ECAP measures predict cochlear implant behavioural thresholds

© 2015 Nicholas SmaleAim: To estimate high rate stimulus behavioural threshold current levels of cochlear implant (CI) users by comparing peak amplitudes and associated latencies of electrically evoked compound action potentials (ECAPs) evoked by different stimulation pulse parameters. Background: CIs create a perception of sound by electrically stimulating the cochlea. The current level for each electrode that corresponds to the minimum amount of stimulation required to create a perception of sound is presently determined through a lengthy and subjective clinical process. This process can be particularly challenging for infants and the long-term deaf. With adequate stimulation, peripheral auditory neurons generate an electrical response called an ECAP. The minimum amount of stimulation required to elicit ECAPs has been assumed to be able to predict behavioural thresholds. This has been found to be untrue, and ECAP-based behavioural thresholds produce sub-optimal speech perception outcomes. There is evidence that differences between ECAP- and behavioural thresholds are caused by differences between individuals’ peripheral auditory neuron survival. Separately, there is evidence that the extent of change in certain ECAP measures with change in stimulation conditions is also affected by peripheral neural survival. We suggest that these ECAP measures may therefore be directly used to estimate differences between ECAP- and behavioural thresholds. Methods: Amplitude growth function series with four different stimulation parameters were created from ECAP thresholds to C-levels for 52 electrodes of 10 adult CI users. ECAPs were recorded intracochlearly through neural response telemetry. The current level difference required to equalise ECAP amplitude, and the change in ECAP peak latency, between stimulation conditions was determined for each electrode. Behavioural thresholds at stimulation rates of 40, 500, 1000 and 2000 pps were determined for each electrode using a 3-interval-forced-choice task. Pearson correlations were performed between equalising CL and low rate changes in behavioural thresholds, both per electrode and per subject mean. Stepwise multiple regression was used to directly estimate higher rate thresholds. Results: ECAP measures significantly improved ECAP threshold based predictions of higher stimulation rate behavioural thresholds (1000 pps: R2 = 0.338, n = 52). Conclusions: This technique is superior to purely ECAP threshold-based fitting, improving predictions of behavioural thresholds. These methods will require validation, but show promise as a new clinical method to create better speech-perception outcomes for certain CI users