Abstract: In this study, we developed a Transformer-based autoencoder (TAE) model for temporal dimensionality reduction of fMRI data, harnessing the Transformer's ability to capture temporal ...