Cognitive dysfunction is a pervasive symptom of multiple sclerosis (MS). Correlational evidence on the relationships between physical activity, sedentary behavior, and cognition has been mixed and limited to a few activity measures. The collinearity of accelerometry-based metrics has precluded an assessment of the full activity spectrum. Here, we aimed to examine the rich set of activity measures using analytic approaches suitable for collinear metrics. We investigated the combination of physical activity, sedentary, and clinicodemographic measures that explain the most variance in composite scores of working memory/processing speed, visual memory, and verbal memory.
We analyzed baseline accelerometry and neuropsychological data (n = 80) from a randomized controlled trial of pedometer tracking. Using partial least squares regression (PLSR), we built three models to predict latent scores on the three domains of cognition using 12 activity metrics, sex, education, and Expanded Disability Status Scale (EDSS) scores. Significance was assessed using linear regression models with model component scores as predictors and cognitive composites as outcomes.
The latent component was significant for working memory/processing speed but was not significant for visual memory and verbal memory after Bonferroni correction. Working memory/processing speed was positively associated with average kilocalories, moderate-to-vigorous physical activity (MVPA), steps, and sex (i.e., higher scores in males) and negatively related to duration of long sedentary bouts and EDSS.
These findings suggest that increasing overall energy expenditure through walking and MVPA, while decreasing prolonged sedentary time may positively benefit working memory/processing speed in people with MS.
This RCT #NCT03244696 was registered on Clinicaltrials.gov (https://www.
gov/ct2/show/NCT03244696).

Copyright © 2023 Elsevier B.V. All rights reserved.