Improved Cochlear Implant Performance Estimation Using Tonotopic-Based Electrocochleography

Abstract

Importance: Cochlear implantation produces remarkable results in postlingual deafness, although auditory outcomes vary. Electrocochleography (ECochG) has emerged as a valuable tool for assessing the cochlear-neural substrate and evaluating patient prognosis.

Objective: To assess whether ECochG-total response (ECochG-TR) recorded at the best-frequency electrode (BF-ECochG-TR) correlates more strongly with speech perception performance than ECochG-TR measured at the round window (RW-ECochG-TR).

Design, setting, and participants: This single-center cross-sectional study recruited 142 patients from July 1, 2021, to April 30, 2022, with 1-year follow-up. Exclusions included perilymph suctioning, crimped sound delivery tubes, non-native English speakers, inner ear malformations, nonpatent external auditory canals, or cochlear implantation revision surgery.

Exposures: Cochlear implantation.

Main outcomes and measures: Speech perception testing, including the consonant-nucleus-consonant (CNC) words test, AzBio sentences in quiet, and AzBio sentences in noise plus 10-dB signal to noise ratio (with low scores indicating poor performance and high scores indicating excellent performance on all tests), at 6 months postoperatively; and RW-ECochG-TR and BF-ECochG-TR (measured for 250, 500, 1000, and 2000 Hz).

Results: A total of 109 of the 142 eligible postlingual adults (mean [SD] age, 68.7 [15.8] years; 67 [61.5%] male) were included in the study. Both BF-ECochG-TR and RW-ECochG-TR were correlated with 6-month CNC scores (BF-ECochG-TR: r = 0.74; 95% CI, 0.62-0.82; RW-ECochG-TR: r = 0.67; 95% CI, 0.54-0.76). A multivariate model incorporating age, duration of hearing loss, and angular insertion depth did not outperform BF-ECochG-TR or RW-ECochG-TR alone. The BF-ECochG-TR correlation with CNC scores was significantly stronger than the RW-ECochG-TR correlation (r difference = -0.18; 95% CI, -0.31 to -0.01; z = -2.02). More moderate correlations existed between 6-month AzBio scores in noise, Montreal Cognitive Assessment (MoCA) scores (r = 0.46; 95% CI, 0.29-0.60), and BF-ECochG-TR (r = 0.42; 95% CI, 0.22-0.58). MoCA and the interaction between BF-ECochG-TR and MoCA accounted for a substantial proportion of variability in AzBio scores in noise at 6 months (R2 = 0.50; 95% CI, 0.36-0.61).

Conclusions and relevance: In this case series, BF-ECochG-TR was identified as having a stronger correlation with cochlear implantation performance than RW-ECochG-TR, although both measures highlight the critical role of the cochlear-neural substrate on outcomes. Demographic, audiologic, and surgical factors demonstrated weak correlations with cochlear implantation performance, and performance in noise was found to require a robust cochlear-neural substrate (BF-ECochG-TR) as well as sufficient cognitive capacity (MoCA). Future cochlear implantation studies should consider these variables when assessing performance and related interventions.

Figure 1.. Tonotopic Organization of Cochlear Microphonic (CM) Responses in a Single Patient

The CM was recorded from even-numbered electrodes of the slim perimodiolar electrode array. Stimulus frequencies of 250, 500, 1000, and 2000 Hz are displayed. The top row represents responses from the most apical electrode (intracochlear electrode 22), the second row from electrode 20, and the bottom row from electrode 2. This figure shows the clear tonotopic arrangement of the cochlea through these CM responses. The electrode with the maximum amplitude from the fast Fourier transformation was selected as the best-frequency electrode for that particular frequency.

Figure 2.. Extraction of the Best-Frequency (BF) Electrocochleography Total Response (BF-ECochG-TR) From a Single Patient

A, Difference and summation waveforms derived from rarefaction and condensation phases at the BF intracochlear site for each frequency, accompanied by their respective fast Fourier transformations. The significant first, second, and third harmonics are indicated with footnote “a,” as determined by response magnitudes surpassing the noise floor by 3 SDs. B, Cochlear microphonic (CM) tuning curve from 250 to 2 kHz. C, Visual depiction of a cochlear implant array consisting of 22 electrode contacts, highlighting the BF location for the range of 250 to 2 kHz. D, Comparison of ECochG-TR measured from the round window (RW-ECochG-TR) and BF-ECochG-TR for 250, 500, 1000, and 2000 Hz. dB re:1μV indicates decibels relative to 1 μV.

Figure 3.. Predictive Capability of Best-Frequency Electrocochleography–Total Response (BF-ECochG-TR) for Speech Perception Performance

A, A robust linear association between round window electrocochleography–total response (RW-ECochG-TR) and BF-ECochG-TR is displayed. Light blue dots represent patients with larger RW-ECochG-TR values than the reference line, indicating insertion trauma confirmed by recorded insertion patterns; these patients were retained in the final analysis. The diagonal line running through the center point indicates the line of equality. B, A strong linear association is shown between RW-ECochG-TR and consonant-nucleus-consonant (CNC) scores at 6 months. C, A marked linear correlation exists between BF-ECochG-TR measured at 250, 500, 1000, and 2000 Hz and 6-month CNC scores. D, A notable linear association is observed between BF-ECochG-TR measured at 500, 1000, and 2000 Hz and CNC scores at 6 months.