Generalizable prediction of hand motor behaviour from spontaneous brain connectivity

Dexterous hand motor behavior emerges from coordinated interactions within a distributed brain network. While task-related neural dynamics have been investigated, recent fMRI studies showed that also spontaneous - i.e. non-task related - brain connectivity can predict task-specific performance. Still, it remains unclear whether spontaneous functional connectivity reflects also the encoding of general aspects of hand motor control. Here, we applied connectome-based predictive modelling (CPM) to resting-state functional connectivity (rs-FC) from the Human Connectome Project (HCP) to predict performance in hand motor tasks. We identified a "core" hand motor network whose intrinsic connectivity predicted not only task-specific measures (dexterity and strength) but generalised its prediction across different effectors and tasks. This "core" model also generalized its predictions to an independent dataset (external validation), including different behavioral measures and rs-fMRI data. In addition, transcranial magnetic stimulation (TMS) over inferior parietal cortex selectively impacted the core model's predictive power in a time-dependent manner, consistent with the known neurophysiological effects of the stimulation protocol. Together, these findings demonstrate that spontaneous brain activity encodes behaviorally relevant information about hand motor control, spanning both low-level features and higher-order representations. By linking spontaneous brain activity to behavioural motor outcomes, our findings pave the way for better understanding how spontaneous connectivity alterations might underlie motor dysfunction in neurological disorders.