🧠 MindCine: Multimodal EEG-to-Video Reconstruction with Large-Scale Pretrained Models

Tian-Yi Zhou*, Xuan-Hao Liu*, Bao-Liang Lu, Wei-Long Zheng†,

Shanghai Jiao Tong University

* equal contribution † denotes the corresponding author

You can also see the poster to breifly know our work.

Brain Decoding Paradigms: Previous vs. Ours.

✨ Update

• 2026/01/21 MindCine is accepted by ICASSP 2026.

🛠️ Environment Setup

Quick Start

# 1. Clone this repo
git clone https://github.com/KevinZhou6/MindCine.git
cd MindCine

# 2. Create the Conda environment
conda env create -f environment.yml

# 3. Activate the environment
conda activate MindCine

👍 Citations

If you find our work useful, please consider citing:

@article{zhou2026mindcine,
  title={MindCine: Multimodal EEG-to-Video Reconstruction with Large-Scale Pretrained Models},
  author={Zhou, Tian-Yi and Liu, Xuan-Hao and Lu, Bao-Liang and Zheng, Wei-Long},
  journal={arXiv preprint arXiv:2601.18192},
  year={2026}
}

😺Acknowledge

We sincerely thank the following outstanding works:

1. EEG2Video — EEG2Video: Towards Decoding Dynamic Visual Perception from EEG Signals.
2. CognitionCapturer - CognitionCapturer: Decoding Visual Stimuli from Human EEG Signals with Multimodal Information.
3. We use the BIOT, LaBraM, EEGPT, CBraMod, Gram to alleviate data scarcity.

🏷️ License

This repository is released under the MIT license. See LICENSE for additional details.