Explainable Disentangled Representation Learning of Recurring Brain Activation Patterns via Variational Autoencoders

Doctoral-proposal contribution to the XAI World Conference 2025 Doctoral Proposals track, framing the EEG microstate research programme as an explainable representation-learning problem. The proposal argues that disentangled latent factors learnt by a variational autoencoder over scalp topographic maps can recover canonical microstate classes while keeping the representation interpretable and probabilistic, in contrast to classical k-means partitioning.