Monitoring functionality and positioning of cochlear implants by impedance spectroscopy and in silico modelling

Cochlear implants allow compensating severe-to-profound hearing loss by directly stimulating the auditory nerve electrically. For optimal performance and maximum patient benefit, accurate placement of the cochlear implant electrode array inside the cochlea is crucial. It can be controlled using imaging methods, such as computed tomography. To minimize radiation exposure, impedance measurements are increasingly used to monitor position and functionality of the cochlear implant electrode. Stimulation electrode impedance can be affected by cochlear implant electrode position, but also biological responses of the cochlea like cell growth on the stimulation electrodes or a change of perilymph composition. To better understand and characterize the impedance behavior of cochlear implants, aiming to provide more precise, localized information about the implant’s environment inside the cochlea, this study explores the use of impedance spectroscopy across a wide frequency range between two stimulation electrodes. A finite element method simulation model of a cochlear implant electrode placed in a simplified straight epoxy cochlea phantom was developed. This simulation model could accurately describe the complex impedance measured between two stimulation electrodes over a frequency range from 20 Hz to 10 MHz (mean relative error of 11.26 %) and especially at a fixed frequency of 100 kHz (mean relative error of 4.20 %). Based on the validated model, an additional cochlea phantom shaped as a conical helix was simulated to investigate the effect of cochlear implant electrode curvature. This snail-shaped model showed that with increasing curvature of the cochlear implant inside the cochlea the absolute value of impedance between two stimulation electrodes decreases. While existing three-dimensional simulation models mainly focus on current distribution or patient-specific cochlea modeling, this study focuses on accurately modeling the cochlear implant, including its lead wires and the forming bilayer on the stimulation electrodes, to correctly describe impedance spectroscopy between two stimulation electrodes. In the long term, this simulation model can serve as a basis for the development of a holistic prediction model which allows for distinguishing different causes for implant failure using impedance spectroscopy.