Learning nonnative speech sounds changes local encoding in the adult human cortex.

Area of Science:

• Neuroscience
• Auditory Perception
• Speech Processing

Background:

• Adults can learn to distinguish nonnative speech sounds, but learning varies.
• Behavioral accuracy improvements correlate with enhanced sound representation separability in cortical speech areas.
• It is unknown if auditory neural populations exhibit uniform or heterogeneous changes during learning.

Purpose of the Study:

• To investigate local population response patterns during nonnative speech sound learning.
• To examine how neural encoding of unfamiliar vocal pitch patterns changes with learning.
• To determine if learning effects are linked to accuracy-dependent or exposure-dependent neural changes.

Main Methods:

• Utilized high-resolution direct neural recordings in human listeners.
• Trained native English speakers to recognize Mandarin Chinese tones.
• Analyzed neural population responses in relation to trial-by-trial accuracy and stimulus exposure.

Main Results:

• Identified distributed neural populations in the superior temporal gyrus and ventrolateral frontal cortex exhibiting learning effects.
• Observed both increases and decreases in tone separability, indicating heterogeneous encoding changes.
• Found that learning effects were partly driven by increased neural response variability and could be predicted by pre-training speech-evoked activity.

Conclusions:

• Nonnative speech sound learning involves widespread, heterogeneous changes in neural representations across distributed brain regions.
• Learning effects are accuracy-dependent and distinct from general exposure effects.
• Intrinsic neural properties may predispose specific populations to behavior-related plasticity.