Reduced functional connectivity supports statistical learning of temporally distributed regularities

Area of Science:

• Cognitive Neuroscience
• Neuroimaging

Background:

• Statistical learning is crucial for adapting to environmental regularities and predicting future events.
• Understanding the neural architecture supporting statistical learning is key to cognitive function.

Purpose of the Study:

• To investigate the whole-brain functional connectivity (FC) supporting statistical learning.
• To characterize the brain network architecture underlying the extraction of temporal regularities.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to measure brain activity during a statistical learning task.
• Behavioral learning scores were correlated with functional connectivity changes.
• Group independent component analysis and seed-based FC analyses were employed.

Main Results:

• Learning performance correlated with activation in the lateral occipital cortex, angular gyrus, precuneus, anterior cingulate cortex, and superior frontal gyrus.
• The superior frontal network showed the strongest correlation with statistical learning performance.
• Reduced FC was observed between the superior frontal gyrus and salience, language, and dorsal attention networks.

Conclusions:

• Weakened functional connections between the superior frontal gyrus and top-down control networks are pivotal for statistical learning.
• This reduced connectivity facilitates the processing of novel information and pattern extraction from the environment.