Directed information measures in neuroscience

En savoir plus sur le livre

Analysis of information transfer has gained traction in neuroscience, revealing a dynamic flow of information that operates alongside the brain's slowly evolving anatomical connectivity. Understanding this transfer is essential for grasping how flexible routing and processing contribute to higher cognitive functions. This work reviews recent advancements in concepts and tools for measuring information transfer, focusing on their application to neurophysiological recordings and interaction analysis. Authored by leading researchers, it explores the current state, future prospects, and challenges in efficiently assessing neuronal information transfer. Key topics include the theoretical quantification and practical estimation of information transfer, spatial and temporal transfer descriptions, multivariate directed measures, and information decomposition among various stimulus/response variables. It also examines the relationship between interventional and observational causality. Applications to neural datasets and references to open-source software underscore the practical relevance of these measures in experimental neuroscience. With cutting-edge mathematical developments, computational techniques, and real data applications, this work serves as a valuable resource for graduate students and researchers keen on understanding information transfer in complex systems.