Boost up the detection sensitivity of ASL perfusion fMRI through support vector machine

Area of Science:

• Neuroimaging
• Functional Magnetic Resonance Imaging (fMRI)
• Biomedical Data Analysis

Background:

• Arterial Spin Labeling (ASL) perfusion fMRI is crucial for measuring brain blood flow.
• ASL data inherently suffers from low signal-to-noise ratio (SNR), complicating group analysis.
• Existing methods may not fully leverage the information within ASL data for sensitive group-level inferences.

Purpose of the Study:

• To develop and evaluate a novel multivariate group analysis approach for ASL perfusion fMRI.
• To enhance the detection sensitivity of brain activation in ASL data.
• To provide a more robust method for group-level analysis of ASL perfusion fMRI.

Main Methods:

• A multivariate method based on group analysis was proposed.
• Support Vector Machine (SVM) learning was used to extract a spatial discriminance map (SDM) for each subject.
• Random Effect analysis (RFX) was applied to individual SDMs for population inference.

Main Results:

• The proposed method was evaluated using fingertapping ASL perfusion fMRI data from 7 subjects.
• Similar activation patterns were observed compared to standard General Linear Model (GLM) based group analysis.
• The multivariate approach demonstrated enhanced sensitivity in detecting activation.

Conclusions:

• The presented multivariate group analysis method effectively improves sensitivity for ASL perfusion fMRI.
• This approach offers a promising alternative for analyzing low SNR ASL data.
• The findings suggest potential for more reliable detection of brain activity in ASL studies.