Lexical Information Guides Retuning of Neural Patterns in Perceptual Learning for Speech.

Area of Science:

• Cognitive Neuroscience
• Auditory Perception
• Speech Processing

Background:

• Speech sound interpretation can vary moment-to-moment.
• Prior experience, or perceptual learning, influences speech interpretation.
• Lexical context plays a role in guiding perceptual learning for speech.

Purpose of the Study:

• To investigate how neural patterns reflect perceptual learning in speech.
• To examine the brain mechanisms underlying lexically guided perceptual learning.
• To determine if neural activity recalibrates to perceived speech categories.

Main Methods:

• Utilized multivoxel pattern analysis (MVPA) on archival fMRI data.
• Trained a classifier to distinguish neural patterns for unambiguous /s/ and /∫/ sounds.
• Tested classifier generalization to ambiguous speech tokens based on individual perception.

Main Results:

• A trained classifier successfully generalized to ambiguous speech tokens.
• Generalization was based on individual participants' trial-by-trial perception.
• Left parietal and right temporal regions were most informative for phonetic categorization.

Conclusions:

• Perceptual learning in speech involves neural recalibration.
• Brain activation patterns adapt to approximate the perceived speech category.
• Findings enhance understanding of neural encoding of speech perception variability.