Mechanisms mediating dynamic changes in neural responses during deep brain stimulation.

BACKGROUND: Deep brain stimulation (DBS) of the subthalamic nucleus (STN) to treat Parkinson's disease (PD) generates local evoked potentials (DLEPs). DLEPs reflect neural activation by DBS, but the mechanisms underlying the dynamic changes in DLEPs during continuous DBS are not understood, and such knowledge is critical to using these signals for DBS programming and closed-loop control. METHODS: We first systematically incorporated short-term synaptic depletion into 4 different synapse types in a biophysically realistic computational model of DLEPs. We then recorded DLEPs in naïve and a virally transfected rat model. Lastly, we conducted a retrospective analysis correlating DLEP dynamics to motor symptom severity in 13 patients with PD. RESULTS: A bespoke biophysical model suggests that synaptic depression, through synaptic vesicle depletion of activated motor cortical synapses onto STN neurons, causes the dynamic changes in DLEP amplitude and latency during continuous DBS. Virally-driven overexpression of endophilin A1 or alpha-synuclein in rat motor cortex reduces the time constant of DLEP amplitude and latency, supporting a critical role of these synapses in mediating DLEP dynamics. Further, a correlation between the change in UPDRS-III scores with antiparkinsonian medications and the time constant of DLEP amplitude is observed in persons with PD, suggesting DLEP dynamics as a biomarker of PD progression. CONCLUSIONS: Collectively, these results suggest short-term synaptic depression of cortical synapses onto STN neurons mediate DLEP dynamics in STN DBS.