Expression Variability of Cognition-Associated Genes Across Neural Cell Types and Species During Aging, with Implications for Predicting Alzheimer's Disease in Humans

As the global population ages, understanding the molecular foundations of cognitive decline and neurodegenerative disorders like Alzheimer's disease (AD) has become a priority. Both aging and AD are associated with reduced neurogenesis and altered neural plasticity, driven by changes in gene expression within neural stem cells (NSCs) and other neural cell types. While previous studies have analyzed cognition-associated gene expression changes in aging neural cells, few have explored these alterations across species or within NSCs. This study investigates age-associated gene expression changes in cognition-related genes in NSCs and other neural cell types across human and mouse datasets, aiming to identify conserved or species-specific patterns. Using gene expression data from human AD patients and controls, as well as from mouse samples, we identified key genes linked to cognitive function and their correlation with cognitive outcomes. To assess the predictive potential of these genes, we developed 24 classification models, with the Decision Tree (DT) and Sequential Neural Network (SNN) with early stopping achieving the best performance. The DT model attained a ROC-AUC score of 0.95897, while the SNN achieved a test accuracy of 90.34 %. These findings highlight critical genetic markers of cognitive decline and underscore their potential for early AD detection and therapeutic targeting.