Auditory attention in a dynamic scene: Behavioral and electrophysiological correlates

Abstract

The ability to direct and redirect selective auditory attention varies substantially across individuals with normal hearing thresholds, even when sounds are clearly audible. We hypothesized that these differences can come from both differences in the spectrotemporal fidelity of subcortical sound representations and in the efficacy of cortical attentional networks that modulate neural representations of the auditory scene. Here, subjects were presented with an initial stream from straight ahead and a second stream (from either left or right), each comprised of four monotonized consonant-vowel syllables. Listeners were instructed to report the contents of either the first stream (holding attentional focus) or the second stream (switching attentional focus). Critically, the direction of the second stream informed subjects whether to hold or to switch attention. Pilot results suggest that when the lateral angle of the second stream is small, task performance is linked to subcortical encoding fidelity of suprathreshold sound (as measured using brainstem frequency-following responses obtained separately in the same subjects). Using a paradigm that allows simultaneous collection of behavioral measures, FFRs, and cortical responses, here we test whether differences in top-down attentional control explain subject variability when the second stream’s lateral angle is large and coding fidelity does not limit performance.

Pre- and post-target cortical processes predict speech-in-noise performance

“…Previous studies showed that such ERP components are strongly modulated by selective attention but only when auditory objects are successfully segregated (Choi et al, 2013;Choi et al, 2014;Kong et al, 2015). Since those early cortical ERP components originate from multiple regions across Heschl's gyrus (i.e., the primary auditory cortex) and its surrounding areas (e.g., posterior superior temporal gyrus) ( C eponien et al, 1998), an efficient and collective way of indexing the neural efficiency in speech unmasking is using a scalp electroencephalographical (EEG) potential at the vertex [e.g., "Cz" of the international 10-10 system for EEG electrode montage: Koessler et al (2009) within a limited time-window (e.g., 100-300 ms range after the stimulus onset].…”

“…Auditory scene analysis relies on the supra-threshold acoustic features that provide binding cues for auditory grouping (Darwin, 1997). These include the spectra (Lee et al, 2013), location (Frey et al, 2014;Goldberg et al, 2014), temporal coherence (Moore, 1990;Shamma et al, 2013;Teki et al, 2011), rhythm (Calderone et al, 2014;Golumbic et al, 2013;Herrmann et al, 2016;Obleser and Kayser, 2019), and timing (Lange, 2009) of the figure and ground. The fidelity of encoding such supra-threshold acoustic features may affect the separation of target speech from background noise.…”

“…Individual differences also exist in how strongly selective attention modulates cortical evoked responses to sounds (Choi et al, 2014). This suggests there may be a correlation between top-down selective attention efficacy and SiN performance.…”

Pre- and post-target cortical processes predict speech-in-noise performance

“…Auditory scene analysis relies on the supra-threshold acoustic features that provide binding cues for auditory grouping ( Darwin, 1997 ). These include the spectra ( Lee et al, 2013 ), location ( Frey et al, 2014 ; Goldberg et al, 2014 ), temporal coherence ( Moore, 1990 ; Shamma et al, 2013 ; Teki et al, 2011 ), rhythm ( Calderone et al, 2014 ; Golumbic et al, 2013 ; Herrmann et al, 2016 ; Obleser and Kayser, 2019 ), and timing ( Lange, 2009 ) of the figure and ground. The fidelity of encoding such supra-threshold acoustic features may affect the separation of target speech from background noise.…”