Related Experiment Video
Updated: May 14, 2026
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
Classification of individuals based on sparse representation of brain cognitive patterns: a functional MRI study.
M Ramezani1, P Abolmaesumi, K Marble
1Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver, BC, Canada. mahdir@ece.ubc.ca
Summary
This study shows that sparse representation analysis with Fisher Linear Discriminant can classify individuals into young and older age groups using functional MRI data from a speech listening task. This method holds promise for diagnosing neurological disorders.
Area of Science:
- Neuroscience
- Machine Learning
- Medical Imaging
Background:
- Neurological disorders alter brain activity patterns detectable by functional imaging.
- Classifying individuals based on functional imaging data is challenging due to high dimensionality and limited sample sizes.
- Functional MRI (fMRI) data offers potential for diagnostic classification.
Purpose of the Study:
- To evaluate a Sparse Representation Analysis (SRA) method combined with Fisher Linear Discriminant (FLD) for classifying individuals based on fMRI data.
- To test the classification of individuals into young and older age groups using fMRI patterns during a speech listening task.
Main Methods:
- Employed Sparse Representation Analysis (SRA) using K-SVD to generate activation sources and sparse modulation profiles from fMRI data.
- Utilized Fisher Linear Discriminant (FLD) framework with these sparse modulation profiles for classification.
- Collected fMRI data from 32 adults (16 young, 16 older) performing a speech listening task under varying noise conditions.
Main Results:
- Successfully classified individuals into young and older age categories based solely on fMRI activation patterns.
- Demonstrated the feasibility of using SRA and FLD for age-based classification from fMRI data.
- Highlighted the potential of the proposed method for real-world diagnostic applications.
Conclusions:
- The combined SRA and FLD approach is a viable method for classifying individuals based on fMRI data.
- This technique shows promise for aiding in the diagnosis of neurological disorders by analyzing brain activity patterns.
- Functional activation patterns during cognitive tasks can effectively differentiate age groups.
