A Neuro-Cognitive Decoupling Framework for Investigating Resilience and Vulnerability in Aging Egyptian Fruit Bats

Abstract: Understanding the wide variability in cognitive aging, where some individuals maintain function despite age-related brain changes while others experience disproportionate decline, remains a critical challenge. This study introduces a novel neuro-cognitive decoupling framework designed to identify individual differences in aging trajectories and pinpoint associated brain regions in the long-lived Egyptian fruit bat (\textit{Rousettus aegyptiacus}). We established a comprehensive pipeline to integrate demographic data, brain Mean Diffusivity (MD) from Diffusion Tensor Imaging (global and 24 ROIs), and cognitive performance metrics from a three-phase spatial memory task in a cohort of 33 bats (epigenetic age 6.62-13.84 years). Our core methodology involved first establishing age-expected normative patterns for both brain MD and cognitive performance using linear regression models that included epigenetic age, sex, and origin colony. We then quantified individual-level 'decoupling indices' as residuals (observed minus predicted values), representing deviations from these norms, and modeled the relationships between brain MD residuals and cognitive residuals. While a critical limitation in the behavioral data extraction necessitated the use of synthetic behavioral data for the final analysis, the neuroimaging pipeline successfully extracted robust global and regional MD values. This proof-of-concept successfully demonstrated the framework's capacity to identify significant associations between brain MD residuals and (synthetic) cognitive residuals, illustrating its potential to uncover specific brain regions whose microstructural integrity disproportionately influences cognitive outcomes independent of chronological age. This residual-based approach offers a powerful, nuanced tool for unraveling mechanisms of cognitive resilience and vulnerability, paving the way for future biological insights once real behavioral data are integrated.

Subjects:	q-bio.NC; q-bio.QM
Cite as:	PX:2508.00045