International Journal of Molecular Sciences · Jul 15, 2026

Multimodal Signatures of Brain Aging: From Descriptive Analyses to Machine Learning-Based Integration

By Caligiuri C, Feroleto C, Morotti M et al.

Understanding age-related neurophysiological changes is crucial for identifying brain aging biomarkers and developing strategies against motor and cognitive decline. To explore aging-related patterns across multiple domains, this study assessed motor and cognitive performance, functional connectivity, and synaptic organization in Young (4 months), Adult (14 months), and Old (24 months) mice. Adult mice exhibited reduced locomotor activity (-50.3%) and forelimb force (-38.3%) compared to Young mice, while Old mice showed decline in spatial (-30.4%) and recognition memory (-40.3%). Golgi-Cox staining revealed region-specific changes in spine density, including an increase in motor cortex layer II/III pyramidal neurons in Old versus Adult mice and reductions in the medial prefrontal cortex and CA1 hippocampal region. Immunofluorescence analysis indicated age-related alterations in VGLUT and VGAT expression across brain regions. Local field potential recordings revealed no significant changes in functional connectivity across age groups. Integration of behavioral and electrophysiological features using machine learning for an exploratory yielded a classification accuracy of 0.798. Although this represented only a modest and non-significant improvement over behavioral features alone, the highest pairwise discrimination was observed between Adult and Old mice (AUC = 0.861). Overall, these findings provide a multilevel descriptive analysis of brain aging, highlighting distinct behavioral and structural alterations alongside preserved functional connectivity.

doi.org/10.3390/ijms27146297 Open access PubMed 42511639 Mechanisms