Related Experiment Video
Updated: Dec 28, 2025
09:14
Exploring the Neural Correlates of Cognitive Reappraisal in Obsessive-Compulsive Disorder Using Task-based Functional Magnetic Resonance Imaging
Published on: March 14, 2025
808
Sub-graph entropy based network approaches for classifying adolescent obsessive-compulsive disorder from
Bhaskar Sen1, Gail A Bernstein2, Bryon A Mueller2
1Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis.
Neuroimage. Clinical
|February 18, 2020
Summary
This study introduces a new method using resting-state fMRI and information theory to accurately classify obsessive-compulsive disorder (OCD) in adolescents. The approach identifies brain network differences, aiding in diagnosis and understanding OCD biomarkers.
Area of Science:
- Neuroimaging
- Computational Psychiatry
- Biomarker Discovery
Background:
- Current obsessive-compulsive disorder (OCD) diagnosis in adolescents relies on clinical interviews, symptom scales, and behavioral observation.
- Functional magnetic resonance imaging (fMRI) offers potential for objective, network-based biomarkers to aid OCD diagnosis.
- Identifying reliable biomarkers from fMRI data is crucial for improving diagnostic accuracy in adolescent OCD.
Purpose of the Study:
- To investigate the clinical diagnostic utility of information-theoretic features from resting-state fMRI for classifying OCD in adolescents.
- To explore univariate, bivariate, and multivariate features, including sub-graph entropy, for identifying OCD-specific brain network patterns.
- To assess the performance of a novel approach using differential sub-graph (edge) entropy for classifying OCD versus healthy controls.
Main Methods:
- Resting-state fMRI data were acquired from 15 adolescents with OCD and 13 healthy controls.
- Time-series data from 85 brain regions were used to compute Shannon wavelet entropy, Pearson correlation matrices, network features, and sub-graph entropy.
- Leave-one-out cross-validation with in-fold feature selection was employed to evaluate classification accuracy, sensitivity, and specificity.
Main Results:
- An information-theoretic approach based on sub-graph entropy achieved high accuracy in classifying OCD versus healthy subjects.
- Differential sub-graph (edge) entropy demonstrated significant predictive power, yielding an accuracy of 0.89, specificity of 1, and sensitivity of 0.80.
- The study identified specific brain regions and network edges as important indicators for OCD in adolescents.
Conclusions:
- Information-theoretic analysis of resting-state fMRI data, particularly sub-graph and edge entropy, provides a robust method for classifying adolescent OCD.
- This novel approach shows potential as a complementary tool for clinicians in diagnosing OCD in adolescents.
- The findings highlight the predictive power of brain network metrics derived from fMRI for identifying OCD.
