Characterization and Classification of ADHD Subtypes: An Approach Based on the Nodal Distribution of Eigenvector...

Area of Science:

• Neuroscience
• Complex Network Theory
• Computational Psychiatry

Background:

• Complex network theory is increasingly used for diagnosing neuropathological conditions like ADHD.
• Current ADHD diagnosis relies on functional MRI, which can be biased or computationally expensive.
• There's a need for deterministic, computationally efficient methods for ADHD subtype identification.

Purpose of the Study:

• To propose a deterministic method for identifying and differentiating common ADHD subtypes.
• To utilize eigenvector centrality as a key complexity measure for ADHD diagnosis.
• To develop a computationally efficient approach for ADHD subtype classification.

Main Methods:

• Applied complex network theory and eigenvector centrality to functional MRI data.
• Utilized a classification tree model to analyze node-wise centrality differences.
• Identified marker nodes across default mode, visual, frontoparietal, limbic, and cerebellar networks.

Main Results:

• Developed a deterministic method for ADHD subtype identification based on eigenvector centrality.
• Node-wise centrality differences effectively diagnosed ADHD subtypes (p < 0.05).
• Identified specific brain networks involved in ADHD neuropathology.

Conclusions:

• Eigenvector centrality provides a deterministic and efficient measure for ADHD subtype diagnosis.
• The proposed method offers a less biased alternative to current diagnostic protocols.
• Findings highlight the involvement of multiple brain regions in ADHD.