Association Between Frequency of Social Contact and Brain Atrophy in Community-Dwelling Older People Without Dementia: The JPSC-AD Study

Background and ObjectivesEpidemiologic evidence has shown that social isolation, a low frequency of social contact with others, is associated with the risk of dementia and late-life depressive symptoms. Therefore, we hypothesized that low frequency of social contact may be involved in brain atrophy, and depressive symptoms may play some role in this relationship. We aimed to evaluate the association between low frequency of social contact and the volumes of various brain regions and to assess the extent to which depressive symptoms mediate these relationships from a large population-based multisite cohort study.MethodsDementia-free community-dwelling Japanese aged 65 years or older underwent brain MRI scans and a comprehensive health examination. Frequency of contact with noncohabiting relatives and friends was determined by asking a single question with 4 categories: everyday, several times a week, several times a month, and seldom. Total and regional brain volumes, intracranial volume (ICV), and white matter lesion volume were estimated using FreeSurfer software. The associations between frequency of social contact and brain volumes per ICV were examined using analyses of covariance. Mediation analyses were conducted to calculate the proportion of the associations explained by depressive symptoms.ResultsWe included 8,896 participants. The multivariable-adjusted mean of the total brain volume in the group with the lowest frequency of social contact was significantly lower compared with that in the group with the highest frequency of social contact (67.3% vs 67.8%), with a significant increasing trend across the groups (p value for trend <0.001). The white matter lesion volume increased significantly with lower frequency of social contact (0.30% in the lowest frequency group vs 0.26% in the highest frequency group, p value for trend <0.001). Lower frequency of social contact was associated with smaller volumes in the temporal lobe, occipital lobe, cingulum, hippocampus, and amygdala (all q values of false discovery rate correction <0.05). The relationships seemed to be partly mediated by depressive symptoms, which accounted for 15%-29% of the observed associations.DiscussionLower frequency of social contact was associated with decreased total and cognitive function-related regional brain volumes. In addition, depressive symptoms partially explained the association in community-dwelling older people without dementia in Japan.