Related Experiment Video
Updated: Jan 12, 2026
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
Abnormal white matter microstructure in tobacco use disorder: A machine learning study based on whole-brain
Lei Jiang1, Yan Kang1, Xu Han2
1State Key Laboratory of Magnetic Resonance Spectroscopy and Imaging, National Center for Magnetic Resonance in Wuhan, Wuhan Institute of Physics and Mathematics, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences, Wuhan 430071, China; University of Chinese Academy of Sciences, Beijing 100049, China.
This study used machine learning to identify brain white matter (WM) differences in individuals with tobacco use disorder (TUD). The approach accurately distinguished TUD participants and found correlations between WM alterations and smoking onset age.
Area of Science:
- Neuroimaging
- Neuroscience
- Machine Learning
Background:
- Cigarette smoking is linked to abnormal white matter (WM) microstructure.
- Traditional analysis methods may produce false positives/negatives in identifying smoking-related WM changes.
- Multivariate machine learning offers enhanced identification of smoking-related WM regions.
Purpose of the Study:
- To apply a multivariate machine learning method for improved identification of smoking-related WM alterations.
- To differentiate individuals with tobacco use disorder (TUD) from non-TUD individuals using diffusion metrics.
- To explore the relationship between brain structure changes and smoking behavior.
Main Methods:
- Utilized whole-brain skeletonized diffusion MRI metrics from 60 TUD and 66 non-TUD participants.
- Trained a linear support vector machine (SVM) classifier to distinguish between groups.
- Determined smoking-related WM regions via discriminative scores and correlated scores with smoking variables.
Main Results:
- The SVM classifier achieved >0.80 accuracy and >0.91 AUC.
- Maximal discriminative weights were found in the anterior corona radiata, posterior thalamic radiation, and genu of the corpus callosum.
- Classification scores positively correlated with earlier age of cigarette use onset.
Conclusions:
- The multivariate approach using whole-brain diffusion metrics outperformed regional analyses.
- Identified WM regions effectively distinguish individuals with nicotine addiction.
- This method advances understanding of smoking-related brain alterations and nicotine addiction.
