Gyral‐Sulcal Net: A Novel Brain Network Representation for Mild Cognitive Impairment Classification

We proposed a novel brain Gyral-Sulcal Net for MCI/CN classification. The Gyral-Sulcal Net integrates the brain's two primary folding patterns, gyri and sulci, into a unified finer-scale anatomical network. It effectively combines structural and functional information as the node features. Based on the Gyral-Sulcal Net, we developed a novel graph transformer network to leverage Gyral-Sulcal Net for MCI/CN classification. Our approach achieved promising results, highlighting the great potential of the Gyral-Sulcal Net in improving diagnostic accuracy.