Brain Structural Correlates of Trajectories to Cognitive Impairment in the Cardiovascular Health Study - Cognition Study (P4.091)

OBJECTIVES: To describe the patterns of gray matter atrophy as a function of mixed membership model trajectories to cognitive impairment. BACKGROUND: We have previously demonstrated that there are three distinct trajectories to cognitive impairment among elderly individuals, and that these basis trajectories can be combined to express the individual impairment development courses of the participants in the cardiovascular health study-cognition study. The purpose of this analysis was to identify areas of brain regional atrophy associated with closeness to two of the trajectories. METHODS: Spoiled gradient recall images from 530 CHS-CS participants were processed using the VBM8 pipeline to obtain voxel-level measures of brain volume. We then correlated individual subject membership trajectory weights with these voxel-level brain volumes. We estimated model parameters with SPM8 for 100 multiply-imputed datasets (to account for the uncertainty in memberships weights), manually performed the post-hoc contrasts, and pooled the results, accounting for between- and within-imputation variability. This process generated an F-statistic and degrees of freedom at every voxel, from which we calculated R2 to create statistical parametric maps. RESULTS: The patterns of atrophy associated with the two trajectories were complementary. Membership to the trajectory associated with peak dementia onset in the 80s included posterior and medial hippocampus, as well as retro spleen Neo cortex. By contrast, membership to the trajectory with peak dementia onset in the late 70s was linked to an area of the hippocampus more anterior and lateral, and with a portion of the posterior cingulate superior and slightly lateral to the retro spleen neocortex. CONCLUSIONS: Trajectories to cognitive impairment in aging are the result, in part, of atrophy in cortical and subcortical regions linked with pathological aging. These data suggest the possibility of predicting cognitive morbidity based on patterns of CNS atrophy. Study Supported by: AG023629, AG15928, AG20098, AG027002, AG05133, and AG027058. Disclosure: Dr. Becker has nothing to disclose. Dr. Popov has nothing to disclose. Dr. Lecci has nothing to disclose. Dr. Gach has nothing to disclose. Dr. Junker has nothing to disclose. Dr. Kuller has nothing to disclose. Dr. Lopez has received personal compensation for activities with Merz Pharma as a consultant.