Competing speech streams are simultaneously represented in the human cortex during attention switching

Area of Science:

• Neuroscience
• Cognitive Science
• Auditory Perception

Background:

• Effective communication in noisy environments relies on sustained and switched attention.
• Neural mechanisms of sustained attention are well-studied, but attention switching remains unclear.

Purpose of the Study:

• To investigate the neural dynamics of auditory attention switching in multi-talker scenarios.
• To understand how the brain disengages from one speech stream and engages with another.
• To explore the role of lexical context updating after attention shifts.

Main Methods:

• Electroencephalography (EEG) recordings from normal-hearing adults in a simulated multi-talker environment.
• Measuring neural tracking of speech streams using Temporal Response Functions (TRF).
• Analyzing EEG alpha power and cortical activity related to lexical prediction.

Main Results:

• Neural tracking of the new speech stream began before disengaging from the old one, indicating transient simultaneous encoding.
• A reduction in EEG alpha power correlated with cognitive effort during attention switches.
• Lexical context updating strategies were modeled using Large Language Models to analyze post-switch processing.

Conclusions:

• Auditory attention switching involves asymmetric engagement and disengagement processes.
• Listeners exhibit a temporary capacity to process two competing speech streams concurrently.
• A reset of lexical context may occur after attention shifts, facilitating flexible speech processing in complex auditory scenes.