higher-order-interactions-eeg

Under the supervision of Professor Mohammad Reza Rahimi Tabar, a leading researcher in complex systems and time series analysis, I analyzed a large set of EEG signals recorded from 109 subjects during simple cognitive tasks(for further references see PhysioNet). The goal was to identify causal and asymmetric connections between seven key brain regions. To achieve this, we applied an advanced method that captures higher-order interactions between different time series. As a result, we reconstructed not only a first-order graph but also a hypergraph, both comprising the seven brain regions.

⚠️ Important Note

To run main_eeg_analysis.ipynb, a large EEG dataset must be downloaded in advance.

• Database source:
  https://physionet.org/content/eegmmidb/1.0.0/

• Direct download link (ZIP):
  https://physionet.org/content/eegmmidb/get-zip/1.0.0/

Moreover, mne and hints package need to be installed:

HiNTS Package Installation(Higher-Order Interactions in N-Dimensional Time Series):

For more detailed examples and tutorials, please refer to authors' source hints package and the corresponding papers:

• Akhshi et al. (2024): GitHub Link
• Tabar et al. (2024): PRX Paper

To install HiNTS, you can either use pip or install it directly from the source.

Option 1: Install via pip

pip install hints-kmcs

Option 2: Install from source

git clone https://github.com/aminakhshi/hints.git
cd hints
python setup.py install