Dear list,
as an extension to the last thread on SSPs I've got a few questions.
As I understood it so far one can include several projectors, such
that noise-covariance SSPs and EOG /ECG SSP may complement each other.
However, Is there actually one preferable way of obtaining
noise-covariance SSPs, that is empty room VS pre-event baseline?
Or does it even make sense to combine the two?
Cheers,
Denis
Hi Denis,
You can apply SSPs to multiple noise sources on top of each other... SSPs operate by taking out data that have a spatial pattern similar to the spatial pattern of the noise that you specify... Hence you have to be careful when making SSPs out of baseline data... If you have long enough chunks of baseline data then usually only noises dominate the first few SSP components but you should inspect to make sure that is the case...
Regards,
Hari
Thanks for the quick reply Hari,
so in the first place creating ECG / EOG / empty room SSPs should
yield a reasonable time-spent-checking / noise reduction trade-off...
In the case of the study I'am currently analysing the event-related
SSP might be difficult, as the paradigm implies a high level baseline
which means that the non-events are already quite meaningful and the
difference between baseline and event is really specific. Also the
breaks between the events are just 2 to 3 seconds.
Denis
2012/9/10 Hari Bharadwaj <hari at nmr.mgh.harvard.edu>: