CNT Pre-processing Toolkit

Toolkit for pre-processing of intracranial EEG data, and an interactive pipeline for pre-processing method evaluation.

Toolkit available in Matlab and Python, compatible with iEEG.org

Python Installation

pip install

pip install git+https://github.com/penn-cnt/CNT_research_tools.git#subdirectory=python

Alternatively, download or clone the toolbox into a local folder via git, switch to folder, and pip install locally:

git clone git@github.com:haoershi/CNT_research_tools.git
cd CNT_research_tools/python
pip install .

Set Conda Environment

An environment could be set through conda to reduce the likelihood of dependency conflicts and making it easier to set up.

Dependencies:

• anaconda

Create a conda environment and activate:

conda env create -n ieegpy -f ieegpy.yml
conda activate ieegpy

If the above command doesn't work, you can manually create an enviornment and install the necessary libraries:

conda create -n ieegpy python=3.9
conda activate ieegpy
pip install -r requirements.txt

Testing

Run pytest to ensure no running issues. (Getting data may not be tested currently) (Need update for the new system)

MATLAB Installation

Dependencies:

• MATLAB >= R2021b
• Signal Processing Toolbox
• Statistics and Machine Learning Toolbox Toolboxes could be installed via Adds-Ons > Get Adds-Ons.
• IEEG MATLAB Toolbox (Can be downloaded at https://main.ieeg.org/?q=node/29, or we've provided with our toolkit)

Add folder matlab in MATLAB working directory.

addpath(genpath('path/CNT_research_tools/matlab'));

Testing

Run unit tests to ensure no running issues: (Getting data may not be tested currently, need update to the current system)

runtests('matlab/test','IncludeSubfolders',true);

Folder Structure

During usage of toolkit, folders users and data would be created under the python/CNTtools or matlab folder to store user login information and data files, respectively.

Login Configuration

The toolkit currently depends on ieeg.org.

To access data, please register first on https://www.ieeg.org.

A usr_ieeglogin.bin password file and a usr_config.json file are required in the user folder before data downloading can run correctly.

Files could be automatically generated throught running login configuration and input of username and password.

session = iEEGPreprocess()
session.login_config()
# input of user information

Functions

The toolbox includes the following functions:

• Download data from ieeg.org
• Standardize channel labels and identify band channels
• Signal filtering
• Data re-referencing
  • Common Average Re-referencing (CAR)
  • Bipolar Re-referencing (BR)
  • Laplacian Re-referencing (LR)
• Pre-whitening
• Feature extraction
• Connectivity calculation
• Plotting and connectivity heatmap

Usage

This toolkit provides a recommended usage pipeline in the form of interactive notebooks in both MATLAB and Python.

For illustration, this toolkit is also used for an systematic evaluation of pre-processing methods, as shown in the demo folder.