In cochlear implant (CI) recipients, positioning the electrode array close to the cochlea’s inner wall (the modiolus), housing the auditory nerve, increases frequency selectivity and thereby improves hearing outcomes. Ideally, electrode positioning would be assessed continuously during electrode array insertion such that the surgeon can take immediate corrective action. To do so, measures of electrical impedance have come into view, as these can be made using the CI’s electrodes and internal circuits. In this study, we introduced a protocol for “three-point impedances” in which bipolar impedances are recorded in response to monopolar stimulation at a neighboring electrode. We assessed the usability of three-point impedances and two existing CI impedance measurement methods (monopolar and four-point impedances) for predicting electrode positioning during CI insertion.
Impedances were recorded during stepwise CI electrode array insertions in cadaveric human temporal bones. The positioning of the electrodes with respect to the modiolus was assessed at each step using cone beam computed tomography. Three-point and four-point impedances strongly correlated with electrode-modiolar distance. In contrast, monopolar impedances were only minimally affected by changes in electrode positioning with respect to the modiolus. An overall linear regression model specificity of 62% was achieved when incorporating all impedance parameters. This specificity could be increased beyond 73% when prior expectations of electrode positioning were incorporated into the model. Simulations using a lumped-element model of an implanted CI confirmed the experimental results. A comparison of experimental and clinical impedance data suggested that the experimental setup reasonably represents a human CI recipient in terms of model conductivities. Overall, this work shows that three-point and four-point impedances could be used to predict electrode positioning in real-time during CI insertion.