2019 Conference on Implantable Auditory Prostheses

14-19 July 2019

Granlibakken, Lake Tahoe

Page 62

P35: REWEIGHTING OF BINAURAL LOCALIZATION CUES IN

BILATERAL COCHLEAR IMPLANT LISTENERS

Maike Ferber, Bernhard Laback, Martin Lindenbeck

Acoustics Research Institute, Austrian Academy of Sciences, Vienna, AUT

For spatial hearing within the horizontal plane, normal-hearing (NH) listeners rely primarily on

the two binaural cues interaural time difference (ITD) and interaural level difference (ILD). They

apply frequency-dependent weights when combining these cues to determine the perceived

azimuth of a sound source. Low-frequency fine structure ITD cues are generally considered to

be most salient. Cochlear implant (CI) listeners, however, rely almost entirely on ILDs, while

ITDs contribute only little or not at all. This has two main reasons: 1) Current envelope-based CI

stimulation strategies using constant-high-rate pulse trains do not encode fine structure ITDs

and available envelope ITD cues are not sufficiently salient. 2) Even when presenting highly

controlled fine structure ITDs using a research interface and the most sensitive pulse rate, CI

listeners’ sensitivity is lower and much more variable across listeners compared to NH listeners’

fine structure ITD sensitivity. A recent study in our lab showed that NH listeners changed their

ITD/ILD weighting when one of the cues was reinforced through visual feedback. Such ITD/ILD

reweighting could be a factor contributing to the low ITD sensitivity of CI listeners, since in daily

experience with current CI-systems, ILDs consistently indicate the sound source locations while

ITDs provide no robust localization cues.

We adapted our NH experiment to bilateral CI listeners stimulated with a research interface. The

experiment was conducted in a virtual environment and consisted of a pretest to establish the

initial ITD/ILD weights, a multi-day training, in which visual feedback reinforced the ITD

locations, and a posttest to measure the effect of training on the weights. We additionally

measured just noticeable differences (JNDs) for ITD and ILD before and after training. ILDs

corresponding to a range of azimuthal locations were derived individually in a separate pretest

and combined with naturally occurring ITDs in various spatially inconsistent combinations, using

a maximum azimuthal inconsi

stency of 24°. Participants’ task was to lateralize 100

-pps and 300-

pps pulse trains presented at a single interaurally place-matched electrode pair. Feedback

consistent with the ITD location was provided.

The data collection is still ongoing; however, preliminary results are promising and suggest an

increase in ITD weights, especially in the 300-pps condition. So far, increased post-training ITD

weights in lateralization do not seem to be reflected in corresponding post-training changes of

ITD and ILD JNDs.

Based on our previous results with NH listeners, we propose a mechanism potentially

contributing to CI listeners’ low ITD sensitivity. Additionally, the present results are promising in

terms of making fine structure ITDs better usable for CI listeners using future CI-systems that

better convey ITD information.

Supported by WTZ project SpaCI (#MULT_DR 11/2017) and EU Marie Curie grant ALT (H2020-

MSCA-RISE-2015 #691229)