Exploiting the brain's network structure in identifying ADHD subjects.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Medical Imaging

Background:

• Attention Deficit Hyperactive Disorder (ADHD) is a prevalent behavioral disorder in children.
• Resting-state functional magnetic resonance imaging (fMRI) offers insights into brain function.
• Brain activity can be modeled as a functional network to identify neurological differences.

Purpose of the Study:

• To develop an automated method for classifying ADHD subjects using fMRI data.
• To investigate differences in brain functional networks between ADHD and control subjects.
• To enhance classification accuracy by focusing on specific brain regions.

Main Methods:

• Functional brain networks were constructed by analyzing voxel-wise correlations in resting-state fMRI data.
• A Principal Component Analysis-Linear Discriminant Analysis (PCA-LDA) classifier was trained using network features.
• A novel brain masking technique was developed to isolate relevant brain regions for feature extraction.
• Graph-motif features, specifically 3-cycle participation maps, were utilized.

Main Results:

• The study successfully modeled brain function as a network, revealing distinct properties in ADHD subjects.
• The proposed brain masking method significantly improved classification accuracy.
• The best classification performance of 69.59% was achieved using 3-cycle map features with the masking technique.
• The classifier was trained on 776 subjects and tested on 171 subjects from the ADHD-200 challenge.

Conclusions:

• The findings demonstrate the potential of using network properties from fMRI data for ADHD classification.
• The developed masking approach enhances diagnostic accuracy by focusing on key brain regions.
• This method shows promise for improving the understanding and diagnosis of ADHD.