What is multivariate analysis in fMRI?
Analysis that uses patterns of activity across multiple voxels to distinguish experimental conditions.
What does MVPA examine
Patterns of neural responses, rather than analyzing single voxel- or region-based values, as is customary in conventional univariate analyses
Describe differences between multivoxel pattern analysis and univariate analysis
Example of a univariate approach
Examining changes in average or peak neural responses across conditions of a study:
e.g. amygdala shows greater activation to fear-inducing than neutral stimuli.
Why are univariate approaches called univariate
Because they only consider one value per condition (e.g. the average signal of a region or voxel) at a time.
Key distinction between univiariate and MVPA
Instead of looking at each voxel separately, or averaging across voxel’s varying signals within a region, MVPA looks for information in the patterns of neural responses across voxels.
Two most widely used varieties of MVPA
- decoding analyses
- representational similarity analysis
Goal of decoding analyses
try to identify what condition elicited a given neural
response.
What do decoding analyses entail
The main method used in decoding analyses
classification
What does classification do in decoding analyses
attempts to predict (i.e. classify) which categories correspond to which
observations—e.g. was a given neural response elicited by
an angry or surprised face?
What does regression do in decoding analyses
treats data as continuous—e.g. how angry was the face that elicited a given
neural response?
What does RSA examine
the relative similarity of the patterns across stimuli; considers the relations among neural response patterns, comparing them to the relations among a stimulus property.
What does a representational dissimilarity matrix show
how similar the neural response patterns elicited by each condition are to each other
What happens in a searchlight analysis
This is a type of MVPA analysis:
a sphere is defined around every voxel, and the pattern of responses in this sphere is
strung out in a vector for each condition.
Benefits of MVPA in comparison to measuring overall response magnitudes
Sensitive to information carried in distributed response patterns: That is, even if a single voxel does not significantly change across conditions when considered
on its own, the signal variability of this voxel may still
contribute to a reliable response pattern that does discriminate between conditions. Univariate blurs this but MVPA picks this up.
What does multivariate pattern analysis use?
Multivariate analysis uses the pattern of activity across multiple voxels to discriminate between conditions.
Instead of treating each voxel independently (univariate), MVPA treats many voxels as a single multivariate feature vector.
What question does MVPA answer?
“Does this region contain information that discriminates conditions?”
(not “Is it more active?”)
What is the main advantage of MVPA?
Higher sensitivity to distributed coding — allows detection of representational differences invisible to univariate analysis.
What is decoding/classification MVPA
Using voxel patterns to train a classifier to distinguish conditions and testing performance on held-out data.
What is Representational Similarity Analysis (RSA)?
A method that compares similarity/dissimilarity of neural patterns across conditions, producing an Representational Dissimilarity Matrix.
Why is MVPA useful at the single-subject level?
Because pattern structure is within-brain, MVPA can detect information without needing group averaging.
What question does classification ask?
How accurately can I predict which condition produced this neural pattern?
What are the key features of classification?
Uses voxel patterns as features
Uses labels (condition A, B, C…) during training
Trains a classifier (e.g., SVM, logistic regression)
Tests on held-out data
Output = accuracy (above chance = information present)
