Models Summary

This page offers a summary of many implemented models. Please note that this list may not be exhaustive. For the definitive and most current list, including detailed class documentation, please consult the :doc:`API documentation <api>`.

We are continually expanding this collection and welcome contributions! If you have implemented a model relevant to EEG, EcoG, or MEG analysis, consider adding it to Braindecode. See the "Submit a New Model" section below for details.

Visualization comparing the models based on their total number of parameters (left plot) and the primary experimental paradigm they were designed for (right plot).

Columns definitions:

• Model: The name of the model.
• Paradigm: The paradigm(s) the model is typically used for (e.g., Motor Imagery, P300, Sleep Staging). 'General' indicates applicability across multiple paradigms or no specific paradigm focus.
• Type: The model's primary function (e.g., Classification, Regression, Embedding).
• Freq (Hz): The data sampling rate (in Hertz) the model is designed for. Note that this might be adaptable depending on the specific dataset and application.
• Hyperparameters: The mandatory hyperparameters required for instantiating the model class. These may include n_chans (number of channels), n_outputs (number of output classes or regression targets), n_times (number of time points in the input window), or sfreq (sampling frequency). Also, n_times can be derived implicitly by providing both sfreq and input_window_seconds.
• #Parameters: The approximate total number of trainable parameters in the model, calculated using a consistent configuration (see note below).

The parameter counts shown in the table were calculated using consistent hyperparameters for models within the same paradigm, based largely on Braindecode's default implementation values. These counts provide a relative comparison but may differ from those reported in the original publications due to variations in specific architectural details, input dimensions used in the paper, or calculation methods.

Submit a new model

Want to contribute a new model to Braindecode? Great! You can propose a new model by opening an issue (please include a link to the relevant publication or description) or, even better, directly submit your implementation via a pull request. We appreciate your contributions to expanding the library!