This repository contains code for applying information bottleneck to optimise steady-state visual evoked potential (SSVEP) based brain computer interfaces (BCIs). This approach should work in other types of BCIs as well. Information bottleneck is information-theoretic optimisation method for solving specific optimisation task. Here, information bottleneck is used to find an optimal classification rule for a BCI. Optimality is viewed in terms of the standard performance measure for BCIs, the information transfer rate (ITR).
The algorithm is introduced in the article: Anti Ingel and Raul Vicente. "Information Bottleneck as Optimisation Method for SSVEP-Based BCI". Frontiers in Human Neuroscience 15 (2021). Please cite this article when using the code.
The code is written in Python and has been tested with Python 3.7. Running the code requires packages numpy, scipy, sklearn, pandas, matplotlib. Code for calculating information bottleneck is needed to run the algorithm. Please download IB.py file from this repository and add it to the src folder before trying to run the code.
The repository contains code for running the algorithm on two different datasets.
Before trying to run the code, please download the required dataset from here. This dataset contains only the extracted features. The original dataset from which these features are calculated can be found here, but this is not necessary to run the code.
Extract data folder to the same folder that contains src. Now optimisation procedure can be run using the script experiments1/4_optimise_itr.py.
Downloaded features can also be calculated from the original dataset. For that download the original dataset and put data into data/original_data folder. The extracted features can be calculated from the original data by running the files 1_original_data_to_csv.py, 2_generate_eeg_data.py, 3_generate_feature_data.py.
Before trying to run the code, please download the required dataset from here. This dataset contains only the extracted features. The original dataset from which these features are calculated can be found here, but this is not necessary to run the code.
Extract data2 folder to the same folder that contains src. Now optimisation procedure can be run using the script experiments2/optimise_itr.py.
Downloaded features can also be calculated from the original dataset. For that code from this repository was used.
For additional information, feel free to contact Anti Ingel (antiingel@gmail.com).