DeepBCI allows you to configure the design and conduct an experiment in real-time EEG/MEG paradigm.

Prerequisites: python, git, optionally conda.

Warning: DeepBCI requires you install some outdated versions of packages. Consider installing it in a virtual enviroment, using tools such as venv or conda. For example, if using conda, create and activate a new environment by running these commands first:

conda create -n DeepBCI python pip
conda activate DeepBCI

Regardless of whether or not you are using a virtual environment, clone this repository and install the package in editable mode by running:

git clone https://github.com/nickware44/DeepBCI.git
cd DeepBCI
pip install -e .

After installation, DeepBCI can be run from anywhere by using this command (when using a virtual environment, it has to be active):

pynfb

Or from the folder you installed it in:

python pynfb/main.py

If you need to use DeepBCI without a console, or run it from anywhere and distribute it, the best option is to freeze it into an executable.

DeepBCI supports building as an executable, using the pyinstaller module. To use it, first install DeepBCI with freeze addon:

pip install -e .[freeze]

Then build the executable from the included spec file (this might take some time):

pyinstaller freeze.spec

The executable can then be found in the dist folder.

DeepBCI supports commandline arguments to streamline opening and running experiments:

usage: pynfb [-h] [-x] [file]

positional arguments:
  file           open an xml experiment file when launched (optional)

optional arguments:
  -h, --help     show this help message and exit
  -x, --execute  run the experiment without configuring (requires file to be specified)

For example, to open an XML file with an experiment that you designed from command line, specify the path to it like so:

python pynfb/main.py your-experiment-file.xml

To run the experiment without configuring, use the -x or --execute option:

python pynfb/main.py -x your-experiment-file.xml

Note: The same command-line options are available for the pynfb executable, but since the console output is hidden, if you make a mistake the application will just not start.

This work was supported by the Center for Bioelectric Interfaces of the Institute for Cognitive Neuroscience of the National Research University Higher School of Economics, RF Government grant, ag. No. 14.641.31.0003.