Linguistic processing of task-irrelevant speech at a cocktail party

Abstract

Paying attention to one speaker in noisy environments can be extremely difficult, because to-be-attended and task-irrelevant speech compete for processing resources. We tested whether this competition is restricted to acoustic-phonetic interference or if it extends to competition for linguistic processing as well. Neural activity was recorded using Magnetoencephalography as human participants were instructed to attended to natural speech presented to one ear, and task-irrelevant stimuli were presented to the other. Task-irrelevant stimuli consisted either of random sequences of syllables, or syllables structured to form coherent sentences, using hierarchical frequency-tagging. We find that the phrasal structure of structured task-irrelevant stimuli was represented in the neural response in left inferior frontal and posterior parietal regions, indicating that selective attention does not fully eliminate linguistic processing of task-irrelevant speech. Additionally, neural tracking of to-be-attended speech in left inferior frontal regions was enhanced when competing with structured task-irrelevant stimuli, suggesting inherent competition between them for linguistic processing.

Do we parse the background into separate streams in the cocktail party?

“…Other studies also confirmed the recognition of some words from a background speech stream, even if the background consisted of multiple voices (Dekerle et al, 2014). Furthermore, signs of spectro-temporal and linguistic processing of task-irrelevant speech streams were found in the auditory cortex, left inferior cortex, and posterior parietal cortex (Brodbeck et al, 2020; Har-shai Yahav & Zion Golumbic, 2021). The prerequisite of background stream segregation might be highly distinctive features, which results in categorical differences, such as different gender of speakers; but this background stream segregation might be unique to speech perception.…”

Capacity and tradeoffs in neural encoding of concurrent speech during Selective and Distributed Attention

“…In interpreting the results of our study, it is important to bear in mind that the neural metric of speech tracking used here (envelope-following response) primarily captures the acoustic representations of input in auditory cortex (Ding and Simon 2012a; Mesgarani and Chang 2012; Zion Golumbic et al 2013; Fiedler et al 2019a; Har-shai Yahav and Zion Golumbic 2021). Therefore, the current results should not be taken as an exhaustive description of all the ways in which attention may affect neural processing of speech.…”

“…This ‘ground-truth’ problem is extremely difficult to address empirically, since we lack a reliable readout of an individual’s internal allocation of attention. Unfortunately, most behavioural and neural measures used in in Selective Attention paradigms cannot provide sufficient temporal resolution to indicate the momentary locus of attention and/or how well task-irrelevant speech has been processed (Har-shai Yahav and Zion Golumbic 2021). While some attempts have been made at utilizing advanced analytical approaches to achieve real-time measures of attention (Miran et al 2018; Jaeger et al 2020), these studies are still limited by the lack of access to the true internal state of attention.…”

“…And yet, these modulatory top-down effects do not eliminate the internal representation of task-irrelevant (or ‘to-be-ignored’) speech. Not only is this speech still encoded in auditory cortex (Ding and Simon 2012a; Horton and Srinivasan 2013; Zion Golumbic et al 2013; O’Sullivan et al 2015; Fiedler et al 2019), but a multitude of behavioural and neural findings provide evidence that some linguistic processing is also applied to task-irrelevant speech (Tun et al 2002; Dupoux et al 2003; Beaman et al 2007; Rivenez et al 2008; Carey et al 2014; Parmentier et al 2014, 2018; Schepman et al 2015; Vachon et al 2019; Dai et al 2021; Har-shai Yahav and Zion Golumbic, 2021). Therefore, alongside the convincing evidence that selective-attention biases neural processing such that to-be-attended speech is preferentially encoded, we still do not have a full understanding of the system’s capacity and limitations for processing additional concurrent speech.…”

“…This 'ground-truth' problem is extremely difficult to address empirically, since we lack a reliable readout of an individual's internal allocation of attention. Unfortunately, most behavioural and neural measures used in in Selective Attention paradigms cannot provide sufficient temporal resolution to indicate the momentary locus of attention and/or how well task-irrelevant speech has been processed (Har-shai Yahav and Zion Golumbic 2021).…”