Authors: Vernon Lawhern (ARL)
Software Language(s): Python, Tensorflow
Software Type: Machine Learning Model Definitions, Visualization Tools
What it does:
Provides tools for training and visualization of learned features from Compact Convolutional Neural Networks (CNNs) for EEG-based Brain-Computer Interfaces (BCI).
Implementation of EEGNet, a compact CNN for neural decoding that generalizes across paradigms better than, and achieves comparably high performance to, the state-of-the-art BCI algorithms.
Why is this important?
Cross-subject, cross-task BCI decoders that require little or no calibration are essential for any real-world application of non-invasive neurotechnology.
Lawhern, V. J., Solon, A. J., Waytowich, N. R., Gordon, S. M., Hung, C. P., & Lance, B. J. (2018). EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces. Journal of neural engineering, 15(5), 056013.