Impedance measures for a better understanding of the electrical stimulation of the inner ear

Quentin Mesnildrey,Olivier Macherey,Philippe Herzog,Frédéric Venail

DOI: https://doi.org/10.1088/1741-2552/aaecff

Abstract:The performance of cochlear implant (CI) listeners is limited by several factors among which the lack of spatial selectivity of the electrical stimulation. Recently, many studies have explored the use of multipolar strategies where several electrodes are stimulated simultaneously to focus the electrical field in a restricted region of the cochlea. Objective: These strategies are based on several assumptions concerning the electrical properties of the inner ear that need validation. The first, often implicit, assumption is that the medium is purely resistive and that the current waveforms produced by several electrodes sum linearly. The second assumption relates to the estimation of the contribution of each electrode to the overall electrical field. These individual contributions are usually obtained by stimulating each electrode and measuring the resulting voltage with the other inactive electrodes (i.e. the impedance matrix). However, measuring the voltage on active electrodes (i.e. the diagonal of the matrix) is not straightforward because of the polarization of the electrode-fluid interface. In existing multipolar strategies, the diagonal terms of the matrix are therefore inferred using linear extrapolation from measurements made at neighboring electrodes. Approach: In experiment 1, several impedance measurements were carried out in vitro and in eight CI users using sinusoidal and pulsatile waveforms to test the resistivity and linearity hypotheses. In experiment 2, we used an equivalent electrical model including a constant phase element in order to isolate the polarization component of the contact impedance. Main results: In experiment 1, high-resolution voltage recordings (1.1 MHz sampling) showed the resistivity assumption to be valid at 46.4 kHz, the highest frequency tested. However, these measures also revealed the presence of parasitic capacitive effects at high frequency that could be deleterious to multipolar strategies. Experiment 2 showed that the electrical model provides a better account of the high-resolution impedance measurements than previous approaches in the CI field that used resistor-capacitance circuit models. Significance: These results validate the main hypotheses underlying the use of multipolar stimulation but also suggest possible modifications to their implementation, including the use of an impedance model and the modification of the electrical pulse waveform.