Related Experiment Video
Updated: Aug 25, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.7K
A brainnetome atlas-based methamphetamine dependence identification using neighborhood component analysis and machine
Yanan Zhou1,2, Jingsong Tang3, Yunkai Sun3
1Department of Psychiatry, National Clinical Research Center for Mental Disorders, The Second Xiangya Hospital of Central South University, Changsha, China.
Frontiers in Cellular Neuroscience
|October 14, 2022
Summary
Machine learning accurately identifies methamphetamine use disorder (MUD) using brain imaging. This approach reveals key brain connectivity changes, aiding in addiction diagnosis and treatment development.
Area of Science:
- Neuroscience
- Psychiatry
- Machine Learning
Background:
- Methamphetamine (MA) addiction poses a significant public health challenge.
- Identifying brain-based biomarkers for MA addiction is crucial for improving treatment outcomes.
- Resting-state functional magnetic resonance imaging (rs-fMRI) offers a non-invasive method to study brain alterations.
Purpose of the Study:
- To develop a predictive model for classifying individuals with methamphetamine use disorder (MUD) using rs-fMRI data.
- To identify brain-based features that can accurately predict MUD.
- To explore the utility of machine learning in identifying neuroimaging biomarkers for MUD.
Main Methods:
- Support vector machine (SVM) classification applied to rs-fMRI data from MUD and healthy control (HC) groups.
- Brain connectivity analysis using the Brainnetome atlas.
- Neighborhood Component Analysis (NCA) for feature selection, identifying 18 discriminative features.
Main Results:
- The SVM classifier achieved high accuracy (88.00%) in differentiating MUD from HCs, with excellent sensitivity (86.84%), specificity (89.19%), and AUC (0.94).
- Top features included alterations in the default mode network (DMN) and thalamic connections within the cortico-striato-thalamo-cortical (CSTC) loop.
- Functional connectivity (FC) between the inferior parietal lobule (IPL) and cingulate gyrus (CG) correlated with the duration of MA use.
Conclusions:
- MUD is associated with significant alterations in brain functional connectivity that are predictive of group membership.
- Machine learning techniques, particularly SVM, are effective tools for identifying neuroimaging biomarkers of MUD.
- These findings support the potential for using neuroimaging and machine learning in the clinical identification and management of methamphetamine addiction.
