Motor Expertise Combines Lower Task Engagement with Enhanced Error Sensitivity

Skilled motor behaviour relies on complex predictive control and error processing. Despite its indispensable position in our everyday lives, neurobiological mechanisms of motor skill acquisition remain poorly understood. The presented study evaluates the functional dynamics and microstructural substrate of expertise-related reorganisation in predictive timing circuits using multimodal functional and quantitative magnetic resonance imaging (MRI) data. Forty-one baseball/softball players, representing a natural model of optimised motor timing programs, were compared against forty-six non-athletic controls using a structural connectome-informed hierarchical Bayesian framework jointly modelling task-based functional MRI dynamics of feedforward predictive control and error-driven modulation. Expertise-related microstructural differences were then characterised using neurite orientation dispersion and density imaging and adiabatic T₁ρ mapping. In non-athletic controls, predictive control was found to engage a distributed cortico-subcortical sensorimotor, visuospatial, attentional and frontoparietal loops, while higher-order associative systems were relatively down-regulated. However, quantitative deviation from the task optimum was associated with lower response across this predictive network. Expertise was characterised by higher behavioural performance but also by attenuated prediction-related functional engagement alongside enhanced error-related modulation. Furthermore, neurite density and orientation dispersion indices were higher in experts across distributed cortical networks, while the cerebellum exhibited a divergent profile characterised by higher neurite density, lower orientation dispersion and faster T₁ρ relaxation. These findings indicate that expertise involves lower task engagement within an established architecture despite higher behavioural performance, together with enhanced error sensitivity and distributed microstructural differences. This pattern is consistent with more efficient neural dynamics complemented with underlying tissue alterations.