Machine and Deep Learning Approaches for Migraine Diagnosis and Classification: Advances and Challenges

Abstract

Migraine is a complex neurological disorder characterised by recurrent headaches, sensory disturbances, and autonomic dysfunction. Despite its high prevalence, accurate diagnosis and effective management remain challenging due to the variability of symptoms and the absence of reliable biomarkers. In recent years, approaches based on artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), have offered new perspectives for improving diagnostic accuracy and disease classification. ML models such as support vector machines and random forests have shown potential in identifying neuroimaging and electrophysiological biomarkers, while convolutional neural networks (CNNs) are capable of extracting meaningful patterns from electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) data. Nevertheless, most studies are limited by small, homogeneous datasets and the lack of external validation, which restricts reproducibility and generalisability. Beyond technical performance, the transition of AI tools from research to clinical settings remains constrained by regulatory issues, limited interpretability, and difficulties in integration with existing healthcare systems. Future studies should focus on developing explainable and multimodal frameworks, supported by cross-institutional collaborations and prospective validation, to ensure both methodological robustness and clinical credibility. In the long term, the integration of such approaches may support earlier and more precise identification of migraine subtypes, guide individualised treatment strategies, and contribute to better patient outcomes.