Auditory cortical images of cochlear-implant stimuli: coding of stimulus channel and current level.

Abstract

This study quantified the accuracy with which populations of neurons in the auditory cortex can represent aspects of electrical cochlear stimuli presented through a cochlear implant. We tested the accuracy of coding of the place of stimulation (i.e., identification of the active stimulation channel) and of the stimulus current level. Physiological data came from the companion study, which recorded spike activity of neurons simultaneously from 16 sites along the tonotopic axis of the guinea pig's auditory cortex. In that study, cochlear electrical stimuli were presented to acutely deafened animals through a 6-electrode animal version of the 22-electrode Nucleus banded electrode array (Cochlear). Cochlear electrode configurations consisted of monopolar (MP), bipolar (BP + N) with N inactive electrodes between the active and return electrodes (0 < or = N < or = 3), tripolar (TP) with one active electrode and two flanking return electrodes, and common ground (CG) with one active electrode and as many as five return electrodes. In the present analysis, an artificial neural network was trained to recognize spatiotemporal patterns of cortical activity in response to single presentations of particular stimuli and, thereby, to identify those stimuli. The accuracy of pair-wise discrimination of stimulation channels or of current levels was represented by the discrimination index, d', where d' = 1 was taken as threshold. In many cases, the threshold for discrimination of place of cochlear stimulation was < 0.75 mm, and the threshold for discrimination of current levels was < 1 dB. Cochlear electrode configurations varied in the accuracy with which they signaled to the auditory cortex the place of cochlear stimulation. The BP + N and TP configurations provided considerably greater sensitivity to place of stimulation than did the MP configuration. The TP configuration maintained accurate signaling of place of stimulation up to the highest current levels, whereas sensitivity was degraded at high current levels in BP + N configurations. Electrode configurations also varied in the dynamic range over which they signaled stimulus current level. Dynamic ranges were widest for the BP + 0 configuration and narrowest for the TP configuration. That is, the configuration that showed the most accurate signaling of cochlear place of stimulation (TP) showed the most restricted dynamic range for signaling of current level. These results suggest that the choice of the optimal electrode configuration for use by human cochlear-prosthesis users would depend on the particular demands of the speech-processing strategy that is to be employed.