Inverse Neural-Kinematic Dynamics Revealed by RQA During Motor Skill Learning

Abstract

Motor skill acquisition involves dynamic co-adaptation of brain and behavior, yet the temporal coupling between neural and kinematic processes during learning remains unclear. Traditional analyses often treat the brain and behavior independently, potentially overlooking their interactive dynamics. In this study, we applied Recurrence Quantification Analysis (RQA), a nonlinear method for assessing temporal structure in complex systems, to both EEG and foot trajectory data collected during a seven-session motor learning task. Twelve healthy adults practiced drawing five shapes with their dominant foot on a digital tablet while EEG was recorded from sensorimotor cortex electrodes. We found a fundamental inverse relationship between neural and behavioral dynamics: as motor performance improved, EEG determinism in the left hemisphere increased by 2.7% (p<0.001), while kinematic determinism decreased by 18.4% (p<0.001), indicating more organized brain activity supporting more flexible movement patterns. EEG laminarity also increased by 1.5% (p<0.001), while kinematic laminarity decreased by 15.2% (p<0.001). Shape-specific analysis revealed that patterns B and M elicited the most pronounced neural-behavioral adaptations. These findings suggest that increasingly structured neural dynamics facilitate adaptive motor execution. Our dual-domain RQA framework offers a powerful tool to quantify learning efficiency and uncover the dynamic coupling between neural reorganization and behavioral adaptation.