Collaborative Research: CCSS: Brain Connectivity Analysis using Communication Theory: Towards Early Detection and Prediction of Cognitive Decline

The human brain is one of the most sophisticated communication networks in nature, and information exchange among distributed brain regions gives rise to cognition and behavior. Along with rapid advances in neuroimaging and electrophysiological recording technologies, brain researchers are increasingly relying on sophisticated computational tools to analyze large, complex datasets. At the same time, driven by breakthroughs in information theory and machine learning, the communications field has developed rich methodologies for system modeling, signal extraction, and information transmission analysis. Motivated by the convergence of communication theory and neuroscience, this interdisciplinary project aims to develop innovative techniques for brain connectivity analysis and early assessment of cognitive health. The resulting technologies will deepen our understanding of brain functions and information processing mechanisms and can be applied directly to the early detection and prediction of cognitive decline and Alzheimer’s disease. Furthermore, by integrating these advances into undergraduate and graduate education, outreach activities, and interdisciplinary training programs, this project will significantly contribute to the development of a highly skilled and diverse workforce in communications, artificial intelligence, and computational neuroscience. The goal of the project is to develop innovative methods for quantitative, information-theoretic modeling and characterization of dynamic brain connectivity and to explore their applications in identifying individuals at risk of cognitive decline. More specifically, (i) this project establishes the conditional equivalence between directed information and existing causality models and develops new methods for characterizing brain effective connectivity and directed information transfer using causalized convergent cross mapping, a state-space reconstruction approach rooted in dynamical systems theory; (ii) this project develops multiscale time–frequency–spatial frameworks for brain connectivity analysis to characterize dynamic neural interactions across multiple temporal and frequency resolutions and identify connectivity alterations associated with cognitive impairment; and (iii) this project develops adaptive discrimination models that move beyond conventional binary classification and enable continuous assessment across the spectrum from normal cognition to mild cognitive impairment, thereby facilitating the early detection and prediction of individuals at risk of cognitive decline before clinical symptoms become recognizable. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.