Risk and protective factors associated with grey matter patterns in older adults

Abstract INTRODUCTION Early dementia detection in the general population is challenged by high cerebral grey matter (GM) heterogeneity preceding behavioural symptoms. Here, we identify distinct GM patterns and their associated factors in a population-based cohort to detect at-risk individuals. METHODS This cross-sectional study examined 746 dementia-free 70-years-old individuals from the Gothenburg H70 Study to identify GM patterns using random forest clustering on MRI measures and examine their associations with sociodemographic, cardiometabolic, cognitive, genetic, and biomarker characteristics. RESULTS Five GM clusters emerged, primarily differentiated by frontoparietal regions. Compared to Cluster 1 (reference), cortical thickness was greater in Clusters 3 and 4, while reduced in Cluster 2, and mixed in Cluster 5. Significant factors included education, cardiometabolic conditions, depression, neurodegeneration, small vessel disease, lipidic and inflammatory alterations. Interestingly, GM patterns reflected cognitive performance. DISCUSSION Cerebrovascular health and related processes could be crucial for GM heterogeneity in late life, with implications for preventive strategies.