Longitudinal Clustering Identifies Different Rates of Cortical Aging in Healthy Older Adults.

Global trajectories of brain aging are well-characterized thanks to many large MRI datasets in the general population. However, regional patterns that may reflect individual differences in brain aging remain less well understood. In this study, we aim to identify differences in brain aging in a longitudinal study of healthy older individuals using a clustering approach and test predictors of cluster membership. Cortical thickness (CTh) is an important biomarker of cortical macrostructure, relevant both in normal aging processes and in many pathologies. CTh was assessed in 176 healthy individuals aged 68-85 years at two time points (mean time interval: 5 ± 1 years) using T1-weighted magnetic resonance imaging (MRI). Regional cortical atrophy was identified based on the annualized percent change (APC) of CTh. Different aging clusters were then classified using hierarchical cluster analysis based on APC values. Two clusters emerged: one with minimal atrophy and the other with more pronounced atrophy of CTh, particularly in left prefrontal and central regions. There were no significant group differences in age or sex between the two clusters. The cluster with more pronounced atrophy was significantly associated with education in the univariate analysis and at borderline significance in the adjusted analysis and showed a non-significant trend of association with cognitive decline. Our results indicate different rates of cortical aging within the context of healthy aging processes.