Generalized neural decoders for transfer learning across participants and recording modalities

Data used for analyzing HTNet generalized neural decoder (see published paper: https://doi.org/10.1088/1741-2552/abda0b). This includes publicly available ECoG data, divided into 4 second segments that centered around move or rest events. Event labels can be found in the last channel/electrode. Projection matrices are also included for every ECoG participant and one EEG participant (for use with EEG data used for our manuscript, available at http://bnci-horizon-2020.eu/database/data-sets (#25)). These matrices provide a mapping from electrodes to common brain regions of interest (see our paper for further details). The EEG projection matrix was obtained from Fieldtrip based on electrode names (Oostenveld et al. Comput Intell Neurosci 2011). This project is designed to be used in conjunction with our publicly released HTNet code (https://github.com/BruntonUWBio/HTNet_generalized_decoding) to replicate the findings from our paper. For this purpose, place EEG and ECoG data files in separate folders labelled 'eeg_dataset' and 'ecog_dataset', respectively. Then put each modality's projection matrices in a subfolder marked 'proj_mat' within each dataset folder.