The neural correlates of statistical learning in a word segmentation task: An fMRI study

Area of Science:

• Neuroscience
• Cognitive Psychology
• Speech Processing

Background:

• Statistical learning is a key mechanism for acquiring complex information, including language.
• Understanding the neural basis of statistical learning in speech is essential for explaining language acquisition.
• Previous research has implicated various brain regions in sequence learning, but specific roles in speech segmentation remain unclear.

Purpose of the Study:

• To investigate the neural correlates of statistical learning in continuous speech segmentation.
• To identify brain areas associated with the ability to discriminate coherent speech sequences from less coherent ones.
• To explore the relationship between individual learning rates and neural activation patterns in the brain.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was employed to measure brain activity.
• Participants listened to continuous speech streams with varying syllable transitional probabilities.
• Behavioral tests assessed the discrimination of statistically coherent (words) and less coherent (partwords) sequences over four runs.

Main Results:

• Behavioral data confirmed participants' ability to distinguish words from partwords based on statistical regularities.
• Significant neural activation was observed in the pars opercularis and pars triangularis of the left inferior frontal gyrus (LIFG).
• Individual learning rates positively correlated with activation in these LIFG regions.

Conclusions:

• The left inferior frontal gyrus (LIFG) plays a critical role in the statistical learning of speech segmentation.
• Findings support the involvement of a frontal/subcortical network in implicit sequence learning, extending to auditory domains.
• This study provides insights into the neural mechanisms underlying how the brain extracts meaningful units from continuous speech input.