[Eeglablist] Using BCILAB with pre-calculated ICA data

[Avital Sternin]

Hello everyone,

I am trying to train a BCI model on independent component data (in order to
monitor the state of the ACC throughout a task) but am getting very
inconsistent results. I am having issues troubleshooting as I am unclear on
the structure that BCILAB uses when storing/calling the data.

1. When training a model: If I run the flt_ica with the TransformData
option then the data field is replaced with the ICA activations and all
subsequent channel-based methods are run on components. Is there a way to
save the dataset in this state so as not to run the ICA every time I train
the model?

2. Is it necessary to run the ICA through BCILAB? If I load a dataset that
has both the EEG data as well as the ICA activation data how do I tell
BCILAB to run on the ICA data? (I want to force it to use EEG.icaact rather
than EEG.data).

Thank you for your help!

Avital Sternin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20131007/ab56a59d/attachment.html>