Effects of Speaker-Identity Cueing on Listening Effort During Speech-in-Noise

The brain is an organ that performs a variety of intricate functions. Specifically, the brain has an amazing ability to recover the complexities of a speech signal within a mixture of sounds. The process of extracting the speech signal from background noise, however, is not necessarily straightforward or easy. Previous studies have developed the concept of “listening effort” as an umbrella term to include all cognitive demand listeners confront to understand speech. From a clinical standpoint, this term suggests that accuracy measurements alone are not sufficient, and a supplementary assessment of how hard a client must try in order to understand speech (especially when the speech is degraded due to background noise) must be conducted. Current research emphasis is on the post-speech-time compensatory processes in recovering speech cues. However, in this study, we claim pre-speech-time attentional processes also create a source of listening effort. To support this idea, we measured the cortical, behavioral, and pupillary responses of 19 normal-hearing participants to SiN conditions when speaker-identity cues were provided before speech. We found that such speaker-identity cues significantly increased alpha oscillations in fronto-temporal cortex during post-cue pre-target time. Cortical evoked responses to target speech exhibited significantly greater amplitude in the cued condition, indicating speaker-identity cues enhance attentional processes. Grand-mean pupil dilation was larger in the cued condition, albeit the difference was not significant. The speaker-identity cues did not alter accuracy significantly, which guaranteed that our comparisons on pupil and EEG responses were not affected by the ratio of correct trials in across-trial averages. Combining these results, we claim that listening effort is not always an inherently bad, fatiguing process, but rather, includes top-down brain mechanisms that help listeners better attend to a target speech signal in background noise.