TEMPORAL PITCH PERCEPTION IN COCHLEAR IMPLANTEES

Andrew E. Vandali1, Cathy Sucher2, David J. Tsang2, Colette M. McKay2,

Jason W. D. Chew3, Hugh J. McDermott2, Richard van Hoesel1

1 The Cooperative Research Centre for Cochlear Implant and Hearing Aid

Innovation (CRC HEAR), Melbourne, Australia.

2 The Department of Otolaryngology, University of Melbourne, Australia

3 The School of Audiology, University of Melbourne, Australia

Temporal pitch perception was investigated in three adult Nucleus 24

cochlear implant users using amplitude modulated stimuli presented on

multiple electrodes (up to 9) in quick succession. Pitch difference

limens (DLs) were measured in a two alternative forced choice (2AFC)

procedure which required subjects to identify which of a pair of stimuli

differing in modulation frequency was higher in pitch. Place of

stimulation was held constant and intensity cues to pitch were randomized

by roving the stimulus level by up to 6 dB. Results showed that

modulation depths of at least 20-40% of the subject’s electrical dynamic

range resulted in pitch DLs as low as 1 to 2 semitones, for some of the

subjects, for modulation frequencies up to approximately 300 Hz. These

results are consistent with data from previous modulation studies using

single electrodes. The effect of introducing phase differences in the

modulation envelopes applied to different regions within a fixed group of

9 adjacent electrodes was also examined. Phase differences of 180 degrees

were found to be detrimental to the task when regions of identical phase

comprised approximately 3 or less adjacent electrodes.

In most sound coding strategies currently used clinically, the depth of

fundamental frequency (F0) modulation within each channel and the phase

relationship between modulation in neighboring channels can vary

substantially across channels and with signal thereby providing

inconsistent, or poorly coded, temporal cues to pitch. To overcome these

analysis/coding problems, a number of experimental processing techniques

were developed (based on the above data) that provided deeper modulation

cues to F0 in the stimulus envelope coincidentally in time across all

activated electrodes. Pitch ranking abilities were measured in five

subjects using these strategies as well as clinical strategies ACE and

CIS. A 2AFC procedure was employed using sung vowel stimuli separated in

F0 by half an octave. Small but significant benefits in pitch ranking for

F0’s less than approximately 300 Hz were found for the experimental

strategies compared to ACE and CIS. In addition, results of speech

recognition tests conducted in quiet and noise demonstrated equivocal

performance between ACE and one of the experimental strategies. However,

it was not clear from these data whether similar benefits to pitch

perception could be expected in everyday situations, such as when

following speaker intonation, when listening to music, or when

discriminating between lexical contrasts in tonal languages.

Investigations in this area are in progress.

Support provided by CRC HEAR and the Garnett Passe and Rodney Williams

Memorial Foundation.

 

back