2019 Conference on Implantable Auditory Prostheses

14-19 July 2019

Granlibakken, Lake Tahoe

Page 271

W41: A BINAURAL ADVANCED COMBINATION ENCODER STRATEGY FOR SOUND

LOCALIZATION WITH THE CCI-MOBILE RESEARCH PLATFORM

Stephen R. Dennison, Alan Kan, Ruth Y. Litovsky

Electrical & Computer Engineering, University of Wisconsin-Madison, Madison, WI, USA

Waisman Center, University of Wisconsin-Madison, Madison, WI, USA

Bilateral cochlear implants (biCIs) restore speech understanding and some spatial hearing

abilities. However, there is still a gap in performance between listening with biCIs and normal

hearing, especially in real-life noisy environments. In particular, recent data from our lab showed

that biCI listeners had difficulty distinguishing between stationary and moving sounds, and the

direction of movement of non-stationary sounds. While some of the limitations in performance

can be attributed to pathology and the surgical challenge of placing two implants at the same

insertion depth, a part of the gap may instead be caused by a lack of synchrony between the

two sound processors. For clinical CIs, each sound processor is controlled by an independent

time clock, and there is no communication across processors. Furthermore, each CI processes

sounds independently in each ear. For n of m sound processing strategies, such as the

Advanced Combination Encoder (ACE) strategy, the processors in each ear could be selecting

different electrodes to stimulate, thereby reducing the fidelity of binaural cues such as interaural

time and level differences (ITD and ILD, respectively).

The CCi-Mobile research processor developed by UT-Dallas aims to solve these two problems;

it has one system clock and both implants are controlled from a central processor. This allows

for precise control of the interaural cues presented via electrical stimulation. The central

processor allows for real-time binaural signal processing which enables the development of truly

binaural processing strategies, and the testing of these strategies in real-time, realistic

conditions. Hence, we developed a binaural ACE (bACE) strategy to test the hypothesis that

synchronized processing and stimulation will improve sound localization performance. In

traditional ACE, the n highest peaks in each time frame are chosen for stimulation. With

independent processing in each ear, the n channels chosen may not always be the same.

Hence, binaural cues, which are compared in the brain on a frequency-by-frequency basis, may

not be transmitted on the same electrodes across the ears, resulting in disrupted binaural

processing. In bACE, channel selection is conducted on inputs from both ears together. The n/2

highest peaks in each ear are selected and the corresponding electrode in the opposite ear is

also chosen for stimulation.

We conducted a localization experiment with bilateral Cochlear users where the stimuli

consisted of stationary and moving sounds that span trajectories of 10°, 20°, or 40°. Three

conditions were compared: ACE and bACE implemented on the CCi-

Mobile, and participant’s

listening with their own clinical processors. Our results will help determine whether synchronized

processing improves discrimination of stationary and moving sounds, and the direction of

moving sounds.

This work was supported by NIH-NIDCD R01DC016839 and NIH-NIDCD R01DC03083 to RYL,

NIH-NIDCD R03DC015321 to AK, and NIH-NICHD U54HD090256 to the Waisman Center.