Nonlinear effects at the electrode-tissue interface of deep brain stimulation electrodes

Abstract

Objective: The electrode-tissue interface provides the critical path for charge transfer in neurostimulation therapies and exhibits well-established nonlinear properties at high applied currents or voltages. These nonlinear properties may influence efficacy and safety of applied stimulation but are typically neglected in computational models. In this study, nonlinear behaviour of the electrode-tissue interface was incorporated in a computational model of deep brain stimulation (DBS) to simulate the impact on neural activation and safety. 
Approach: Nonlinear electrode-tissue interface properties were incorporated in a finite element model of DBS electrodes in vitro and in vivo, in the rat subthalamic nucleus, using an iterative approach. The transition point from linear to non-linear behaviour was determined for voltage and current-controlled stimulation. Predicted levels of neural activation during DBS were examined and the region of linear operation was compared with the Shannon safety limit. 
Main results: A clear transition of the electrode-tissue interface to the nonlinear region was observed for both current and voltage-controlled stimulation. The transition occurred at lower values of the activation overpotential for simulated in vivo than in vitro conditions (91mV and 165mV current-control; 110mV and 275mV voltage-control), corresponding to applied currents of 30µA and 45µA, voltage of 330mV. Incorporation of nonlinear properties resulted in activation of a higher proportion of neurons under voltage- but not current-controlled stimulation. Under current-controlled stimulation, the transition to Faradaic charge transfer occurred at stimulation amplitudes as low as 30µA, which corresponds to a charge density of 2.29µC/cm2 and charge of 1.8nC, well below the Shannon safety limit.
Significance: The results indicate that DBS electrodes may operate within the nonlinear region at clinically-relevant stimulation amplitudes. This affects the extent of neural activation under voltage-controlled stimulation and the transfer to Faradaic charge transfer for both voltage- and current-controlled stimulation with important implications for targeting of neural populations and safety.