Quantifying Age-Related Changes in Brain and Behavior: A Longitudinal versus Cross-Sectional Approach.

Abstract

Cross-sectional versus longitudinal comparisons of age-related change have often revealed differing results. In the current study, we used within-subject task-based fMRI to investigate changes in voxel-based activations and behavioral performance across the life span in the Reference Ability Neural Network cohort, at both baseline and 5 year follow-up. We analyzed fMRI data from between 127 and 159 participants (20–80 years) on a battery of tests relating to each of four cognitive reference abilities. We applied a Gaussian age kernel to capture continuous change across the life span using a 5 year sliding window centered on each age in our participant sample, with a subsequent division into young, middle, and old age brackets. This method was applied separately to both cross-sectional approximations of change and real longitudinal changes adopting a comparative approach. We then focused on longitudinal measurements of neural change to identify regions expressing peak changes and fluctuations of sign change across our sample. Our results revealed several regions expressing divergence between cross-sectional and longitudinal measurements in each domain and age bracket; behavioral comparisons between measurements showed differences in change curves for all four domains, with processing speed displaying the steepest declines. In the longitudinal change measurement, we found lack of support for age-related frontal increases across analysis types, instead finding more posterior regions displaying peak increases in activation, particularly in the old age bracket. Our findings encourage greater focus on longitudinal measurements of age-related changes, which display appreciable differences from cross-sectional approximations.