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.

When faced with situations where many people talk at once, individuals can employ different listening strategies to deal with the cacophony of speech sounds and to achieve different goals. In this fMRI study, we investigated how the pattern of neural activity is affected by the type of attention applied to speech in a simulated “cocktail party.” Specifically, we compared brain activation patterns when listeners “attended selectively” to only one speaker and ignored all others, versus when they “distributed their attention” and followed several concurrent speakers. Conjunction analysis revealed a highly overlapping network of regions activated for both types of attention, including auditory association cortex (bilateral STG/STS) and frontoparietal regions related to speech processing and attention (bilateral IFG/insula, right MFG, left IPS). Activity within nodes of this network, though, was modulated by the type of attention required as well as the number of competing speakers. Auditory and speech-processing regions exhibited higher activity during distributed attention, whereas frontoparietal regions were activated more strongly during selective attention. These results suggest a common “attention to speech” network, which provides the computational infrastructure to deal effectively with multi-speaker input, but with sufficient flexibility to implement different prioritization strategies and to adapt to different listener goals.

Managing attention in multispeaker environments is a challenging feat that is critical for human performance. However, why some people are better than others in allocating attention appropriately remains highly unknown. Here, we investigated the contribution of two factors-working memory capacity (WMC) and professional experience-to performance on two different types of attention task: selective attention to one speaker and distributed attention among multiple concurrent speakers. We compared performance across three groups: individuals with low (n = 20) and high (n = 25) WMC, and aircraft pilots (n = 24), whose profession poses extremely high demands for both selective and distributed attention to speech. Results suggests that selective attention is highly effective, with good performance maintained under increasingly adverse conditions, whereas performance decreases substantially with the requirement to distribute attention among a larger number of speakers. Importantly, both types of attention benefit from higher WMC, suggesting reliance on some common capacity-limited resources. However, only selective attention was further improved in the pilots, pointing to its flexible and trainable nature, whereas distributed attention seems to suffer from more fixed and severe processing bottlenecks.

Paying attention to one speaker in noisy environments can be extremely difficult. This is because task-irrelevant speech competes for processing resources with attended speech. However, whether this competition is restricted to acoustic-phonetic interference, or if it extends to competition for linguistic processing as well, remains highly debated. To address this debate, here we test whether task-irrelevant speech sounds are integrated over time to form hierarchical representations of lexical and syntactic structures.Neural activity was recorded using Magnetoencephalography (MEG) during a dichotic listening task, where human participants 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 (Non-Structured), or syllables ordered to form coherent sentences (Structured). Using hierarchical frequency-tagging, the neural signature of different linguistic-hierarchies within the Structured stimuli – namely words, phrases and sentences – can be uniquely discerned from the neural response.We find that, indeed, the phrasal structure of task-irrelevant stimuli was represented in the neural response, primarily in left inferior frontal and posterior parietal regions. Moreover, neural tracking of attended speech in left inferior frontal regions was enhanced when task-irrelevant stimuli were linguistically structured. This pattern suggests that syntactic structurebuilding processes are applied to task-irrelevant speech, at least under these circumstances, and that selective attention does not fully eliminate linguistic processing of task-irrelevant speech. Rather, the inherent competition for linguistic processing resources between the two streams likely results in the increased listening effort experienced when trying to focus selective attention in multi-speaker contexts.Significance statementThis study addresses the fundamental question of how the brain deals with competing speech in noisy environments. Specifically, we ask: when one attempts to focus their attention on a particular speaker, what level of linguistic processing is applied to other, task-irrelevant speech? By measuring neural activity, we find evidence that the phrasal structure of task-irrelevant speech is indeed discerned, indicating that linguistic information is integrated over time and undergoes some syntactic analysis. Moreover, neural responses to attended speech were also enhanced in speech-processing regions, when presented together with comprehensible yet task-irrelevant speech. These results nicely demonstrate the inherent competition for linguistic processing resources among concurrent speech, providing evidence that selective attention does not fully eliminate linguistic processing of task-irrelevant speech.

No abstract