Longitudinal Associations between the Neighborhood Built Environment and Cognition in US Older Adults: The Multi-Ethnic Study of Atherosclerosis.

Few studies have examined associations between neighborhood built environments (BE) and longitudinally measured cognition. We examined whether four BE characteristics were associated with six-year change in global cognition and processing speed. We obtained data on 1816 participants without dementia from the Multi-Ethnic Study of Atherosclerosis. BE measures included social destination density, walking destination density, proportion of land dedicated to retail, and network ratio (street connectivity). Global cognition was measured with the Cognitive Abilities Screening Instrument (CASI) and processing speed with the Digit Symbol Coding test (DSC). Multivariable random intercept logistic models tested associations between neighborhood BE at 2010-2012 and maintained/improved cognition (versus decline) from 2010-2018, and mediation by minutes of physical activity (PA)/week. The sample was an average of 67 years old (standard deviation = 8.2) (first cognitive measurement) and racially/ethnically diverse (29% African American, 11% Chinese, 17% Hispanic, 44% White). Compared to individuals with no walking destinations in the 1-mile surrounding their residence, those with 716 walking destinations (maximum observed) were 1.24 times more likely to have maintain/improved DSC score (Odds ratio: 1.24; 95% confidence interval: 1.03-1.45). No other associations were observed between BE and cognition, and PA minutes/week did not mediate the association between walking destination density and DSC change. This study provides limited evidence for an association between greater neighborhood walking destinations and maintained/improved processing speed in older age and no evidence for associations between the other BE characteristics and cognition. Future studies with finer grained BE and cognitive measures and longer-term follow up may be required.