Neural Correlates of Category Learning in Monkey Inferior Temporal Cortex

Area of Science:

• Neuroscience
• Visual Cognition
• Machine Learning

Background:

• Area TE is crucial for learning visual categories based on perceptual similarity.
• Understanding how neural activity changes during category learning is essential.

Purpose of the Study:

• To investigate whether category learning alters neural activity in area TE.
• To examine changes in neuronal selectivity and network recruitment after training.

Main Methods:

• Two male monkeys were trained to categorize natural images of cats and dogs.
• Multielectrode arrays were used to record neural activity in area TE before and after training.
• Neural activity was compared during a passive viewing task.

Main Results:

• Category training improved the accuracy of abstract category decoding.
• Individual neurons (single units) in area TE showed increased category selectivity.
• The proportion of category-selective neurons and the duration of their sustained responses increased.

Conclusions:

• Visual category learning enhances category separability in area TE.
• This enhancement is driven by changes in individual neuron stimulus selectivity and increased network recruitment.