Skip to content

GEDAI v1.3

Choose a tag to compare

@neurotuning-personal neurotuning-personal released this 09 Dec 20:55
· 259 commits to main since this release

Introduces the possibility to reject very "bad" epochs during denoising (i.e. prevent GEDAI from cleaning them), and cropping the data. Bad epochs can be removed using their ENOVA (Explained Noise Variance), which is bounded between 0 and 1. The ENOVA is the variance explained of the noise removed by GEDAI, expressed as a proportion of the variance of the original EEG data. The default ENOVA threshold (i.e. 0.9) removes epochs with more than 90% noise variance, while a more agressive ENOVA threhold (e.g. 0.3) would remove all epochs with > 30% noise.