Social Networks and MRI measures of Brain Injury in Racially and Ethnically Diverse Populations

Abstract

AbstractBackgroundSocial networks may influence late‐life cognition directly via affecting brain structures or indirectly through enhancing cognitive resilience – i.e., the ability to maintain cognitive function in the presence of neurodegeneration. This study assesses the associations of social networks with markers of neurodegeneration, and determines whether social networks modify the associations between these markers and cognitive function.MethodWe pooled data from three harmonized cohorts of Kaiser Permanente Northern California members: the Kaiser Healthy Aging and Diverse Life Experiences (KHANDLE)study, the Study of Healthy Aging in African Americans (STAR), and the LifeAfter90 (LA90) Study. Social network sizes were calculated by summing the number of living children, spouse, and close relatives and friends. Neuroimaging markers included total brain volume (TBV), hippocampal volume (HCV), and white matter hyperintensity volume (WMH). Volumetric measures were normalized using a residual approach, which regressed intracranial volume (ICV) on volumetric measures to predict the ICV‐adjusted volume. Executive function and verbal episodic memory were measured by the Spanish and English Neuropsychological Assessment Scales and were z‐standardized. Linear regression models estimated the associations between social network size and each neuroimaging marker, adjusting for age, sex, race and ethnicity, education, cohort, and Activities of Daily Living (ADLs). We included an interaction term between social networks and neuroimaging markers to evaluate whether the relationships between imaging markers and cognitive function differed by social networks.ResultThe pooled sample’s average age was 76 (SD 11) years (59% female; 47% Black) (Table 1). In the fully adjusted models, the mean differences in injury markers did not differ by social network size (Figure 1). However, the associations of HCV and WMH with executive function (βsocial network*WMH = 0.09, 95% CI 0.01,0.18; βsocial network*HCV = ‐0.08, 95% CI ‐0.15, ‐0.01) were stronger among participants with smaller social networks (Figure 2).ConclusionIn this cross‐sectional study of racially and ethnically diverse older adults, larger social networks attenuated the associations between neurodegeneration markers and executive function. These findings suggest that larger social networks may promote cognitive resilience.