ASD Classification from fMRI Connectomes

Mini-project comparing SVM (vectorized connectomes) and Graph Convolutional Networks (GCN) for classifying Autism Spectrum Disorder from resting-state fMRI.

Joint work with Alexander Vatamidis Norrstam. Site-aware cross-validation, graph-theoretic metrics, and PyTorch Geometric GCNs benchmarked against a classical ML pipeline.