Preattentive processing in the spatial unmasking of speech
Listeners benefit when target speech and masking speech are spatially separated. Research suggests that under complex listening conditions, when target and masker are perceptually confusable, spatial separation provides a cue that facilitates bottom-up segregation of target and masker, and top-down selective attention to the target. However, it has been difficult to distinguish between the contributions of bottom-up and top-down processes through behavioral measures alone. The present experiment used event-related potentials (ERPs) to examine the role of attention in the spatial unmasking of speech. Vocoded target words were presented with 2-talker vocoded masking babble when target and masker were 1) co-located, and 2) perceptually separated by adding an identical copy of the masker at a separate location that preceded the onset of the co-located masker by 4 ms. When listeners were attending to the sounds in a target detection task, auditory evoked potentials elicited by target words were only observed when target and masker were perceptually separated and masking was released, replicating previous findings (Zobel et al., 2016). Importantly, a similar modulation of ERPs was observed when listeners directed attention away from the auditory modality to engage in a challenging visual 2-back task. These findings provide strong evidence that under complex listening conditions, the perception of spatial separation between target and masker facilitates bottom-up, preattentive processing of target speech.