Researchers from the Graduate School, Universitas Padjadjaran, collaborated with national and international institutions to apply cutting-edge machine learning and big data analytics in studying Attention-Deficit/Hyperactivity Disorder (ADHD). Analyzing 5,937 brain voxels sourced from neuroimaging data, the team employed advanced feature selection tools—such as Boruta, Random Forest combined with DALEX explainability, and deep neural networks—as well as dimensionality reduction and clustering (PCA, KMeans, MCLUST). Their work identified critical brain regions associated with ADHD, including the Fusiform Gyrus, Thalamus, and Superior Temporal Gyrus, and demonstrated that deep neural networks using ReLU activation achieved superior classification accuracy across various performance metrics.

This breakthrough aligns with several key Sustainable Development Goals (SDGs):

SDG 3 (Good Health and Well-Being) — by advancing more precise, data-driven methods for ADHD diagnosis, potentially improving early intervention and patient outcomes.

SDG 9 (Industry, Innovation, and Infrastructure) — by demonstrating the value of AI-powered analytics in neuropsychiatric research.

SDG 4 (Quality Education) — through informing educators and clinicians with enhanced diagnostic insight for supporting students affected by ADHD.