Characterizing brain aging patterns and their relationship to cognitive decline is crucial, especially in the eighth decade of life when dementia risk accelerates. Previous research often relies on cross-sectional data, which confounds age effects with lifelong patterns. This study utilizes longitudinal data to overcome this limitation and investigate individual differences in cortical aging patterns in a large sample of healthy older adults. The study focuses on uncovering the dimensionality of cortical atrophy, its relationship to cognitive decline, and how this contrasts with cross-sectional correlations. Understanding these patterns is essential for developing effective preventative and interventional strategies for cognitive decline.

Existing research indicates that cortical aging is not uniform, with some regions showing greater age-related atrophy than others. Univariate analyses, focusing on single brain regions, are insufficient to capture the complexities of this process. Cross-sectional studies have identified clusters of regions with shared morphometric characteristics, but these cannot fully reflect the dynamics of within-individual change over time. The authors cite a previous study on multivariate longitudinal changes in cortical structure in an age-heterogeneous sample with mild cognitive impairment, identifying five atrophic patterns. However, the question of whether these patterns replicate in healthy older adults remains unanswered.

Participants (N=1091 initially) were drawn from the Lothian Birth Cohort 1936 and followed up at approximately ages 73, 76, and 79. Structural MRI data were acquired at waves 2-4 (1376 scans total). Cortical grey matter was parcellated into 34 regions per hemisphere using FreeSurfer. Longitudinal growth curves for each region were estimated using structural equation modeling (SEM). Exploratory factor analysis (EFA), using a Schmid-Leiman transformation, was then performed on the estimated latent correlation matrices of both intercepts and slopes. The optimal factor solution (number of factors) was selected based on eigenvalue magnitudes, loading patterns, and root mean square of residuals. To check for hemisphere consistency, the analysis was repeated for the right hemisphere. Finally, confirmatory factor analysis (CFA) was used to test the identified factor structure and investigate its relationship with cognitive decline, using Factor of Curves growth curve SEMs for general cognitive ability and specific domains (visuospatial, processing speed, memory). APOE e4 carrier status was also considered.

Longitudinal correlations of cortical volumetric changes were substantially higher (average r = 0.805) than cross-sectional correlations (average r = 0.350). EFA revealed a three-factor structure: a general factor of cortex-wide atrophy (explaining 66% of variance), and two more specific factors representing fronto-temporal and occipito-parietal atrophy (additional 20% variance). The general atrophy factor was strongly associated with declines in general cognitive ability (r = 0.431, p < 0.001) and in visuospatial ability (r = 0.415, p = 0.002), processing speed (r = 0.383, p < 0.001), and memory (r = 0.372, p < 0.001). Sensitivity analyses excluding participants with dementia or stroke diagnoses showed similar results. Overall cortical volume showed a significant decline of 0.87% per annum.

The findings highlight the importance of longitudinal studies for understanding brain aging. The three-factor model of cortical atrophy provides a more nuanced view than cross-sectional analyses, capturing the coordinated nature of age-related changes. The strong association between the general atrophy factor and cognitive decline across domains suggests a shared underlying neurobiological mechanism. The identification of distinct regional patterns suggests that different aspects of cognition may be differentially vulnerable to age-related brain changes. This work provides valuable insights for understanding the neurobiology of cognitive aging and informs strategies for prevention and intervention.

This study demonstrates that cortical aging in the eighth decade of life is characterized by three major dimensions of atrophy. The most prominent dimension is a general factor of cortex-wide atrophy strongly linked to cognitive decline. Future research could explore the genetic and environmental factors that contribute to these patterns, potentially leading to improved diagnostic and therapeutic approaches.

The study is limited to a specific age range and birth cohort, which may limit generalizability. While participants were initially free of dementia, the study could not fully account for preclinical stages of neurodegenerative disease. The cross-sectional nature of the cognitive assessments within each wave might not fully capture the precise timing and nature of cognitive decline. Future research could address these limitations by broadening the age range and considering more frequent cognitive assessments.