Real-life behavioral and neural circuit markers of physical activity as a compensatory mechanism for social isolation

Social isolation and loneliness increase human mortality like known health risk factors such as obesity, alcohol consumption or smoking 15 cigarettes per day. Lack of social contact also impairs momentary affective well-being, impacts the structural and functional integrity of emotion regulatory brain networks and is a potent risk factor for mood disorders. Social distancing directives during the COVID-19 pandemic have exacerbated this public health problem and highlighted the importance of finding remedial strategies. One promising strategy to mitigate the negative affective consequences of lack of social contact is physical activity, a known protective factor for affective well-being and mental health with neural mechanistic links to emotion regulatory brain regions. In this study, we hypothesized that physical activity can compensate for the negative affective effects of lacking social contact in daily life and that individuals at increased neural and psychological risk for depression benefit most from this compensatory mechanism.

The authors situate their work within evidence that social isolation and loneliness elevate mortality risk and impair affective well-being, with observed alterations in emotion-regulatory brain networks and default mode network connectivity linked to loneliness and depression. Prior research supports physical activity as protective for mental health and affective well-being with neural links to emotion regulation. The pandemic context emphasized the need for strategies to counteract social isolation’s impact. References cited include studies on mortality risk from loneliness, neural correlates of social contact benefits, neurobiology of loneliness, DMN associations with perceived isolation, and links between loneliness and mental health problems.

Design: Multimodal epidemiological cohort with ambulatory assessment (accelerometry and smartphone-based ecological momentary assessment) and neuroimaging, plus a pandemic-period replication cohort. Ethics: Conducted per Declaration of Helsinki (2013). Institutional review board approvals (Heidelberg University, Medical Faculty Mannheim; study 1: 2014-555N-MA; study 2: 2019-733N). Written informed consent obtained. Participants compensated. Participants: Study 1: community-based cohort of 317 healthy young adults aged 18–28 years (57.09% female), recruited Sep 2014–Nov 2018, monitored for 7 days. Of these, 175 underwent resting-state fMRI after the ambulatory week; 76 provided follow-up data on perceived loneliness during the first COVID-19 lockdown. Study 2 (replication): 30 healthy adults aged 18–63 years, recruited Dec 2019–Jul 2022, monitored for 6 months during the COVID-19 pandemic in Germany. Participant flow: 356 assessed, 39 excluded (28 technical problems, 2 low EMA compliance, 9 missing questionnaires) yielding 317. For fMRI, 176 acquired, 1 excluded for movement, 175 analyzed. For follow-up, 76 provided COVID-19 data. Replication: 35 assessed, 5 excluded (4 technical issues, 1 low EMA), yielding 30. Measures and data collection: Physical activity measured via accelerometers (hip-worn in study 1; wrist-worn in study 2). Affective valence and real-life social contact repeatedly assessed via smartphone e-diaries; prompts were either location-based or random, with continuous GPS to assign contexts (for example, home, work) using a day reconstruction method. Affective valence was measured with two items (unwell–well; discontent–content), showing reliability via Spearman-Brown coefficients ρ=0.80 (within-person) and ρ=0.94 (between-person), and convergent validity (within r=0.66; between r=0.88). Additional questionnaires assessed socioeconomic status, physical activity (habitual hours/week), social network size, trait neuroticism (NEO-FFI-30), UCLA Loneliness Scale, trait anxiety (STAI-T), and schizotypal traits (SPQ). fMRI: resting-state scans analyzed to compute default mode network (DMN) within-network connectivity using the 100-region, 7-network Schaefer-Yeo parcellation; data processed with CONN toolbox v19c in MATLAB R2020a. Power analysis: Based on simulation studies for multilevel models, n≈200 with ≥30 level-1 observations provides 80% power to detect an effect size of 0.08; thus n=317 was deemed sufficient. Statistical analysis: Software: SAS v9.4 for statistical models; CONN toolbox and MATLAB for fMRI. Main within-person multilevel model (Study 1): outcome = momentary affective valence; predictor = momentary social contact (alone vs in company); moderator = momentary physical activity (mean milli-g in the 60 min prior to prompt). Covariates: time of day (linear and quadratic) and current location (level 1); sex, age, body mass index (level 2). Between-participant analyses: multiple linear regression predicting trait loneliness from social network size (predictor) and habitual physical activity (moderator: hours/week). Loneliness during COVID-19 lockdown modeled via ordinal logistic regression with proportional odds using participant-specific random slopes from the main interaction (social contact × physical activity) as predictors. Neural moderation: three-way multilevel interaction testing whether DMN connectivity further moderated the compensatory effect (social contact × physical activity × DMN connectivity). Study 2 replication employed the same main model during the pandemic period.

- Physical activity significantly moderated the relationship between momentary social isolation and decreased affective valence in daily life: β = 0.01; 95% CI = 0–0.02; P = 0.020. Greater recent physical activity reduced the affective loss associated with being alone. - Dose to offset deficit: Approximately 349 milli-g (g/1,000) of physical activity accumulated across 1 hour (e.g., walking ~3 miles per hour) was estimated to fully compensate the affective deficit associated with lacking social contact. - Replication: The compensatory effect was replicated in the pandemic-era sample (study 2) (p = 0.03; 95% CI = 0.02–0.04; P < 0.001). - Neural moderation: Individuals with higher resting-state DMN within-network connectivity (a risk phenotype for loneliness/depression) showed stronger compensation of the social-affective deficit through physical activity: β = 0.14; 95% CI = 0.01–0.26; P = 0.029. - Between-individual findings: Participants with small social networks who engaged in high habitual physical activity reported lower trait loneliness than those with low habitual activity (β = 0.05; 95% CI = 0.001–0.092; P = 0.046). - Pandemic relevance: Individuals with a more pronounced compensatory mechanism pre-pandemic were less likely to frequently feel lonely during the first COVID-19 lockdown (OR = 0.92; 95% CI = 0.85–0.99; P = 0.021). - Robustness under constraints: Benefits persisted under pandemic-like constraints—effects observed with only light physical activity (β = 0.04; 95% CI = 0–0.8; P = 0.040) and with physical activity performed at home (B = 0.08; 95% CI = 0.01–0.15; P = 0.032).

The study demonstrates, using intensive longitudinal e-diary and accelerometer data, that physical activity can compensate for the loss of affective well-being associated with lack of social contact in real life. This extends previous questionnaire-based or single-factor field studies by revealing a dynamic interplay between social contact and physical activity influencing affect in daily life. Approximately one hour of walking at about three miles per hour can offset the social-affective deficit linked to social isolation, and benefits persist at lower doses and when activity occurs at home, highlighting practical applicability even under restrictions like those during the COVID-19 pandemic. Importantly, compensation effects were stronger among individuals with neural risk markers (higher within-DMN connectivity), smaller social networks, and heightened loneliness during the pandemic. These findings identify likely responders and suggest an accessible strategy to mitigate negative consequences of social isolation, informing preventive recommendations for affective dysfunction in the post-pandemic context.

Physical activity is reproducibly linked to better affective well-being among people lacking social contact in daily life, with especially strong benefits in individuals at neural and psychological risk for affective disorders. The results suggest an effective, accessible approach to mitigate the adverse effects of social isolation and loneliness, with potential to improve public health and enrich preventive recommendations for managing social isolation in the post-pandemic world.

Affective valence was assessed via an established two-item mood scale suited for everyday-life research, providing insight into mood changes but not specific emotions (e.g., anxiety, anger); future work should examine emotion-specific effects. The observational, real-life design affords high ecological validity but precludes causal inference; associations and temporal order were observed, yet unmeasured third variables cannot be excluded. Future studies should test causality with experimental manipulations such as just-in-time adaptive interventions embedded in daily life.