Melodic Contour Identification by Cochlear-Implant Listeners With Asymmetric Phantom Pulses Presented to Apical Electrodes.

Abstract

Objectives: (a) To compare performance by cochlear-implant listeners on a melodic contour identification task when the fundamental frequency (F0) is encoded explicitly by single-pulse-per-period (SPP) pulse trains presented to an apical channel, by amplitude modulation of high-rate pulse trains presented to several electrodes, and by these two methods combined, (b) to measure melodic contour identification as a function of the range of F0s tested, (c) to determine whether so-called asymmetric phantom stimulation improves melodic contour identification relative to monopolar stimulation, as has been shown previously using pitch-ranking tasks.

Design: Three experiments measured melodic contour identification by cochlear-implant listeners with two different methods of encoding fundamental frequency (F0), both singly and in combination. One method presented SPP pulse trains at the F0 rate to an apical channel in either partial-bipolar or monopolar mode. The second method applied amplitude modulation at F0 to high-rate (~2000 pulses per second) pulse trains on six adjacent electrodes. For this "MOD" stimulation, the channel envelopes were misaligned so as to simulate the effects of the bandpass filters in the commercial signal-processing strategy.

Results: In experiment 1, the SPP stimulation used the asymmetric phantom method: pseudomonophasic pulses were applied in partial-bipolar mode to electrodes 1 and 3, with 25% of current returned via an extra-cochlear electrode, and with the short high-amplitude phase anodic with respect to electrode 1. The MOD stimuli were presented to a set of basal electrodes. Performance for SPP stimulation was better, both when presented alone and when combined with MOD stimulation, relative to MOD stimulation alone. Performance was also better when the range of F0s present in the stimuli spanned a low range (97 to 194 Hz) than when they spanned a medium (161 to 322 Hz) or a high range (242 to 484 Hz). Experiment 2 was similar to experiment 1 except that the MOD stimuli were presented to a set of six apical electrodes. Performance with SPP stimulation alone was again significantly better than with MOD stimulation, but the difference between combined and MOD stimulation was not significant. Experiment 3 compared performance of SPP stimulation applied in asymmetric phantom mode to monopolar stimulation of the most-apical electrode using symmetric biphasic pulses. No differences were found between these two types of stimulation, either presented in isolation or with MOD stimulation of nearby apical electrodes.

Conclusions: The results show that F0 encoding by SPP stimulation was better than with MOD stimulation, that it was robust to possible interference from MOD-stimulated electrodes, but that performance with combined stimulation was not better than with SPP alone. Contrary to previous data from pitch-ranking studies, we found no evidence that asymmetric phantom improved melodic contour identification compared with the monopolar stimulation used in commercial strategies.