Related Experiment Video
Updated: Apr 27, 2026
17:06
Co-analysis of Brain Structure and Function using fMRI and Diffusion-weighted Imaging
Published on: November 8, 2012
26.0K
Discriminative analysis of multivariate features from structural MRI and diffusion tensor images.
Muwei Li1, Yuanyuan Qin2, Fei Gao3
1College of Electronics and Information Engineering, Sichuan University, Chengdu 610064, China.
Magnetic Resonance Imaging
|June 28, 2014
Summary
Combining diffusion tensor imaging (DTI) and T1-based MRI markers improves Alzheimer's disease (AD) detection. This approach accurately differentiates AD patients from healthy individuals using machine learning, offering a promising tool for early diagnosis.
Area of Science:
- Neuroimaging
- Machine Learning
- Neurology
Background:
- Magnetic resonance imaging (MRI) and machine learning aid in identifying anatomical patterns for Alzheimer's disease (AD) differentiation.
- T1-based imaging markers, particularly gray matter volumes, show discriminative potential but overlook white matter abnormalities.
- Diffusion tensor imaging (DTI) has explored white matter but lacks systematic studies on region-of-interest (ROI)-based features or combined T1 and DTI analysis.
Purpose of the Study:
- To evaluate the discriminative power of tract-based DTI features for Alzheimer's disease.
- To assess the feasibility of combining DTI-derived features with T1-based gray matter volumes for enhanced AD classification.
- To investigate the diagnostic accuracy of a combined feature set using machine learning.
Main Methods:
- Tract-based fractional anisotropy (FA) maps from DTI were analyzed.
- Support vector machine (SVM) was employed to evaluate discriminative capability.
- A combined feature set including tract-based FA and gray matter volumes was assessed via cross-validation.
Main Results:
- Tract-based FA maps showed discriminative capability for Alzheimer's disease.
- The combined feature set (tract-based FA + gray matter volumes) achieved high classification performance.
- Classification yielded 94.3% accuracy, 95.0% sensitivity, 93.3% specificity, and an AUC of 0.96.
Conclusions:
- Tract-based FA features are feasible for recognizing anatomical changes in Alzheimer's disease.
- Combined tract-based FA and gray matter volume features offer a robust method for AD classification.
- This integrated neuroimaging approach can complement existing classification methods for Alzheimer's disease.
