Journal Logo

Special Edition on Cochlear Implants

A Model of a Nucleus 24 Cochlear Implant Fitting Protocol Based on the Electrically Evoked Whole Nerve Action Potential

Franck, Kevin H.

Author Information
  • Free

Abstract

Current Methods of Cochlear Implant Fitting

Accurate implant fitting requires two psychophysical judgments per active electrode. Cochlear Corporation’s CI24M device has 22 stimulating electrodes. The resulting 44 psychophysical judgments can be difficult for the adult patient due to fatigue and the varying-pitch stimulus. Additionally, loudness percepts change with stimulation experience, necessitating repeated psychophysical assessment.

Traditional fitting techniques lose efficacy and efficiency with children who are unable to provide appropriate conditioned responses or loudness judgments. This population is rapidly expanding as cochlear implant candidacy extends to include infants. For some children, complete psychophysical data may require several months of appointments. Although progress in auditory skills may be advancing during this time, the child’s stimulation levels may be poorly balanced.

Studies have shown significant differences in levels and standard errors of measurement obtained using variations of clinical psychophysical implant fitting techniques (Skinner, Holden, Holden & Demorest, 1995). Dawson, Skok, and Clark (1997) found that loudness imbalance due to improperly set psychophysical levels has a significant effect on speech recognition performance.

Another factor to consider is that percepts evoked by single-electrode fitting stimulation are different from percepts evoked by multiple-electrode live-voice stimulation. Such across-electrode loudness summation effects are not considered in the traditional psychophysical implant fitting technique.

Methods for assessing fitting levels have changed little since the multichannel cochlear implant was introduced almost two decades ago. Increasing difficulties are being experienced in the clinic as the number of processing parameters rapidly increases, and changing processing parameters often requires reassessment of fitting levels. There is a need for more efficient fitting techniques that rely less on the patient’s loudness judgments.

The Electrically Evoked Whole NerveAction Potential

The electrically evoked whole nerve action potential (EAP) is a measure of the synchronous primary afferent neural response to electrical stimulation. EAPs generated by stimulation of intracochlear electrodes have been recorded from nonstimulating intracochlear electrodes (Abbas et al., 1999; Brown, Abbas, Borland, & Bertschy, 1996; Brown, Abbas, & Gantz, 1990, 1998; Brown, Hughes, Luk, Abbas, Wolaver, & Gervais, 2000;Finley, Wilson, van den Honert, & Lawson, Reference Note 2; Franck & Norton, 2001; Gantz, Brown, & Abbas, 1994;Hughes, Brown, Abbas, Wolaver, & Gervais, 2000).

High correlations have been found between EAP thresholds and behavioral threshold levels (T-levels), and weaker though significant correlations have been found between EAP thresholds and maximum comfortable loudness levels (C-level) per subject (Brown et al., 1996, 1998., 2000; Franck & Norton, 2001; Hughes et al., 2000). Pooled, correlations with psychophysical fitting levels improve when a subject-dependent constant is added to the EAP predictions (Brown et al., 2000; Franck & Norton, 2001; Hughes et al., 2000). An efficient EAP fitting protocol has been proposed that takes advantage of the high EAP threshold to T-level correlation (Hughes et al., 2000).

Studies have reported positive correlations between the slope of the EAP growth function and the number of surviving spiral ganglion cells in an animal model (Hall, 1990). In human cadavers, the number of spiral ganglion cells is correlated with C-level and dynamic range measures taken while the subjects were alive (Kawano, Seldon, Clark, Ramsden, & Raine, 1998). Modest correlations (r = 0.32, p < 0.01) have also been found between EAP growth function slope and psychophysical dynamic range at individual electrode sites (Franck & Norton, 2001). As the dynamic range is the difference between T- and C-levels, this indicates the potential clinical utility of EAP growth function slope measurement in the determination of cochlear implant fitting C-levels. In addition, knowledge of the location of surviving populations of neurons may provide other important information that will enhance the effectiveness of cochlear implant fitting procedures.

The EAP-based fitting protocol modeled in this study differs from the Hughes et al. (2000) protocol by the addition of the use of EAP growth function data and fitting in a live-voice mode. The purpose of this study is to explore a model of an EAP-based fitting protocol retrospectively on subjects whose EAP responses were previously measured (Franck & Norton, 2001), using the Neural Response Telemetry capabilities of Cochlear Corporation’s CI24M cochlear implant.

Methods

Subjects

This study retrospectively examined data from Franck and Norton (2001) and Franck (Reference Note 3) where details regarding the 12 subjects’ demographics, hearing loss etiology and onset, and speech perception performance were explained in detail.

Loudness Psychophysics for Device Fitting

T- and C-level data were measured with a procedure modified from the ascending loudness judgments with knob (Skinner et al., 1995). The procedure was modified so that T- and C-levels were established twice. Details regarding this method for assessing T- and C-levels were explained in detail in Franck and Norton (2001) and Franck (Reference Note 3).

EAP Measurement

EAP thresholds and growth function slopes from Franck and Norton (2001) and Franck (Reference Note 3) were used in this study. EAP recordings were made using the Neural Response Telemetry software (version 2.01), developed by Dillier and Lai (Reference Note 1), that takes advantage of Cochlear Corporation’s CI24M implant 2-way telemetry system. The subtraction procedure was used to separate the large electrical artifact produced by the stimulation from the response of the neural tissue. See Brown et al. (1990) for a detailed description.

EAP-Based Fitting Protocol

Across the electrode array, the predicted contour of the T-levels (Tcontour) was estimated as the EAP thresholds (VisualThreshold). The predicted T-level contour was not meant to represent the absolute level of threshold stimulation. The predicted dynamic range (DR) between the target T- and C-levels was estimated using parameters from the linear regression (GrowFactor, GrowConstant) and the slope of the EAP growth function (GrowM), measured in microvolts per clinical units (μV/CL). See Equation 1.

Regression parameters were derived from half of the subjects, chosen arbitrarily (alternating), so that the model could be tested on the other half. The split-half design was used to separate the data from which fitting parameters were derived from the data to which the parameters were applied. See Equation 2 for the C-level contour (Ccontour). Example T- and C-level contours are shown in Figure 1A. The T- and C-level contours represented the model’s best estimate of relative stimulation levels. However, these contours did not predict the absolute stimulation levels of the T- and C-level targets.

F1-8
Figure 1:
Retrospective EAP-based fitting protocol. Solid and dashed lines are model and target levels, respectively. A: T- and C-level contours are estimated using EAP data. B: T-level contour is globally adjusted to match target T-levels. C: C-level contour is globally adjusted to match target C-levels. D: predicted and target T- and C-levels are shown together.

The next step to determine the model T-levels was to fit the predicted T-level contour to the absolute T-level target. In this retrospective model, the magnitude of the difference between average T-level contour values and average T-level target values was used as the constant (Tshift), as shown in Equation 3 and Figure 1B.

Similarly, to determine the model C-levels, the magnitude of the difference between average C-level contour values and average C-level target values was used as the constant (Cshift), as shown in Equation 4 and Figure 1C. The final modeled T- and C-levels are shown in Figure 1D.

Results

Regression parameters used to predict the psychophysical dynamic range from the EAP growth function slope from the split-half pool were 0.67 CL2/μV and 31 CL for GrowFactor and GrowConstant, respectively. Parameters from the entire data pool (Franck & Norton, 2001) were 1.25 CL2/μV and 26 CL, respectively. Differences between parameters for the split-half and Franck and Norton (2001) are attributed to the small n(Franck & Norton, 2001).

EAP-Based Fitting Protocol

The model of the EAP-based fitting protocol from half of the subjects was used to estimate T- and C-levels for the other half of the subjects. The modeled and psychophysical T- and C-levels are shown in Figure 2. Differences between measured and modeled data are quantified in Table 1.

F2-8
Figure 2:
Comparison of T- and C-levels from traditional and model EAP-based fitting methods. Solid and dashed lines are measured and modeled levels, respectively.
T1-8
Table 1:
Differences between measured and modeled data.

Discussion

The psychophysical fitting parameters are closely modeled by the EAP-based fitting protocol. This supports the investigation of the clinical use of an EAP-based cochlear implant fitting algorithm.

To realize the model, the values of GrowFactor and GrowConstant should be determined from as large a dataset as possible. Until more data are measured, the parameters from the full dataset of Franck and Norton (2001) could be used. In addition, the clinical manifestation of ramifications of the EAP threshold’s inability to predict the absolute threshold (the need for Tshift and Cshift) would have to be addressed. With the following procedures, this could be accomplished using stimulation in a live-voice mode. A two-step fitting procedure is required because Win-DPS only allows live-voice adjustment as a percentage of the dynamic range, rather than by constant level shifts.

To fit the predicted T-level contour to the subject’s threshold loudness level, Tcontour would be set as C-levels at a subthreshold intensity, such as 100 CL using the Win-DPS software. T-levels would be set at 50 or 100 CL units below the starting C-levels. In a live-voice mode with a loud broadband stimulus, the C-levels would be adjusted with global percentage modifications until the subject reports first hearing. A constant, artificial dynamic range assures that the global modification results in a constant shift across the array. The absolute shift would estimate the Tshift parameter.

Using a new measurement screen, the starting C-level profile, Ccontour, would be set as C-levels near the threshold levels determined in the previous step. T-levels would be set to 50 or 100 CL units below the starting C-levels. In a live-voice mode with a loud broadband stimulus, the C-levels would be adjusted with global percentage modifications until the subject reports maximum comfortable loudness. This would estimate the Cshift parameter.

The C-level settings from both measurement screens would be combined and programmed as the user’s map.

The intensity of broadband noise that would be used in the clinic to adjust T- and C-levels is determined by the input operating range, which is a function of the sensitivity setting. For the CI24M implant, sensitivity can be set from 0 to 20. Each increase in the sensitivity level decreases the input level by 1.5 dB. The dynamic range of input levels is 32 dB. At a sensitivity setting of 10, the C-level where the automatic gain control becomes active is 74 dB SPL. Therefore, when fitting the cochlear implant with the EAP-based protocol, if the sensitivity setting was set at 10, T- and C-levels would be established with a 74 dB SPL broadband live-voice signal. If the sensitivity setting were set to 20, T- and C-levels would be established with a 59 dB SPL signal.

EAP measures may be useful for estimation of psychophysical levels in cochlear implant fitting. This may simplify and increase the accuracy of cochlear implant fitting, especially in populations where behavioral responses are inconsistent or unreliable. This EAP-based implant fitting protocol only requires two loudness judgments from the subject, as opposed to 44 loudness judgments with the traditional method. As opposed to the traditional fitting protocols, this EAP-based fitting protocol allows adjustment in live-voice mode, accounting for across-electrode loudness summation effects perceived by many implant users. In consequence, the EAP-based fitting protocol allows clinicians to fit the cochlear implant in the manner closest to how the implant is used during normal operation.

Acknowledgments:

This work was supported by a personnel training grant (T32 DC00033) from the National Institute on Deafness and Other Communication Disorders, National Institutes of Health, the University of Washington Department of Speech and Hearing Sciences, and the Seattle Children’s Hospital and Regional Medical Center.

References

1. Abbas, P. J., Brown, C. J., Shallop, J. K., Firszt, J. B., Hughes, M. L., Hong, S. H., & Staller, S. J. (1999). Summary of results using the Nucleus CI24M implant to record the electrically evoked compound action potential. Ear and Hearing, 20, 45–59.
2. Brown, C. J., Abbas, P. J., & Gantz, B. (1990). Electrically evoked whole-nerve action potentials: Data from human cochlear implant users. Journal of the Acoustical Society of America, 88, 1385–1391.
3. Brown, C. J., Abbas, P. J., Borland, J., & Bertschy, M. R. (1996). Electrically evoked whole nerve action potentials in Ineraid cochlear implant users: Responses to different stimulating electrode configurations and comparison to psychophysical responses. Journal of Speech and Hearing Research, 39, 453–467.
4. Brown, C. J., Abbas, P. J., & Gantz, B. J. (1998). Preliminary experience with neural response telemetry in the nucleus CI24M cochlear implant. American Journal of Otology, 19, 320–327.
5. Brown, C. J., Hughes, M. L., Luk, B., Abbas, P. J., Wolaver, A., & Gervais, J. (2000). The relationship between EAP and EABR thresholds and levels used to program the Nucleus 24 speech processor: Data from adults. Ear and Hearing, 21, 151–163.
6. Dawson, P. W., Skok, M., & Clark, G. M. (1997). The effect of loudness imbalance between electrodes in cochlear implant users. Ear and Hearing, 18, 156–165.
7. Franck, K. H., & Norton, S. J. (2001). Estimation of psychophysical levels using the electrically evoked compound action potential measured with neural response telemetry capabilities of Cochlear Corporation’s CI24M device. Ear and Hearing, 22, 289–299.
8. Gantz, B. J., Brown, C. J., & Abbas, P. J. (1994). Intraoperative measures of electrically evoked auditory nerve compound action potential. American Journal of Otology, 15, 137–144.
9. Hall, D. R. (1990). Estimation of surviving spiral ganglion cells in the deaf rat using the electrically evoked auditory brainstem response. Hearing Research, 45, 123–136.
10. Hughes, M. L., Brown, C. J., Abbas, P. J., Wolaver, A. A., & Gervais, J. P. (2000). Comparison of EAP thresholds with MAP levels in the Nucleus 24 cochlear implant: Data from children. Ear and Hearing, 21, 164–174.
11. Kawano, A., Seldon, H. L., Clark, G. M., Ramsden, R. T., & Raine, C. H. (1998). Intracochlear factors contributing to psychophysical percepts following cochlear implantation. Acta Otolaryngologica (Stockholm), 118, 313–326.
12. Skinner, M. W., Holden, L. K., Holden, T. A., & Demorest, M. E. (1995). Comparison of procedures for obtaining thresholds and maximum acceptable loudness levels with the nucleus cochlear implant system. Journal of Speech and Hearing Research, 38, 677–689.

Reference Notes

1. Dillier, N., & Lai, W. K. (1998). Neural Response Telemetry (Version 2.01). Zurich, Switzerland.
2. Finley, C., Wilson, B., van den Honert, C., & Lawson, D. T. (1997). Intracochlear evoked potentials in response to pairs of pulses: Effects of pulse amplitude and interpulse interval. Speech Processors for Auditory Prostheses (contract N01-DC-5–2103), Quarterly Progress Report #6.
3. Franck, K. H. (1999). The electrically evoked whole-nerve action potential: Fitting applications for cochlear implant users. Unpublished Doctoral Dissertation, University of Washington, Seattle, WA.
© 2002 Lippincott Williams & Wilkins, Inc.