Effects of sensorineural hearing loss on visually guided attention in a multitalker environment

Abstract

This study asked whether or not listeners with sensorineural hearing loss have an impaired ability to use top-down attention to enhance speech intelligibility in the presence of interfering talkers. Listeners were presented with a target string of spoken digits embedded in a mixture of five spatially separated speech streams. The benefit of providing simple visual cues indicating when and/or where the target would occur was measured in listeners with hearing loss, listeners with normal hearing, and a control group of listeners with normal hearing who were tested at a lower target-to-masker ratio to equate their baseline (no cue) performance with the hearing-loss group. All groups received robust benefits from the visual cues. The magnitude of the spatial-cue benefit, however, was significantly smaller in listeners with hearing loss. Results suggest that reduced utility of selective attention for resolving competition between simultaneous sounds contributes to the communication difficulties experienced by listeners with hearing loss in everyday listening situations.

FIG. 1

Across-subject mean pure-tone thresholds for left and right ears in the normal-hearing (NH) and hearing loss (HL) listener groups. Error bars indicate SDs.

FIG. 2

A Schematic of the layout of the five loudspeakers relative to the listener’s head. B Schematic of the four attention conditions. Sounds were presented from the five loudspeakers simultaneously and were divided into five contiguous time segments. The forward-speech target (T) was presented during one random time segment from one random loudspeaker, and in that time segment, the other four loudspeakers contained maskers. In all nontarget time segments, maskers were presented from all loudspeakers. While the auditory stimuli were the same across conditions, visual cues (gray regions) varied in order to manipulate attention. In the no cue condition, there was no visual cue; in the other three conditions, the cue indicated where or when to listen (or both).

FIG. 3

Mean level variations over time for the five-digit target sequences (averaged across the 125 target sequences after time-normalization).

FIG. 4

Across-subject mean performance (top panel) and cue benefits (bottom panel) for the three groups of listeners (HL hearing loss, NH normal-hearing, NH −3 dB normal-hearing with an attenuated target). The four bars for each group represent the four attention conditions. Error bars indicate SEs of the means.

FIG. 5

Across-subject mean performance (top panel) and cue benefits (bottom panel) as a function of digit position for the three groups of listeners (HL hearing loss, NH normal-hearing, NH −3 dB normal-hearing with an attenuated target). The four different lines and symbols represent the four attention conditions. Error bars indicate SEs of the means.