Disparity level identification using the voxel-wise Gabor model of fMRI data

Area of Science:

• Neuroscience
• Visual Perception
• Computational Neuroscience

Background:

• Disparity perception is fundamental to understanding the physical world.
• While macaque electrophysiology and human fMRI studies map disparity-tuning neurons and visual areas, characterizing population responses to disparity using fMRI remains challenging.

Purpose of the Study:

• To develop and evaluate computational models for characterizing neural population responses to visual disparity using fMRI.
• To identify critical visual areas involved in representing disparity information.

Main Methods:

• Constructed three voxel-wise encoding Gabor models to predict voxel responses to varying disparity levels.
• Employed a decoding method to identify novel disparity levels from population cortical responses.
• Compared the performance of a fine-coarse model (FCM) against single and uncrossed-crossed models.

Main Results:

• The fine-coarse model (FCM) significantly outperformed the single and uncrossed-crossed models in fitting voxel responses to disparities.
• FCM demonstrated high accuracy in predicting voxel responses within the V3A complex.
• FCM achieved high accuracy in identifying novel disparity levels from population responses in the V3A complex.

Conclusions:

• The fine-coarse model (FCM) provides a superior method for characterizing voxel responses to disparities compared to alternative models.
• The V3A complex is identified as a critical visual area for the neural representation of disparity information.