Frequency-to-Place Mismatch and Cochlear Implant Outcomes by Electrode Type

Abstract

Speech recognition outcomes with a cochlear implant (CI) are highly variable. One factor suggested to correlate with CI-aided speech recognition is frequency-to-place mismatch, or the discrepancy between the natural tonotopic organization of the cochlea and the electric frequency allocation of the CI electrodes within the patient's cochlea. To evaluate the association between frequency-to-place mismatch and speech recognition outcomes in a large cohort of postlingually deafened adult CI users, while controlling for various clinical factors known to be associated with those outcomes. This retrospective cohort study used data from a CI program at a tertiary medical center and included CIs from postlingually deafened adult CI users. After excluding patients whose data were not logged, patients with implantations occurring between 2016 and 2023 were included in the analysis. The data were extracted in November 2023. Results of the Consonant-Nucleus-Consonant (CNC) monosyllabic word recognition test measured in the CI-aided alone condition 1 month, 3 months, 6 months, and 12 months after activation served as the main outcome. The independent variables included frequency-to-place mismatch, electrode array type, mean modiolar distance, electrode position, age at implantation, biological sex, contralateral hearing abilities, time since CI activation (test interval), and daily device use. In 498 CIs from 447 postlingually deafened adults (mean [SD] age, 63.1 [17.1] years; 271 [54.4%] CIs with male users), frequency-to-place mismatch was negatively correlated with CI-aided speech recognition outcomes, but the association was only significant for precurved electrode arrays and not straight electrode arrays. In the linear mixed effects model for straight electrode arrays, only test interval (β = 1.14 [95% CI, 0.90-1.38]) and daily device use (β = 0.90 [95% CI, 0.42-1.38]) were correlated with the improvement of word recognition over the first year of device use. In the model for precurved electrode arrays, mismatch at 1500 Hz (β = -0.011 [95% CI, -0.011 to -0.006]), scalar location (β = 16.37 [95% CI, 9.01 to 23.74]), test interval (β = 1.18 [95% CI, 1.18-1.41]) and daily device use (β = 1.65 [95% CI, 1.15-2.14]) all were significantly associated with the improvement of word recognition over the first year of device use. In this cohort study of postlingually deafened adult CI users, including both straight and precurved electrode arrays, daily device use and time since CI activation were found to be significantly associated with improved CI-aided speech recognition outcomes. Frequency-to-place mismatch at 1500 Hz and scalar location were associated with word recognition only for precurved arrays. These findings suggest that, for patients with straight arrays, any interference produced by frequency-to-place mismatch may be overcome by adaptation, which could be supported by daily device use. However, for patients with precurved arrays, daily device use may not be sufficient to completely overcome electrode placement factors, including scalar location and mismatch.