PSYC 4600 Lecture Notes - Lecture 16: Extrapolation, Voxel, Brodmann Area 20
23 Dec 2016
School
Department
Course
Professor
Document Summary
Uses pattern analysis and classification to get a more fine brain understanding of brain activity. Lecture 16: classification as a more individualized one. It requires multiple examples to determine that certain pixel patterns fit within the dimensions of that person. Multi-voxel pattern analysis (mvpa: pattern classification in the brain fmri volumes are simply 3d images of brain activity. Can we predict the type of image (face, house, cat, show, chair, etc. ) a subject is currently viewing based on the pattern of activation across ventral temporal cortex: study: haxby et al. Present each type of image multiple times and record average response in each voxel in ventral temporal cortex. Repeat first half of experiment with new stimuli and calculate new set of patterns. Within-category correlations were higher than between-category correlations: even when excluding maximally-responsive voxels. Face pattern in first half strongly predictive of face pattern in second half: even when excluding voxels from ffa.
