Categorical learning revealed in activity pattern of left fusiform cortex

Area of Science:

• Neuroscience
• Cognitive Science
• Brain Imaging

Background:

• The brain encodes object categories, but how learning shapes these neural representations is unclear.
• Understanding how we learn to categorize complex stimuli like faces is crucial for cognitive neuroscience.

Purpose of the Study:

• To investigate whether enhanced categorical discrimination or feature analysis drives face/non-face categorization learning.
• To examine how neural representations of object categories change with learning in the human brain.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to measure brain responses before and after training.
• Participants underwent training with either a categorization task or a feature analysis task.
• Stimuli included gray-scale images for training and two-toned Mooney images for testing, with a defined face/non-face boundary.

Main Results:

• Categorical training induced a shift in the left fusiform gyrus from graded to categorical neural representations.
• This shift correlated with improved face/non-face discrimination, including increased face selectivity and reduced false alarms.
• The right fusiform cortex's activity correlated with categorization pre-training but was unaffected by learning.

Conclusions:

• The left fusiform cortex plays a key role in learning to categorize faces.
• Learning-induced changes in neural representation are critical for accurate categorization.
• Findings suggest a re-evaluation of hemispheric roles in face categorization.