Emerging native-similar neural representations underlie non-native speech category learning success

Area of Science:

• Neuroscience
• Cognitive Science
• Linguistics

Background:

• Learning non-native speech sounds in adulthood is difficult, with varied success rates.
• The brain mechanisms behind these individual differences in learning efficacy are not well understood.

Purpose of the Study:

• To investigate if training makes adult learners' brain representations of non-native sounds resemble those of native speakers.
• To determine if this neural similarity predicts learning success in adults.

Main Methods:

• Used functional magnetic resonance imaging (fMRI) on English listeners learning Mandarin tones.
• Applied inter-subject neural representational similarity (IS-NRS) analysis and predictive modeling.
• Compared neural representations of learners with those of native Mandarin speakers.

Main Results:

• Learners developed neural representations similar to native speakers in speech perception areas after training.
• The degree of neural similarity significantly predicted learning speed and outcomes.
• Inter-subject neural representational similarity (IS-NRS) was a stronger predictor than other neural measures.
• Successful learning involved multidimensional, cost-efficient neural representations linked to feedback sensitivity in the frontostriatal network.

Conclusions:

• Experience-dependent neuroplasticity supports successful adult speech learning.
• Emergent native-like neural representations are key markers of learning success.
• Findings can inform the design of individualized, feedback-based training for efficient speech acquisition.