Sulcal Graph

Sulcal graphs represent the brain's cortical folding patterns as networks, enabling quantitative analysis of individual brain variations and population-level comparisons. Current research focuses on improving the accuracy and efficiency of sulcal graph identification and labeling using deep learning models, particularly 3D convolutional neural networks, incorporating techniques like contrastive learning, adversarial training, and diffusion models for pre-training and fine-tuning. These advancements aim to facilitate the discovery of biomarkers for neurological and psychiatric disorders by providing robust and explainable methods for analyzing brain structure, ultimately improving diagnostic capabilities and our understanding of brain-behavior relationships. The development of multi-graph matching techniques further enhances the ability to compare and analyze sulcal patterns across large populations.