Clinical experiences with intraoperative electrocochleography in cochlear implant recipients and its potential to reduce insertion trauma and improve postoperative hearing preservation

Abstract

Access to low-frequency acoustic information in cochlear implant patients leads to better speech understanding in noise. Electrocochleography (ECochG) can provide real-time feedback about the health of the cochlea during the insertion process with the potential to reduce insertion trauma. We describe our experiences of using this technique. Data from 47 adult subjects with measurable residual hearing and an Advanced Bionics (Valencia, CA) SlimJ (46) or MidScala (1) electrode array were analyzed. ECochGs were recorded intraoperatively via the implant. The surgeon adjusted the course of the electrode insertion based on drops in the ECochG. The final array position was assessed using postoperative imaging and pure tone thresholds were measured before and after surgery. Three different patterns of ECochG response amplitude were observed: Growth, Fluctuating and Total Loss. Subjects in the growth group showed the smallest postoperative hearing loss. However, the group with fluctuating amplitudes showed no meaningful correlation between the ECochG responses and the postoperative hearing loss, indicating that amplitude alone is insufficient for detecting damage. Considering the phase of the signal additionally to the amplitude and reclassifying the data by both the phase and amplitude of the response into three groups Type I-Type III produced statistically significant correlations between postoperative hearing loss and the grouping based on amplitude and phase respectively. We showed significantly better hearing preservation for Type I (no drop in amplitude) and Type II (drop with a concurrent phase shift), while Type III (drop without concurrent phase shift) had more surgery induced hearing loss. ECochG potentials measured through the implant could provide valuable feedback during the electrode insertion. Both the amplitude and phase of the ECochG response are important to consider. More data needs to be evaluated to better understand the impact of the different signal components to design an automated system to alert the surgeon ahead of damaging the cochlea.

Conflict of interest statement

I have read the journal’s policy and the authors of this manuscript have the following competing interests: GG is employer of Advanced Bionics. TL and AB are members of the medical/audiological advisory boards for Advanced Bionics and other cochlear implant manufacturers. The commercial affiliation of GG does not alter our adherence to all PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Cochlear microphonic recordings for each subject made intraoperatively via the cochlear implant showing the large variability in responses.

Total amplitude of the responses ranged from ~5 μV to over 200 μV (blue line). Light grey area shows the noise floor. The subject ID in each axis is coloured according to the group categorisation (green = Growth, yellow = Fluctuating, red = Total Loss, see also Fig 5).

Fig 2. 3D visualization of cochlea and electrode position from subject ID 28, showing a potential soft contact with the basilar membrane and some slight buckling in the middle part of the array.

Fig 3. Top view from left to right.

ID 24 (Scala tympani without touching the basilar membrane, relatively deep insertion in a small cochlea); ID 27 (translocation of MidScala array); ID 28 (soft contact with the basilar membrane and some slight buckling in the middle part of the array); ID 35 (good position, little less than full insertion).

Fig 4. Median hearing loss, calculated as the difference between pre-operative hearing and 4 weeks post-surgery, at each frequency for 47 subjects.

Fig 5. Cochlea microphonic waveform examples for each of the Growth (n = 11), Fluctuating (n = 25) and Total loss (n = 4) groups and the threshold calculations used to define them.

Fig 6. Average low frequency hearing loss after 4 weeks for the three groups of subjects categorized by the intraoperative cochlear microphonic amplitude.

Fig 7. Cochlear Microphonic amplitude (blue) and phase (red) over insertion time.

The top panel shows a Type I example where the amplitude does not drop by more than 5 dB. In the middle panel (Type II) each drop is accompanied by large phase changes. The insertion in the lower panel is classified as Type III because the two drops around an insertion time of 100 s don’t come with significant phase changes.

Fig 8. Average low frequency hearing loss after 4 weeks for the three groups of subjects with no drop > 5 dB (Type I), a drop > 5 dB with concurrent phase shift (Type II) and a drop > 5 dB without concurrent phase shift (Type III).