STN-GPe DBS

This package implements an STN-GPe loop subject to STN stimulation, where amplitude was modulated by two control policies.

This package implements a computational model of subthalamopallidal network and investigates STN beta power under high-frequency stimulation. I am a final-year undergraduate, and this is my Capstone project focused on the dynamics of basal ganglia (BG) under parkinsonian condition, particularly the STN beta, and how deep brain stimulation (DBS) can restore normal BG dynamics.

The network consists of conductance-based STN and GPe neuron models connected by random synaptic projections. The function neuron_*.m, synapse_*.m, and network_*.m implement differential equations of corresponding neurons, synapses, and the complete network, receiving relevant variables as input and their time derivatives as output. The network equations are fed into RK3.m, which applies third-order Runge-Kutta method to solve the model, implemented in the solve_network.m script.

Based on this STN-GPe circuit model, high-frequency square pulse sequence was applied to each STN neuron. The solve_network_DBSparams.m implements open-loop, continuous pulses for different fixed amplitude values. The solve_network_PD.m and solve_network_thrs.m implement closed-loop modulation of pulse amplitude, PD control and a threshold-based modulation policy, respectively.

This project is my priliminary exploration in the field of computational modeling of DBS. Future studies will be more directed to research questions of significance, extending the model to downstream nucleus and studying the stimulation effects on network model with varying parameters.