Abstract
BACKGROUND: Aging involves heterogeneous brain grey matter (GM) loss patterns that may overlap with dementia-related changes. We evaluated cognitively unimpaired older adults to identify specific GM patterns, their clinical and cognitive profiles, and longitudinal trajectories. METHODS: We analyzed 746 participants from the Gothenburg H70 Study using random forest clustering based on MRI measures of cortical thickness and subcortical volume across 41 regions. We examined associations with clinical and MRI variables, APOE status, and CSF Alzheimer biomarkers (n = 286), and assessed 5-year cognitive and brain trajectories using regression models. RESULTS: Five clusters emerged, mainly differing in frontoparietal regions. Compared to Cluster 1 (reference), Cluster 2 showed diffuse GM loss, higher odds of diabetes (OR = 2.54, 95%CI [1.27-5.06]) and risk alcohol consumption (OR = 1.83, 95%CI [1.13-2.97]), poorer episodic memory (β=-0.190, p = 0.014) and visuospatial abilities (β =-0.209, p = 0.044), and greater longitudinal decline in MMSE (β(slope) = - 0.448, p = 0.035) and increase in white matter hyperintensity volume (β(slope) = 1.843, p = 0.004). Clusters 3 showed thicker GM and lower BMI (OR = 0.57, 95%CI [0.35-0.94]). Cluster 4 had preserved GM, lower smoking habits (OR = 0.62, 95%CI [0.40-0.95]), triglycerides levels (OR = 0.55, 95%CI [0.32-0.95]) and depression (OR = 0.17, 95%CI [0.05-0.56]), higher education (OR = 2.52, 95%CI [1.08-5.87]) and better cognition in multiple domains. Cluster 5 had mixed GM pattern and higher odds of heart disease (OR = 3.44, 95%CI [1.48-8.01]). CONCLUSIONS: Cardiovascular and psychosocial factors are key determinants of GM integrity and cognition. Targeting these risks factors may preserve brain health in late life.