Computer Simulation of the Electrical Stimulation of the Human Vestibular System: Effects of the Reactive Component of Impedance on Voltage Waveform and Nerve Selectivity

Abstract

The vestibular system is responsible for our sense of balance and spatial orientation. Recent studies have shown the possibility of partially restoring the function of this system using vestibular implants. Electrical modeling is a valuable tool in assisting the development of these implants by analyzing stimulation effects. However, previous modeling approaches of the vestibular system assumed quasi-static conditions. In this work, an extended modeling approach is presented that considers the reactive component of impedance and the electrode-tissue interface and their effects are investigated in a 3D human vestibular computer model. The Fourier finite element method was employed considering the frequency-dependent electrical properties of the different tissues. The electrode-tissue interface was integrated by an instrumental electrode model. A neuron model of myelinated fibers was employed to predict the nerve responses to the electrical stimulus. Morphological changes of the predicted voltage waveforms considering the dielectric tissue properties were found compared to quasi-static simulations, particularly during monopolar electrode configuration. Introducing the polarization capacitance and the scar tissue around the electrode in combination with a power limitation leads to a considerable current reduction applied through the active electrode and, consequently, to reduced voltage amplitudes of the stimulus waveforms. The reactive component of impedance resulted in better selectivity for the excitation of target nerves compared to the quasi-static simulation at the expense of slightly increased stimulus current amplitudes. We conclude that tissue permittivity and effects of the electrode-tissue interface should be considered to improve the accuracy of the simulations.

Unpacking Galvanic Vestibular Stimulation using simulations and relating current flow to reported motions: Comparison across common and specialized electrode placements

“…Further, computational modeling efforts in vestibular stimulation in general, are sparse. For invasive vestibular implant, a nite element (FE) model of the human inner ear coupled to a neuron model has been used to investigate electrode locations and waveforms [25,55]. Induced EF via magnetic induction has been estimated in the context of non-invasive magnetic stimulation [26].…”