Cortical contribution to subthalamic activity during chronic electrical stimulation

Abstract

Objective: In Parkinson's disease (PD), the motor cortex appears to play an important role in the generation of abnormal activity in the motor loop network involving basal ganglia, thalamus and cortex. When activity in this network is disrupted, motor symptoms appear but are alleviated by deep brain stimulation (DBS). We use a computational model to examine cortical contribution during DBS. Exploring brain mechanims underlying the effect of DBS may help us propose less invasive but efficient (cortical) stimulation protocols. Method: We used a multi-scale, population based, mathematical model of the subthalamic nucleus (STN) -the main excitatory structure of the basal ganglia- and the external globus pallidus, which receive cortical and striatal input. The model simulates healthy (stable, low-amplitude) and pathological (5 Hz oscillatory synchronized) activity within these two nuclei. In the pathological state, constant or oscillatory (low-frequency) cortical inputs to the STN are simulated and the ability of DBS to suppress abnormal oscillations (bursts) is explored. Results: In the presence of 10 Hz cortical inputs to the STN, DBS does not suppress abnormal bursts present in the STN and correlated with tremor generation (Figure 1). On the contrary DBS suppresses abnormal bursts when cortical input is constant (not shown). Conclusion: Results support the view that a functional decoupling takes place between resonant cortical input and STN during DBS in PD. Indeed cortical input to the STN has a frequency similar to that of the STN, enhancing STN abnormal oscillatory behavior. This decoupling may originate from the descending cortical spikes colliding with the antidromic depolarization rising from the STN and induced by DBS. This mechanism could be responsible for clinical improvements. In principle such a functional decoupling may be achieved through cortical stimulation. This could be done in practice by switching from low- to high-pass the filter properties of cortical synapses, thus removing the low-frequency cortical contribution