Skip to content

Steps to Reproduce Results

This document outlines the steps to reproduce the results presented in this repository.

Important Info

Make sure you're running these steps in a system that can handle the computational load, as EEG data processing and GAN training can be resource-intensive. GPUs are recommended for training GAN models as well as having sufficient RAM for data processing.

  1. Clone the Repository: Start by cloning the repository to your local machine.

    git clone https://github.com/josephzhengg/amplab-digtwins.git
cd amplab-digtwins

  2. Install Dependencies: Install the required Python packages using pip.

    pip install -r requirements.txt

  3. Data Acquisition: Download the raw EEG data from the Healthy Brain Network (HBN) dataset (release 10) found here and place all raw data files in the raw/ directory.

  4. Data Preprocessing: Use the filter.ipynb and processed_filter.ipynb notebooks to preprocess the raw EEG data.

    Notebooks Used

    • filter.ipynb: Performs low-pass filtering and artifact removal using ASR. (Saves to processed/ directory)

    • processed_filter.ipynb: Performs channel selection and further filtering of processed data. (Saves to filtered_processed/ directory)

  5. Task Separation: Utilize the separate_tasks_resting_state.ipynb notebook to separate different tasks (e.g., eyes open vs. eyes closed) from the resting-state EEG data.

  6. GAN Training: Train GAN models on the resting-state EEG data using the resting_state_gans.ipynb notebook. The generated models will be saved in the gan_data/ directory.

  7. Generate Synthetic Data: Use the trained GAN models to generate synthetic EEG data samples, which will be saved in the synthetic_data/ directory (this will automatically be done in the GAN training notebook).

Every single step and notebook is designed to flow seamlessly into the next, so following these steps in order should allow you to reproduce the results effectively. Again, this repo wants to encourage experimentation, so feel free to modify the notebooks and explore different configurations!