Electrophysiologically-based electrode selection has the potential to improve speech perception in cochlear-implant...

There is large variability in speech perception outcomes across cochlear-implant (CI) users. One factor that contributes to this variation is the neural modulation encoding in the periphery, which can vary along the implant electrode array. Since temporal envelope cues are crucial for speech perception with a CI, previous studies have shown that there is potential of deactivating electrodes based on poorer behavioral measures of neural modulation processing to improve speech perception. However, behavioral measures are typically time-consuming and require active feedback from the CI recipient. A potentially useful objective measure of neural modulation processing is the electrically-evoked auditory steady-state response (eASSR). Recently, the across-array variation of eASSRs has been shown to strongly correlate with speech perception in noise in CI users. In the present study, we demonstrate feasibility to measure 40-Hz eASSRs with clinical pulse rates across multiple electrodes. Next, we investigated whether objective electrode-selection based on individual across-array eASSR patterns has the potential to improve speech perception in CI users. 40-Hz eASSRs were recorded across the whole implant electrode array by means of EEG. A custom-built EEG system with a resolution of 262 kHz was used in order to be able to remove CI-stimulation electrical artifacts. Next, the across-array variation of eASSR amplitudes was used as a basis for individualized electrode-selection. Two experimental MAPs with 11 electrodes (MAPs A and B) were created for each participant: MAP A retained electrodes that were considered better at conveying temporal envelope cues, and MAP B those that were considered poorer at it. Speech perception performance with the clinical and the two experimental MAPs was assessed using speech perception tasks in quiet and in noise, and correlated with eASSR pattern metrics after accounting for the effect of tonotopical changes on speech perception due to electrode deactivation. Results showed that MAP A consistently performed better than MAP B in all three listening conditions, but both still perform worse than the clinical MAP. Furthermore, results suggested that higher eASSR amplitudes overall and greater across-array variation were associated with better performance with MAP A, even after controlling for tonotopical effects. No such associations were observed for MAP B. These findings suggest that automatic, objective electrode-selection strategies based on local neural modulation encoding of cochlear regions along the electrode array is useful for individual clinical CI fitting, with the potential to improve speech perception outcomes in CI recipients.