The effect of deep brain stimulation on cortico-subcortical networks in Parkinson's disease patients with freezing of gait: Exhaustive exploration of a basic model

Current treatments of Parkinson's disease (PD) have limited efficacy in alleviating freezing of gait (FoG). In this context, concomitant deep brain stimulation (DBS) of the subthalamic nucleus (STN) and the substantia nigra pars reticulata (SNr) has been suggested as a potential therapeutic approach. However, the mechanisms underlying this approach are unknown. While the current rationale relies on network-based hypotheses of intensified disinhibition of brainstem locomotor areas to facilitate the release of gait motor programs, it is still unclear how simultaneous high-frequency DBS in two interconnected basal ganglia nuclei affects large-scale cortico-subcortical network activity. Here, we use a basic model of neural excitation, the susceptible-excited-refractory (SER) model, to compare effects of different stimulation modes of the network underlying FoG based on the mouse brain connectivity atlas. We develop a network-based computational framework to compare subcortical DBS targets through exhaustive analysis of the brain attractor dynamics in the healthy, PD, and DBS states. We show that combined STN+SNr DBS outperforms STN DBS in terms of the normalization of spike propagation flow in the FoG network. The framework aims to move toward a mechanistic understanding of the network effects of DBS and may be applicable to further perturbation-based therapies of brain disorders. Parkinson's disease patients with freezing of gait (FoG) may be treated by deep brain stimulation, which produces effects mediated by brain networks. Currently, the approach of combined DBS of the subthalamic nucleus and the substantia nigra pars reticulata is investigated for being particularly beneficial for patients with axial symptoms, but the exact mechanisms of this effect are unknown. Here, we present a network-based computational framework using a basic excitable model that enables us to simulate the complete activity patterns of the brain network involved in FoG. These simulations reveal network mechanisms underlying STN+SNr DBS and its efficacy in alleviating FoG. The proposed framework can capture the influence of the DBS target sites on cortico-subcortical networks and help to identify suitable stimulation targets.