The Phenomenon of Functional Hair Cells Disconnected from Neurons in Cochlear Implant Subjects

Background: Electrocochleography (ECochG) can be used to assess the residual cochlear function in cochlear implant (CI) subjects. The summed magnitude of the ECochG responses to tones of different frequencies shows a high degree of correlation with word score outcomes in adults (r=0.67) and older children (r=0.60). Our hypothesis is that these correlations are due to hair cell rather than neural activity, because the neural activity should reflect the audiometric results which are not predictive of outcomes with implants, while knowledge of hair cell activity could be new information. The phenomenon would be indicative of ‘cochlear synaptopathy’, which has been shown in animals but not previously in humans.

Objectives: To describe the functional substrate of the cochlea in CI subjects

Methods: ECochG to a series of tone frequencies was collected intraoperatively from the round window of adult subjects (n=166). Recordings were analyzed to separate hair cell from neural responses in order to identify instances of hair cell activity with no corresponding neural activity.

Results: In 90/373 recordings (24%), where neural activity should be observed if present (i.e., with magnitudes >0.5 µV and extending to nearly 100 µV), there was no evidence of neural activity. The prevalence of such recordings varied by the etiology of hearing loss.

Conclusions: Evidence for cochlear synaptopathy was found in CI subjects, so the hypothesis that functional hair cells not revealed in the audiogram can indicate neural survival available for electrical stimulation is a plausible hypothesis for the correlation between ECochG results and speech perception outcomes.