LUMINA: Laplacian-Unifying Mechanism for Interpretable Neurodevelopmental Analysis via Quad-Stream GCN

Researchers developed LUMINA, a new Graph Convolutional Network architecture that improves AI-driven diagnosis of neurodevelopmental disorders using fMRI brain data. The system achieved 84.66% accuracy for ADHD and 88.41% for autism spectrum disorder detection by addressing traditional GCN limitations in capturing neural connection dynamics.

• →LUMINA uses a Quad-Stream GCN with bipolar RELU activation to overcome traditional GCN smoothing limitations that blur important neurological contrasts.
• →The system demonstrated superior diagnostic performance with 84.66% accuracy for ADHD and 88.41% for autism spectrum disorder on standard datasets.
• →The dual-spectrum graph Laplacian filtering mechanism preserves diverse neural connection characteristics that conventional GCN models tend to lose.
• →Testing was conducted through 5-fold cross validation on ADHD200 and ABIDE datasets with 144 and 579 subjects respectively.
• →The research addresses critical neurodevelopmental disorders in childhood by improving AI-driven fMRI brain data analysis.