AI-Based Early Detection of Migraines Using a Fusion Machine Learning Model

Migraine is a neurological disorder characterized by unpredictable and often severe attacks that can seriously disrupt daily life. Although some people experience an aura as an early warning, others miss the opportunity to get timely treatment. This research investigates the potential for migraine detection from Electroencephalogram (EEG) signals recorded from temporal lobe electrodes. A novel AI-based dual-branch fusion model has been proposed that leverages a Multilayer Perceptron (MLP) trained on engineered EEG features alongside a Convolutional Neural Network (CNN) trained on spectrogram representations of the same signals. This hybrid model takes advantage of both structured and unstructured data modalities to improve predictive accuracy. The experimental findings indicate that the fusion model offers an accuracy of 95%, which makes it a potential candidate for biomedical applications. The proposed approach lay a foundation for future neurotechnology that is capable of identifying EEG patterns in migraine and other related neurological conditions, contributing to the advancement of AIdriven diagnostic tools in smart healthcare.