How Noise (un)affects Fine-Grain ECAP Recordings

Background

Recordings of evoked compound action potentials (ECAP) using Cochlear Implants (CI) are affected by noise, but the impact can be reduced by averaging across sufficient (30-50) repetitions. The compromise between noise reduction and measurement duration is mainly limited by the maximum repetition rate (100-200 Hz) ensuring a constant electro-physical environment.

Objectives

Pulses with quasi-continuous increase of stimulation amplitudes (Fine-Grain) result in an overall duration being dependent on repetition rate, magnitude and number of steps for stimulation amplitude increase. By keeping the measurement duration fixed, a variation of these parameters affects accuracy and the perceived loudness increase of CI users.

Methods

Test-retest reliability of ECAP thresholds and magnitude of the noise floor are regarded. The expected accuracy based on theoretical calculations is compared to the achieved accuracy for human data (13 subjects, 22 ears). The repetition rate was varied (40Hz, 60Hz, and 76Hz) while magnitude and number of steps for amplitude increase were adjusted in order to achieve the same measurement duration.

Results

Theoretical calculations predict for decreased repetition rates an increased noise floor and reduced test-retest accuracy for ECAP-thresholds (10% deviation). Human data confirms the prediction, but shows less dependency on repetition rate.

Conclusions

We conclude that the FineGrain method is very robust concerning unsystematic noise. We thank the staff of Instituto Arauz in Buenos Aires for collecting the data.