Multivariate pattern analysis of fMRI data: What can it reveal about visual representation?
Speaker: James Haxby , Dartmouth
Date: April 25 2008
Time: 2:00PM to 3:00PM
Location: 32-D507
Host: Polina Golland, CSAIL
Contact: Polina Golland, x38005, polina@csail.mit.edu
Relevant URL: Functional brain imaging has revealed a complex, macroscopic
organization in the functional architecture of the ventral object
vision pathway. Numerous studies have found regions of ventral
temporal cortex that consistently demonstrate category-related
response preferences, most notably a region that responds maximally
during face perception, the fusiform face area (FFA). Faces and
numerous other object categories, however, also evoke distinct
patterns of response across wider expanses of ventral temporal
cortex, including distinct patterns of response in cortical regions
that respond submaximally to the category being viewed, suggesting
that the representations of faces and other objects extend beyond the
regions defined by category preference. Multivariate pattern
analysis (MVPA) uses machine learning pattern classifiers to analyze
fMRI data. Whereas conventional univariate methods were designed to
find clusters of voxels with similar response properties, MVPA is
designed to detect reliable patterns of differences among voxel
responses that have a higher spatial frequency than category-
selective regions. Whereas conventional univariate analysis of fMRI
data characterizes local activity on a one-dimensional scale of
magnitude, MVPA analyzes local activity as a multidimensional
vector. Consequently, MVPA detects patterns that can be related to
the underlying population encoding for stimuli. Moreover, the
similarity of neural responses to different stimuli or experimental
conditions can be analyzed in a high dimensional space for neural
representation. The similarity structure of neural representations
varies by region. The similarity of patterns of response to visual
stimuli is correlated with psychological similarity, suggesting that
these methods now allow us to use fMRI to investigate neural
representations of visual stimuli as multidimensional and relate them
to multidimensional spaces for representing their physical and
psychological properties.
See other events that are part of Biomedical Imaging and Analysis 2007/2008
See other events happening in April 2008
