Can Natural Image Autoencoders Compactly Tokenize fMRI Volumes for Long-Range Dynamics Modeling?
Paper โข 2604.03619 โข Published โข 9
Masked-pretraining checkpoint for the TABLeT encoder, trained on UK Biobank fMRI data.
If you use this model, please cite:
@article{kim2026tablet,
title = {Can Natural Image Autoencoders Compactly Tokenize fMRI Volumes for Long-Range Dynamics Modeling?},
author = {Kim, Peter Yongho and Park, Juhyeon and Park, Jungwoo and Choi, Jubin and Seo, Jungwoo and Cha, Jiook and Moon, Taesup},
journal = {arXiv preprint arXiv:2604.03619},
year = {2026},
url = {https://arxiv.org/abs/2604.03619}
}
Paper: arxiv.org/abs/2604.03619