Cochlear health alters the polarity effect and spike-initiation sites in guinea pigs

While cochlear implants (CIs) historically use cathodic-leading pulses for stimulation, studies in humans found that anodic-leading pulses are perceived louder than cathodic-leading ones. Modeling studies proposed that cathodic pulses excite the spiral ganglion neurons (SGNs) more peripherally than anodic pulses. Thus, the anodic-benefit in human CI listeners is thought to reflect degenerated peripheral processes. We used an animal model to test the contributions of peripheral dendrites and central axons of SGNs to polarity-effectiveness in CI stimulation. We mechanically lesioned the SGN (∼400 µm diameter; n = 18 cochleae) and introduced a 9-day degeneration time (n = 13 cochleae) to mimic human SGN degeneration. These lesions were compared to 20 control ears. We stimulated via a guinea-pig adjusted CI with symmetric, biphasic pulses (monopolar mode) of alternating leading-phase polarity (50µs/phase). Electrically-evoked compound action potential recordings to anodic- and cathodic-leading pulses were separated in the analysis to calculate the polarity effect. We confirmed the cathodic-benefit for cochleae with healthy SGN (lower threshold, larger amplitudes, dynamic ranges, and steeper slopes). Longer latencies (50-70µs) to cathodic than anodic monophasic and biphasic pulses confirmed the proposed peripheral (cathodic) and central (anodic) spike-initiation sites. The cathodic benefit persisted after acute lesioning, which prolonged latencies for anodic- but not for cathodic-leading pulses - consistent with remaining structures being excited by both polarities. After chronic degeneration, the threshold showed polarity-specific changes, leading to an anodic-benefit. The observed decline in cathodic-effectiveness with reduced neural health confirmed theoretical considerations for human CI users with stimulus-polarity and degeneration-type dependent changes in spike-initiation site.