Prospects for using nonlinear characteristics of single neuron activity in motor control models

This article explores the prospects of using nonlinear characteristics of single-neuron activity in the subthalamic nucleus (STN) for describing motor control in patients with Parkinson’s disease (PD). The study analyzes the activity of 308 neurons recorded from 16 PD patients during externally triggered and self-initiated movements. In addition to traditional linear parameters (such as firing rate and interspike interval variation), nonlinear metrics-including entropy (Sample Entropy, Approximate Entropy) and mutual informationwere evaluated. The results demonstrate that nonlinear characteristics can reveal differences in neuronal activity patterns between movement types and patient groups (with druginduced and without dyskinesias). Using a random forest algorithm, a predictive model was developed that accurately (0.82) classifies movement type based on neuronal activity parameters, with entropy and kurtosis measures being the most informative features. These findings highlight the importance of nonlinear metrics for assessing motor function and personalizing neurostimulation strategies in PD.