Seed correlation analysis based on brain region activation for ADHD diagnosis in a large-scale resting state data set

Area of Science:

• Neuroimaging
• Neuroscience
• Medical Diagnostics

Background:

• Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder often linked to brain functional connectivity disruptions.
• Resting-state functional MRI (rs-fMRI) shows promise as a diagnostic biomarker for ADHD.
• Investigating functional connectivity in resting-state brain networks is crucial for understanding ADHD.

Purpose of the Study:

• To develop an effective seed-correlation analysis procedure for identifying ADHD biomarkers.
• To investigate potential biomarkers within resting-state brain networks for ADHD diagnosis.
• To refine seed selection for improved brain interconnection analysis in ADHD research.

Main Methods:

• Analysis of rs-fMRI data from 149 children diagnosed with ADHD.
• Implementation of a two-step hierarchical analysis to extract functional connectivity features.
• Evaluation using linear classifiers and random sampling validation.

Main Results:

• A data-driven method (ReHo) identified four key brain regions for seed-correlation analysis.
• The proposed method achieved an 83.24% success rate in identifying ADHD patients.
• This approach outperformed traditional seed selection methods.

Conclusions:

• The study demonstrates the feasibility of diagnosing ADHD using rs-fMRI data analysis.
• Data-driven models provide a precise and reliable method for seed identification.
• Tailoring models to specific datasets may yield better diagnostic performance than generalized models.