The increasing ubiquity of information and communication technologies has integrated digital environments into everyday physical spaces, enabling high-frequency shifts between digital and physical contexts, with potential implications for health. To investigate the emerging pattern of rapid shifts between environments and activities, digital and spatial behavioral patterns were analyzed using concepts from time-geography—particularly fragmentation—and normalized entropy measures from the digital phenotyping literature. This exploratory study used a standardized, pooled dataset, comprising two studies that collected objective smartphone logs of digital activity, GPS-based spatial context data, and daily self-reports of anxiety and mood. Associations were found between increases in fragmentation levels and mental health outcomes, with the direction of these associations often varying by gender and spatial context. Increased fragmentation of digital activities and digital activities within mobility episodes correlated with heightened anxiety in females but lower anxiety in males. This trend was reversed in the home context, where males with high fragmentation of digital activity reported more negative affect, while females did not. These findings point to complex associations between digital activity patterns and mental health, warranting further investigation.

Objectives Stress is known to be a modifiable key determinant of sleep quality. This study aims to investigate the impact of daytime physiological stress and nighttime environmental stressors on sleep architecture using a rule-based algorithm and wearable sensors in a real-world setting. Methods Twenty-one participants in Jerusalem were monitored over a 7-day period using the Dreem Headband, a wearable device that measures sleep via electroencephalography. Daytime stress was quantified through the Moment of Stress algorithm, utilizing electrodermal activity from the Empatica Embrace Plus bracelet. To assess potential sources of nighttime stress, we used indoor sensors to measure noise and temperature levels in the bedroom. We applied linear mixed-effects models with a random intercept at the individual level to assess the effects of these factors on sleep outcomes. Results An increase in the number of Moments of Stress per day from the 10th to the 90th percentile was related to a 6.57-point increase in the rapid eye movement (REM) sleep percentage and a 5.74-point decrease in the deep (N3) sleep percentage, indicating a shift in sleep architecture under heightened stress. Each additional minute of noise exposure above 65 dB(A) was linked to 1.20 more minutes of wake after sleep onset (WASO) (95% CI: 0.54-1.86) and to a 0.10-point increase in light (N1) sleep percentage (95% CI: 0.03-0.16). No association was detected between temperature and sleep. Conclusions Overall, these preliminary findings suggest that daytime physiological stress and nighttime noise may be associated with alterations in sleep architecture in naturalistic settings.

Background: Air pollution is a major public health concern associated with increased respiratory morbidity and mortality worldwide. It negatively impacts respiratory health, yet limited research based on accurate assessments with multiple sensors exists on its short-term effects. In this study we investigate the short-term association between exposure to multiple air pollutants and lung function in adults during daily mobility. Methods: Data on daily activities of 199 participants of the MobiliSense cohort living in the metropolitan area of Paris were collected between 2018 and 2020. Participants were equipped of two portable ambulatory monitors of air pollutants, a GPS receiver and an accelerometer during their mobility. Exposure to black carbon (BC), nitrogen dioxide (NO₂), nitrogen monoxide (NO), carbon monoxide (CO), ozone (O₃), and particulate matter (PM₂.₅) was recorded continuously. Lung function was assessed using spirometry tests conducted in the morning and evening over three days (N = 2,504), measuring forced expiratory volume in 1 second (FEV₁), forced vital capacity (FVC), and FEV₁/FVC ratio. Mixed-effects linear models were applied to assess the association between pollutant exposure at varying time lags (15 minutes to 6 hours prior to spirometry testing) and lung function outcomes. Results : Multipollutant models showed that increased exposure to BC and PM 2.5 was associated with a reduced lung function. A 1 μg/m 3 increase in BC within 1 or 2 hours prior to testing was associated with a decrease in FEV1 by 0.016 (95% CI -0.024, -0.008) and 0.021 (95% CI -0.034, -0.007) respectively. Similarly, increases in BC exposure over 2 hours to 4 hours were associated with a decrease in the FEV1/FVC ratio. Additionally, PM2.5 exposure 15 or 30 minutes or 1 hour before testing was linked to a 0.60 (95% CI -1.30, -0.03), 0.70 (95% CI -1.39, -0.09) and 0.50 (95% CI -1.10, -0.01) percentage points reduction in the FEV1/FVC ratio. Ozone (O3) was positively associated with FEV1 and FVC. No associations were found for other pollutants or time windows. Conclusion: This study provides evidence that short-term exposure to air pollutants – particularly BC and PM 2.5 – can impair lung function. The findings demonstrate that even brief increases in BC and PM 2.5 during daily mobility are associated with measurable reductions in FEV1 and the FEV1/FVC ratio. By assessing multiple pollutants across short exposure windows (15 minutes to 6 hours), this study strengthens causal inference regarding rapid respiratory effects and underscore the health relevance of transient pollution peaks encountered in urban environments, particularly from traffic emissions. Clinical trial number : not applicable.

Current ubiquitous information and communication technologies are reshaping human behaviours and could have profound implications for mental health. To investigate the dynamic patterns of daily interactions with the digital environment-particularly via smartphones-a novel, ecological approach is implemented which incorporates the use of innovative sensors. Data was collected from 31 healthy individuals from the general population living in Jerusalem who were tracked using objective smartphone logs of digital activity and daily self-reports of anxiety and depression symptoms. Significant associations were found between temporal patterns of smartphone usage and momentary anxiety symptomatology. Higher levels of overall phone usage in the morning hours were a significant predictor of anxiety while high levels of phone usage in the rest of the day were associated with decreased anxiety symptoms. A more nuanced analysis of usage types revealed that high levels of social and process-related digital activity in the morning were associated with a reduction in anxiety symptoms, with evidence of a dose-response relationship. These findings highlight the importance of context, user motivation, and the complex relationship between smartphone usage and mental health which warrant further research.

Past studies have investigated black carbon (BC) exposure during daily travels and often referred to it as individuals' daily BC exposure, while people spend most of their time indoors. The aim of this study was to investigate the determinants of in-home BC exposure in Grand Paris, France, using data from the MobiliSense cohort. The personal BC concentration of 278 participants was measured continuously at a 10-second resolution with a MicroAethalometer device (MicroAeth AE51) strapped on their shoulder for 4 days. We applied a multilevel model with a random intercept at the household level, since more than one individuals from the same house were participating in the study, to predict the hourly in-home BC concentration. The use of home appliances and the characteristics of the dwelling, along with outdoor air pollution concentrations, were modelled as predictors. Our findings suggest that some heating systems, such as open fireplaces and stoves using woods, and outdoor pollutants, were predictors of in-home BC concentration. Our study suggests that in-home pollution can be reduced by improving home features, such as heating systems, or by reducing the infiltration of outdoor air pollutants into indoor microenvironment either by using air filters or redesigning the ventilation system.

PURPOSE:This systematic review aimed to identify where children and adolescents are physically active using studies combining GPS, GIS, and accelerometry, and to examine how physical activity (PA) contexts vary by age and across countries. METHODS:A systematic search was conducted across four databases. Observational studies objectively assessing PA contexts in children and adolescents were included. Records were screened using AI-assisted active learning by two independent reviewers. Data extraction and study quality assessment were conducted independently by two reviewers, and findings were synthesized narratively by age group and country. RESULTS:Thirty-five studies from 12 countries were included. Children accumulated PA across a broad range of contexts, including recreational areas, home, and school, whereas adolescents' PA was more concentrated in school and home settings. Studies including mixed-age samples often reported higher PA in residential and recreational areas, partly reflecting differences in spatial definitions and classification. Across all age groups, school and home environments consistently contributed to PA, while commercial locations contributed little. Cross-country synthesis showed variation in dominant PA contexts, with residential areas appearing more prominent in the United States, school-based activity in Denmark and Switzerland, and more distributed patterns in countries such as the Netherlands and the United Kingdom. CONCLUSION:Children and adolescents are physically active across multiple contexts, but the relative importance of these settings varies by age and country and is influenced by methodological differences. Distinguishing age groups and improving consistency in spatial definitions are essential for accurately identifying PA contexts and informing context-specific interventions.

BackgroundExposure to circadian entrainers, such as sunlight, positively impacts sleep architecture, while exposure before bedtime to circadian disruptors, such as artificial light and smartphone use, can negatively affect sleep. However, real-world evidence from longitudinal observational studies that simultaneously capture these factors alongside electroencephalography-derived sleep stages remains limited. ObjectiveThis study aimed to investigate the effects of specific environmental and behavioral factors on sleep metrics and architecture by using sensor-based measurements over 7 consecutive days. Specifically, it examined day-to-day associations between (1) daytime sunlight exposure and (2) prebedtime artificial light exposure and smartphone use with selected sleep outcomes on the following night. MethodsA total of 21 participants from the Jerusalem metropolitan area were monitored continuously using the Dreem wearable electroencephalography for sleep staging, HOBO data loggers for light exposure, the wGT3X+ triaxial accelerometer for physical activity, and a dedicated mobile app to record smartphone usage. Sleep outcomes included total sleep time (TST), sleep onset latency (SOL), and the proportions of light sleep (N1) and deep sleep (N3). Sunlight exposure was defined as the number of hours above 1000 lux during daytime, and artificial light and smartphone use before bedtime were quantified as the duration of exposure accumulated in the 2 hours preceding sleep onset. Linear mixed-effects models with a random intercept at the individual level estimated the associations between these exposures and next-night sleep outcomes, adjusting for step count and other individual covariates. ResultsThe average TST was 420 (SD 85) minutes, and SOL averaged 17.6 (SD 18) minutes. Light sleep (N1) represented 6.6% (SD 2.1%) of sleep, and deep sleep (N3) accounted for 20.1% (SD 7.6%). Each additional hour of daytime sunlight exposure was associated with an increase of 10.67 (95% CI 0.6-20.7) minutes in TST the following night and with a 0.3 (95% CI –0.6 to –0.0) percentage-point decrease in light sleep (N1) percentage. No associations were found between evening artificial light exposure and sleep outcomes, while each minute of smartphone use before bedtime was linked to an increase in SOL of 0.2 (95% CI 0.0-0.4) minutes. ConclusionsThese findings emphasize the importance of daylight exposure for circadian alignment and the potential sleep-disruptive effects of evening digital engagement. This study demonstrates the feasibility and value of integrating wearable electroencephalography and environmental and behavioral sensors in naturalistic settings to uncover behavioral and environmental correlates of sleep architecture.

Social interactions and daily mobility are two behaviors that mutually transform throughout life. Understanding how they interact is particularly important for elderly people, who are more at risk to be spatially and socially isolated. From data collected among a sample of 225 people aged 60 and over living in the Paris region (France), we explore how their social interactions are embedded into their activity space using the new concept of “sociability space” (the geographical portion of the activity space consisting of places visited with social network member). Sociability places are found to less numerous, less diversified, more spatially dispersed and more decentered from residence than non-sociability places. Regarding traditional social network measures (degree, global clustering, etc.), stronger and more numerous correlations are observed with the structural characteristics of sociability spaces than with non-sociability spaces. Finally, variations in structural characteristics of sociability spaces according to people’s socio-demographic and residential profiles largely differ from those observed in structural characteristics of non-sociability spaces. This empirical study shows that “sociability space” indicators are complementary to social network indicators and provide a more precise picture of geography of social interactions than activity space considered as a whole. Such approach may help to identify particular groups at risk of social isolation, such as people living alone and those with a low level of education, and inspire public policies aiming at both promoting daily mobility and reducing social isolation.

Studies often investigate the long-term impact of social contacts on mental health in older adults, neglecting momentary effects. This research, grounded in the consideration of daily activity, explores how time-varying social contacts associate with momentary depressive symptoms among 216 older adults in the Île-de-France region. Employing a geographically-explicit ecological momentary assessment approach (GEMA), we collected participants' depressive symptoms, mobility locations, and social contacts data via smartphone surveys, GPS receivers, and mobility survey over 7 days. Bayesian mixed models with random effects at individual and daily levels, considering time autocorrelation, were employed. Participants engaging with social contacts exhibited lower depression not only immediately but also in the following hours. Interestingly, a longer duration of time spent with social contacts did not lead to a sharper decrease in depression levels. Notably, larger decreases were observed when the number of social contacts increased from one to two, especially with friends or family members.

The individual exposure to environmental noise in cities is usually assessed at the residential neighbourhood level with static, year-averaged strategic maps. This representation may underestimate noise exposure, given the mobility of individuals within the city and proximate sources of exposure. Our study employs high-resolution sensor analysis to observe how personal noise exposure differs from modelled noise map metrics, identify socioeconomical and behavioural determinants of exposure, and explore the impact of reallocating certain behaviours to others on daily personal noise exposure (LAeq,24h). Data on daily activities of 259 participants of the MobiliSense cohort living in the metropolitan area of Paris were collected between 2018 and 2020. Participants were equipped of a personal monitor for sound pressure, and of a GPS receiver and an accelerometer. Modes of transport were collected during a mobility survey. Results showed that noise exposure based on personal monitoring during space-time behaviours differed from modelled noise levels at residence. Participants were exposed to values below the recommended critical value for health of 55 dB(A) in urban areas in only 36

INTRODUCTION:Stress is nearly ubiquitous in everyday life; however, it imposes a tremendous burden worldwide by acting as a risk factor for most physical and mental diseases. The effects of geographic environments on stress are supported by multiple theories acknowledging that natural environments act as a stress buffer and provide deeper and quicker restorative effects than most urban settings. However, little is known about how the temporalities of exposure to complex urban environments (duration, frequency and sequences of exposures) experienced in various locations - as shaped by people's daily activities - affect daily and chronic stress levels. The potential modifying effect of activity patterns (ie, time, place, activity type and social company) on the environment-stress relationship also remains poorly understood. Moreover, most observational studies relied quasi-exclusively on self-reported stress measurements, which may not accurately reflect the individual physiological embodiment of stress. The FragMent study aims to assess the extent to which the spatial and temporal characteristics of exposures to environments in daily life, along with individuals' activity patterns, influence physiological and psychological stress. METHODS AND ANALYSIS:A sample of 2000 adults aged 18-65 and residing in the country of Luxembourg completed a traditional and a map-based questionnaire to collect data on their perceived built, natural and social environments, regular mobility, activity patterns and chronic stress at baseline. A subsample of 200 participants engaged in a 15-day geographically explicit ecological momentary assessment (GEMA) survey, combining a smartphone-enabled global positioning system (GPS) tracking and the repeated daily assessment of the participants' momentary stress, activities and environmental perceptions. Participants further complete multiple daily vocal tasks to collect data on vocal biomarkers of stress. Analytical methods will include machine learning models for stress prediction from vocal features, the use of geographic information systems (GIS) to quantify dynamic environmental exposures in space and time, and statistical models to disentangle the environment-stress relationships. ETHICS AND DISSEMINATION:Ethical approval (LISER REC/2021/024.FRAGMENT/4-5-9-10) was granted by the Research Ethics Committee of the Luxembourg Institute of Socio-Economic Research (LISER), Luxembourg. Results will be disseminated via conferences, peer-review journal papers and comic strips. All project outcomes will be made available at https://www.fragmentproject.eu/.

BACKGROUND:Social networks are known to protect against depressive symptoms in older adults. However, most research relies on retrospective self-reported depression measures and cross-sectional data, which may introduce bias. Ecological momentary assessment with longitudinal data overcomes these limitations by repeatedly measuring the subject's experience in the present moment. This study examined how social network characteristics relate to momentary depressive symptoms and their daily fluctuations in older adults. METHODS:We analysed data from 216 older adults in Paris, France, using the Healthy Aging and Networks in Cities and Promoting Mental Well-Being and Healthy Aging in Cities studies. Social network characteristics included network size and frequency of in-person and digital interactions per week. Depressive symptomatology was assessed using a daily smartphone survey of the Center for Epidemiological Studies-Depression over a week. Linear mixed-effect models estimated associations between social network characteristics and momentary depressive symptoms, while multivariable linear models examined relationships with daily symptom fluctuations. RESULTS:Network size and frequency of contact from digital communications per week were not associated with fewer depressive symptoms; however, there was suggestion that having more in-person contact was related to fewer depressive symptoms (exp(β) = 0.90, 95% CI 0.82 to 1.00). Moreover, having a larger social network (exp(β) = 0.91, 95% CI 0.85 to 0.98) and more in-person contacts (exp(β) = 0.96, 95% CI 0.93 to 0.98) were associated with less fluctuations in daily depressive symptoms, but no association for the frequency of contact from digital communications was observed. CONCLUSION:Findings from this study suggest that larger social networks and more in-person contact may promote more stable and better mental health among older adults.

Urban heat islands, combined with extreme heat waves, can pose a public health risk. During the 2003 heat wave in Paris, strong correlations were observed between nighttime outdoor air temperatures and mortality [1]. However, previous studies only focus on outdoor nighttime air temperatures when citizens are sleeping, without linking these observations with the heat stress they may have been exposed to during the day or in their apartment. Similarly, studies [2] highlight the relationship between air temperature and mortality during heat waves at the city level. Building on this, we aim to demonstrate the viability of using heat stress as a metric to assess its impact on individual physiological responses.This is one of the principals aims of the “Heat waves, urban Health islands, Health: a mobile sensing approach” (H3Sensing ANR research) project. Citizen science methods will be used in order to measure heat stress exposure over several days as well as individual physiological responses. Mobile measurements of microclimatic parameters [3] allow us to characterize and map heat stress exposure [4] in Greater Paris. Stationary measurements in apartments and individual surveys will complete the data set and combined with measured physiological data.Beginning in March, 2025, the citizen science study will last for two periods of 4 days in spring and in summer with a target group of 180 Paris inhabitants. The proposed communication will present the study methodology in terms of measurement device prototyping and data processing as well as preliminary analyses from the study.

Past epidemiological studies using fixed-site outdoor air pollution measurements as a proxy for participants’ exposure might have suffered from exposure misclassification.In the MobiliSense study, personal exposures to ozone (O3), nitrogen dioxide (NO2), and particles with aerodynamic diameters below 2.5 µm (PM2.5) were monitored with a personal air quality monitor. All the spatial location points collected with a personal GPS receiver and mobility survey were used to retrieve background hourly concentrations of air pollutants from the nearest Airparif monitoring station. We modeled 851343 minute-level observations from 246 participants.Visited places including the residence contributed the majority of the minute-level observations, 93.0%, followed by active transport (3.4%), and the rest were from on-road and rail transport, 2.4% and 1.1%, respectively. Comparison of personal exposures and station-measured concentrations for each individual indicated low Spearman correlations for NO2 (median across participants: 0.23), O3 (median: 0.21), and PM2.5 (median: 0.27), with varying levels of correlation by microenvironments. Generally, a large degree of individual variability in the correlation between personal and fixed-site measurements was found for all air pollutants. Results from mixed-effect models indicated that personal exposure was very weakly explained by station-measured concentrations (R2 < 0.07) for all air pollutants. The fit of the model was relatively high for O3 in the active transport microenvironment (R2: 0.25) and for PM2.5 in active transport (R2: 0.16) and in the separated rail transport microenvironment (R2: 0.20). Model fit slightly increased with decreasing distance between participants’ location and the nearest monitoring station.Our results demonstrate a relatively low correlation between personal exposure and station-measured air pollutants, confirming that station-measured concentrations as proxies of personal exposures can lead to exposure misclassification. However, distance and the type of microenvironment are shown to affect the extent of misclassification.

Several epidemiological studies have documented associations between air pollution exposure and cardiovascular responses, including adverse effects of air pollutants on blood pressure (BP). However, previous studies only considered the effect of specific air pollutants on resting BP, and did not sufficiently consider the independent effects of various air pollution species as well as their overall mixture effect. We addressed this gap in our MobiliSense sensor-based study among 273 participants living in the Grand Paris region. Participants wore personal monitors to assess personal exposure to particles [black carbon and particulate matter smaller than 2.5 μm in diameter (PM2.5)] and gaseous pollutants [ozone (O3), nitrogen monoxide (NO), carbon monoxide (CO), and nitrogen dioxide (NO2)] along with noise exposure. Participants were asked to measure their blood pressure (BP) at rest in the mornings and evenings for three days. Multilevel models with a random intercept at the individual level explored the relationship between air pollution exposure (averaged over the day) and change in resting BP from morning to evening. We also used the quantile G-computation method to estimate the joint effect of the mixture of targeted air pollutants on resting BP. Sensitivity analyses examined the associations between air pollution exposure averaged at different temporal scales before evening BP measurements and the outcome. A quantile increase in the mixture of air pollutants (PM2.5, NO2, NO, CO, and O3) over the day did not affect changes in systolic BP [-0.33 mmHg (95% CI: -3.31, 2.65)] and diastolic BP [-0.53 mmHg (95% CI: -2.66, 1.60)] from morning to evening. When shorter time exposure windows were considered (from a few minutes to a few hours), both NO and the mixture showed positive associations with the morning-to-evening DBP change in only some of the models. Future studies with sufficient repeated BP measurements for more participants should test the association at varying temporal scales (minutes to days) to better understand how air pollution exposure influences resting BP.

Introduction The residential environment is hypothesized to influence sleep quality within urban settings. Factors associated with the residential environment include air and noise pollution, area socioeconomic status, green and blue spaces, and other neighborhood features. This study seeks to quantify the association of selected environmental factors with sleep quality in the daily lives of 211 older adults residing in the Paris metropolitan area with sensor-based methods. Methods Participants’ sleep and physical activity were monitored over a 7-day period using 2 accelerometers. Ecological momentary assessment (EMA) surveys were administered 4 times a day to assess depressive and anxiety symptoms. Environmental factors surrounding participants’ residential addresses, including noise and air pollution, walkability, green and blue space availability, median income, and population density, were computed using geoprocessing methods. Hierarchical mixed models with a random intercept at the individual level were fitted to estimate the adjusted association between residential environmental factors and sleep outcomes [total sleep time (TST), sleep efficiency (SE), and wake after sleep onset (WASO)]. Potential effect modification of or mediation by physical activity and depression and anxiety levels were explored in the analyses. Results We observed an effect size of 1.4 more minutes of sleep for each increase of one thousand euro in neighborhood median income (Confidence Intervals: 0.35, 2.45). The average adjusted difference in total sleep time between the 10th and 90th percentiles of neighborhood median income was 23.6 minutes of sleep. Other environmental factors and depression and anxiety levels did not exhibit correlations with sleep outcomes. Conclusions The results reveal a positive association between median income at the residential level and TST, while no associations were identified for SE and WASO. In conclusion, these findings underscore the impact of neighborhood socioeconomic status on total sleep time within the context of urban living, highlighting the need for further research.

BACKGROUND:The composition of daily time-use physical behaviours-such as sedentary behaviour (SB), light physical activity (LPA), moderate-to-vigorous physical activity (MVPA), and sleep may be crucial for overall health and wellbeing. This study examined the associations between these time-use behaviours and both evaluative wellbeing (life satisfaction) and experienced wellbeing (momentary happiness, anxiousness, and tiredness). Evaluative wellbeing reflects an individual's overall life assessment, while experienced wellbeing captures real-time affective states. We investigated these associations by reallocating time among behaviours and assessing the predicted impact on wellbeing outcomes. METHODS:Time-use behaviours were obtained from 211 adults who wore Axivity AX3 accelerometers on their wrists for seven days. Participants also completed a survey to assess demographics and life satisfaction, before using a custom smartphone app to report their real-time happiness, anxiousness, and tiredness levels over seven days (at three random times each day). Time-use data were processed using UK Biobank machine learning algorithms. We employed Bayesian multilevel compositional analysis to investigate how time-use behaviours, and reallocating time between behaviours, were associated with both life satisfaction and momentary affective states. RESULTS:Increasing sedentary time (relative to other behaviours) over the week of observation was negatively associated with happiness and positively associated with anxiousness aggregated at the day level. Conversely, increasing the proportion of MVPA (relative to other behaviours) was associated with reduced anxiousness and tiredness. Substitution analysis showed that reallocating 20 min of SB to MVPA increased happiness by 0.12 units, 95% CI [0.01, 0.22] and reduced anxiousness by 0.20 units, 95% CI [-0.34, -0.07]. Additionally, reallocating 20 min of time spent in LPA to MVPA reduced tiredness by 0.16 units, 95% CI [-0.28, -0.03]. All affective states are reported on a 0-10 scale. No associations were found between time-use behaviours and life satisfaction. CONCLUSION:Our study shows that time-use behaviours, particularly reducing sedentary time and increasing physical activity, were more strongly linked to experienced wellbeing. Studies that focus solely on examining time-use behaviours and long-term wellbeing outcomes, such as life satisfaction (common in population studies), may overlook the dynamic interplay and immediate impacts of behaviours on wellbeing. While some associations were present, most of the tested relationships were weak or non-significant, suggesting that contextual factors like social and environmental conditions may play a greater role in shaping wellbeing. The next step is to explore sequential associations, such as behaviours occurring immediately before or after a momentary affect response is recorded.