A Topological Fingerprint Encodes Motor Skill at Rest

In this study, we investigated whether the architecture of brain interactions at rest maintains a representation of individual behavioral skills. Specifically, we aimed to identify a minimal set of topological features that capture an electrophysiological mechanism underlying the encoding of a motor skill. We tested whether brain network topology at rest could model individual performance in a motor task, such as manual dexterity, in 86 subjects of either sex from the Human Connectome Project. Using a machine learning procedure, we identified an optimal fingerprint that accurately modeled individual manual dexterity, encompassing four participation index-based connector hubs in the alpha frequency band, involving the parietal cortex. A vulnerability analysis, in which we simulated disconnections of the involved hubs, revealed that two of them were critical, resulting in a significant drop in predictive performance. We combined these features to propose a functional "refocusing" mechanism: hubs progressively prune connections external to their modules when dexterity increases, while maintaining an internal representation of dexterity performance. Such inhibition and maintenance are well aligned with the role of the alpha band reported in the literature. These findings suggest that the architecture of interactions at rest, by combining few topological features in the alpha band, encodes stable behavioral traits, such as motor skills.