Physical activity, brain tissue microstructure, and cognition in older adults

OBJECTIVE: To test whether postmortem MRI captures brain tissue characteristics that mediate the association between physical activity and cognition in older adults. METHODS: Participants (N = 318) were older adults from the Rush Memory and Aging Project who wore a device to quantify physical activity and also underwent detailed cognitive and motor testing. Following death, cerebral hemispheres underwent MRI to quantify the transverse relaxation rate R(2), a metric related to tissue microstructure. For analyses, we reduced the dimensionality of the R(2) maps from approximately 500,000 voxels to 30 components using spatial independent component analysis (ICA). Via path analysis, we examined whether these R(2) components attenuated the association between physical activity and cognition, controlling for motor abilities and indices of common brain pathologies. RESULTS: Two of the 30 R(2) components were associated with both total daily physical activity and global cognition assessed proximate to death. We visualized these components by highlighting the clusters of voxels whose R(2) values contributed most strongly to each. One of these spatial signatures spanned periventricular white matter and hippocampus, while the other encompassed white matter of the occipital lobe. These two R(2) components partially mediated the association between physical activity and cognition, accounting for 12.7% of the relationship (p = .01). This mediation remained evident after controlling for motor abilities and neurodegenerative and vascular brain pathologies. CONCLUSION: The association between physically activity and cognition in older adults is partially accounted for by MRI-based signatures of brain tissue microstructure. Further studies are needed to elucidate the molecular mechanisms underlying this pathway.