S. Tsugaya, A. Nakamura, T. Nomura
Electroencephalographic (EEG) studies of human quiet stance demonstrate beta-band event-related desynchronization (beta-ERD) during the micro-fall phase of postural sway, followed by event-related synchronization (beta-ERS; post-movement beta rebound) during the micro-recovery phase. These modulations correlate with intermittent ankle muscle inactivation that exploits the stable manifolds of an unstable upright equilibrium; however, the underlying neurocircuit mechanisms remain elusive. Here, we investigated this rhythmogenesis using an embodied spiking neural network model of the cortico-basal ganglia-thalamic (CBGT) circuitry integrated with a physical inverted pendulum. In this closed-loop system, continuous sensory feedback is integrated into the striatum, while the motor cortex executes decisions via drift-diffusion-like population competition, where the decision time (DT) represents the intermittent control-off period. We demonstrate that simulated cortical LFPs exhibit characteristic sway-phase-locked beta-ERD and beta-ERS exclusively when corticostriatal synaptic weights are functionally balanced to implement intermittent control. Conversely, forced-choice continuous control that ceaselessly generates feedback torque fails to replicate these modulations, sustaining flat network states devoid of control-off periods (DT). Structural dissections reveal that while sensory drive remains continuous, sway-phase-locked beta modulations are an emergent property generated fundamentally by bidirectional thalamocortical loops and the intrinsic dynamics of the GPe-STN pacemaker circuit. Our findings suggest that CBGT-mediated, sway-phase-locked beta activity serves as a hallmark of healthy intermittent motor selection. This computational framework provides a crucial bridge linking pathological alterations in basal ganglia dynamics and the loss of behavioral intermittency to the postural impairments observed in clinical populations such as Parkinson's disease.