Cholesky-Space-for-Brain-Computer-Interfaces

This repository is the official implementation for the paper submission titled:

"Cholesky Space for Brain–Computer Interfaces"

1. Environment Setup

1.1 Requirements

The following dependencies are required to run the code:

• Python 3.10
• PyTorch 2.2.2 + CUDA 12.1
• numpy == 1.24.4
• pandas == 2.2.3
• matplotlib == 3.9.0
• scikit-learn == 1.5.0
• moabb == 1.2.0

1.2 Installation Steps

We recommend installing PyTorch first manually (to ensure CUDA compatibility) before other dependencies:

# Install PyTorch with CUDA 12.1 support
pip install torch==2.2.2 torchvision==0.17.2 torchaudio==2.2.2 --index-url https://download.pytorch.org/whl/cu121

Then install the remaining dependencies via requirements.txt:

pip install -r requirements.txt

2. Datasets

Currently, the code supports the following dataset:

• BCIC IV 2a (BNCI 2014-001) (~743MB)

Important Note

✅ No manual download required!
The raw dataset will be automatically downloaded from the official source during the first run of the script.

3. How to Run

3.1 Default Run

To train all models on all supported datasets with default hyperparameters, simply execute:

python main.py

This command will:

1. Automatically download and preprocess the dataset.
2. Train all baseline/models in the pipeline.
3. Save training logs and results to the specified directory (see main.py for details).

3.2 Custom Run (Specific Datasets/Models)

To train on specific datasets or use specific models, modify main.py as follows:

1. Open main.py and locate the two key lists (e.g., DATASETS and MODELS).
2. Uncomment the lines corresponding to your target dataset/model.
3. Re-run the script:

   python main.py

4. Citation

If you find this code or our work helpful for your research, please consider giving this repository a ⭐ Star and citing our paper:

@ARTICLE{10922209,
  author={Wang, Xingfu and Qi, Wenxia and Yang, Wenjie and Wang, Wei},
  journal={IEEE Transactions on Neural Networks and Learning Systems}, 
  title={Cholesky Space for Brain–Computer Interfaces}, 
  year={2025},
  volume={36},
  number={8},
  pages={15424-15435},
  keywords={Electroencephalography;Manifolds;Brain modeling;Decoding;Feature extraction;Motors;Emotion recognition;Covariance matrices;Computational efficiency;Vectors;Brain–computer interface (BCI);Cholesky space;electroencephalogram (EEG);Riemannian manifold},
  doi={10.1109/TNNLS.2025.3542801}}