Head Movements Allow Listeners Bilaterally Implanted With Cochlear Implants to Resolve Front-Back Confusions

Abstract

Objectives: We report on the ability of patients fit with bilateral cochlear implants (CIs) to distinguish the front-back location of sound sources both with and without head movements. At issue was (i) whether CI patients are more prone to front-back confusions than normal hearing listeners for wideband, high-frequency stimuli; and (ii) if CI patients can utilize dynamic binaural difference cues, in tandem with their own head rotation, to resolve these front-back confusions. Front-back confusions offer a binary metric to gain insight into CI patients' ability to localize sound sources under dynamic conditions not generally measured in laboratory settings where both the sound source and patient are static.

Design: Three-second duration Gaussian noise samples were bandpass filtered to 2 to 8 kHz and presented from one of six loudspeaker locations located 60° apart, surrounding the listener. Perceived sound source localization for seven listeners bilaterally implanted with CIs, was tested under conditions where the patient faced forward and did not move their head and under conditions where they were encouraged to moderately rotate their head. The same conditions were repeated for 5 of the patients with one implant turned off (the implant at the better ear remained on). A control group of normal hearing listeners was also tested for a baseline of comparison.

Results: All seven CI patients demonstrated a high rate of front-back confusions when their head was stationary (41.9%). The proportion of front-back confusions was reduced to 6.7% when these patients were allowed to rotate their head within a range of approximately ± 30°. When only one implant was turned on, listeners' localization acuity suffered greatly. In these conditions, head movement or the lack thereof made little difference to listeners' performance.

Conclusions: Bilateral implantation can offer CI listeners the ability to track dynamic auditory spatial difference cues and compare these changes to changes in their own head position, resulting in a reduced rate of front-back confusions. This suggests that, for these patients, estimates of auditory acuity based solely on static laboratory settings may underestimate their real-world localization abilities.

Figure 1

Conceptual illustration of the front-back confusions and their resolution with head movements. (A) binaural difference cues are the same for stimuli in front and behind the listener (B) if the listener turns the head to the right, binaural difference cues will migrate towards the left if the sound source is stationary in front of the listener and (C) to the right if the sound source is behind the listener.

Figure 2

Schematic illustration of the experimental apparatus. Stimuli were presented from loudspeakers spaced 60° apart. In previous experiments where all 24 loudspeakers were used, (for a separation of 15°) the root-mean-squared (RMS) localization error in the frontal hemifield for normal hearing listeners was approximately 6°. RMS error for cochlear implant listeners in the same set-up was approximately.29°

Figure 3

Localization measured as loudspeaker identification for 7 normal hearing listeners. Stimuli were3-s duration, bandpass-filtered Gaussian noise (2 kHz - 8 kHz). The radius of each bubble is proportional to the total number of responses, pooled across listeners. Correct responses are plotted along the positively-sloped diagonal in blue. Reversals are plotted along the negatively-sloped diagonal in red.

Figure 4

Localization measured as loudspeaker identification for 7 bilaterally implanted listeners. Otherwise same as Fig. 3.

Figure 5

Summary of the change in performance due to head movements for each of the 7 CI patients calculated as performance with head movements minus performance without head movements. After head movements, all listeners showed an increase in correct localization (blue) and a decrease in front-back reversals (red).Head movement produced no clear trend across listeners for other errors.

Figure 6

Localization measured as loudspeaker identification for 7 patients bilaterally implanted with MED-EL cochlear implants, with the implant at the worse ear turned off. Otherwise same as Fig. 2.