How and when predictability interacts with accentuation in temporally selective attention during speech comprehension

Area of Science:

• Cognitive Neuroscience
• Psycholinguistics
• Speech Processing

Background:

• Speech comprehension involves integrating predictive information with acoustic signals.
• Understanding the temporal dynamics of attention in language processing is crucial.
• The interplay between top-down and bottom-up attention mechanisms remains an active research area.

Purpose of the Study:

• To investigate the interaction between top-down prediction and bottom-up acoustic signals in speech comprehension.
• To determine the timing of this interaction during language processing.
• To examine how attention allocation is modulated by predictability and accentuation.

Main Methods:

• Electroencephalography (EEG) was used to record brain activity.
• Mandarin Chinese spoken sentences with manipulated word predictability and accentuation were presented.
• A linguistic attention probe was employed to assess attentional focus.
• Event-related potentials (ERPs) like N1 and N400, and gamma-band power were analyzed.

Main Results:

• Attention probes and acoustic cues (accentuation, low predictability) modulated early attention (N1 latency).
• Prediction and accentuation showed a complementary interaction on attention allocation.
• Highly predictable words, especially when accented, showed reduced N400 and enhanced gamma-band activity.
• Early attention (N1) correlated with later semantic processing (gamma-band power).

Conclusions:

• Top-down and bottom-up attention interact early in speech comprehension, preceding lexical-semantic processing.
• This interaction influences attentional resource allocation and semantic integration.
• Findings contribute to models of language comprehension, highlighting the dynamic interplay of predictive and sensory information.