Psychophysical measures from electrical stimulation of the human cochlear nucleus

Auditory performance on basic psychophysical tasks was measured in ten deaf patients with electrodes positioned near their cochlear nucleus. The device is called the auditory brainstem implant (ABI). Electrodes were placed during surgery to remove an acoustic neuroma, which results in the removal of the VIII nerve and, thus deafness. In patients who received auditory sensation from electrical stimulation we measured auditory performance on standard psychophysical tasks: thresholds, loudness growth, intensity discrimination, temporal integration, temporal modulation detection, gap detection, and forward masking. Plots of threshold as a function of frequency or biphasic pulse duration were markedly different from those of patients with cochlear implants. The difference in threshold functions is probably partly due to the biophysical difference in the neural elements stimulated. Another possibility is that part of the difference is due to the highly abnormal spatial pattern of activation in the cochlear nucleus from electrical stimulation, which prevents normal spatial integration of activity. The usable range of electrical amplitudes above threshold is comparable with that of cochlear implants, typically 10-15 dB. Little temporal integration occurs over a range of stimulus durations from 2-1000 ms. When compared at equivalent loudness levels, gap detection thresholds are similar to, or a bit longer than, gap thresholds in normal-hearing listeners and cochlear implant patients. Forward masking recovery functions are similar to those of normal listeners and cochlear implant patients. Patients' ability to detect amplitude modulation as a function of modulation frequency is similar to that of cochlear implant patients and normal listeners. Thus, direct electrical stimulation of the brainstem produces temporal resolution that does not significantly differ from that of normal listeners when compared in equivalent amplitude units. This implies that the limiting factors for these tasks are more centrally located, and not directly related to threshold mechanisms. Thus, a properly designed speech processor could preserve the important temporal features of speech for these patients.