayed midpoint timing being a marker of circadian misalignment [34]. This misalignment can contribute to poor sleep quality, difficulty waking up, and even daytime fatigue. In alignment with existing literature, our study confirms that morning sunlight exposure, especially before 10 a.m., is linked to earlier sleep timing and an earlier midpoint of sleep. Studies have demonstrated that exposure to morning sunlight, particularly before 10 a.m., is associated with earlier sleep onset and a more aligned circadian rhythm. This highlights the critical role of morning light as an external cue (zeitgeber), which helps to reset the internal biological clock and mitigate the effects of social jetlag and circadian misalignment [35]. Despite its importance, limited epidemiological data is evaluating the direct impact of sunlight exposure on sleep parameters. This study highlights the substantial influence of sunlight exposure on the midpoint of sleep, with morning light exposure being the most impactful.

Morning sunlight, in particular, helps regulate the secretion of melatonin, a hormone crucial for sleep regulation, thereby improving sleep onset and sleep quality. Increased sunlight exposure also correlates with lower levels of daytime sleepiness and improved alertness during the day. In contrast, reduced exposure to sunlight, especially early in the morning, can have profound effects on the circadian system, delaying the onset of sleep and prolonging the midpoint of sleep. This delay creates a cascade of negative effects, including poorer sleep quality, increased daytime fatigue, and decreased cognitive performance [35, 36]. Our data showed that individuals exposed to sunlight before 10 a.m. had significantly lower PSQI scores, reflecting better overall sleep quality.

While the association between afternoon light exposure (after 3 p.m.) and sleep was weaker, we observed a significant reduction in the midpoint of sleep with greater late-afternoon exposure. This could potentially be explained by two mechanisms. First, afternoon sunlight exposure might signal the approaching end of the daylight period, reinforcing the circadian cycle. Alternatively, increased physical activity during afternoon sunlight exposure may contribute to improved sleep parameters, as physical activity is known to positively impact sleep efficiency and quality [37]. Although our study did not find a direct association between sunlight exposure and sleep efficiency, these factors may interact to influence overall sleep health.

Another important pathway through which sunlight impacts sleep is through its role in stimulating vitamin D production in the skin. Vitamin D plays a vital role in various physiological processes, including the regulation of sleep [38, 39]. Studies suggest that adequate levels of vitamin D are associated with improved sleep quality [38, 40,41,42] Menezes-Júnior and Sabião [10, 43], potentially due to its role in modulating inflammatory pathways and regulating mood [44,45,46]. Thus, sunlight exposure may enhance sleep both directly by aligning circadian rhythms and indirectly by boosting vitamin D synthesis, which in turn contributes to better sleep quality and overall health.

The robust methodology employed in our study, which included face-to-face interviews conducted during the COVID-19 pandemic, a household survey, and a representative random sample of residents from various socioeconomic strata, strengthens the findings. To develop theories that can bolster the analysis’s underlying presumptions, the hypotheses were also meticulously defined by recent scientific literature and expressed in terms of counterfactuals.

A key limitation of this study is the assessment of the sleep midpoint as an overall measure, without distinguishing between the midpoint of weekdays and weekends. This distinction could have provided more detailed indicators of the variability caused by social jetlag, which typically arises from differing sleep schedules on workdays versus free days. However, the overall midpoint used in this study still offers valuable information, representing the average sleep behavior over the entire week. This broader measure can capture general circadian alignment and habitual sleep timing, making it a useful indicator for long-term health outcomes. Another limitation is that exposure to artificial light, including screen time, was not controlled in this study. Light exposure, especially from screens in the evening, could influence both sleep onset and the circadian rhythm, potentially affecting the sleep midpoint and its interpretation. Therefore, future research should not only explore the specific impacts of weekday versus weekend midpoint differences but also control for environmental factors like artificial light exposure and screen time, which may influence these relationships. Additionally, sleep efficiency was estimated using the Pittsburgh Sleep Quality Index (PSQI), which does not directly measure total sleep duration or nocturnal awakenings, as it relies on self-reported bedtime, wake-up time, and sleep latency. More precise assessments of sleep efficiency require objective measures such as actigraphy or polysomnography, which were not feasible in this study due to data collection limitations during the COVID-19 pandemic. Future research should incorporate these objective methods to improve the accuracy of sleep efficiency assessments.

Conclusion

This study underscores the significant influence of sunlight exposure, particularly in the early morning, on the regulation of sleep, specifically the midpoint of sleep and sleep quality. Given the increasing prevalence of circadian misalignment in modern society, promoting greater exposure to natural sunlight, especially before 10 a.m. can serve as a practical, non-pharmacological intervention to improve sleep, and reduce the negative effects of social jetlag and related health risks.