Hi Brian,
applying FASTER to continuous data is certainly possible - if you don't select any epoching options, the  ICA and later steps will be applied to the continuous data. As you say, there will be increased computation time and the possibility of bad blocks of data, which can bias the ICA with a lot of ICs relating to these bad blocks, which mucks up the statistics a bit. The "unmarkered epoching" option in the epoching section of the GUI was introduced to do exactly as you suggested, split the data into e.g. 1s epochs. This allows the use of the epoch rejection tool for removing sections of data with lots of contamination, allowing the ICA a better chance of a sensible decomposition.
Hope that helps a bit, any further questions  don't hesitate to ask,
Thanks for your interest,

On 19 October 2011 23:18, Brian Roach <brian.roach@ncire.org> wrote:

I'm interested in potentially applying steps from the FASTER routine to continuous EEG data (like resting state data).  The first step if applied to continuous data, and then I am wondering if skipping step 2 but still applying step 3 would work.  Specifically, how important would it be to have epochs vs continuous data passed to ICA?  Since EEGlab's ICA is spatial and the outlier IC detection steps examine spatial, spectral, and timecourse components of ICs without respect to trials, I think this would work OK.  However, has anyone tried it?  Are there potential issues with the EOG correlations that I should consider?  I appreciate that a large block of continuous data means more input to ICA and potentially greater computation time, but that is OK with me as long as it seems like a valid approach.  Without applying steps 2 and 4, there are potentially bad blocks of data that wouldn't be rejected.  Perhaps a better alternative is to cut continuous data into 1s "epochs" for step 4.  If I used that same approach and applied step 2, I would probably cut out chunks of the continuous EEG.

If anyone has tested anything like this and has advice, I would appreciate it!


The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Ciosco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
Faster-eeg-list mailing list

Hugh Nolan,
Trinity Centre for Bioengineering,
Printing House,
Trinity College Dublin.

Tel: +353861297722