Page 1 of 1

Normalizing by mean instead of variance

Posted: 31 Oct 2012, 09:40
by boulay
Several members of my lab use % ERD as feedback, and they have attributed some importance to the actual percentage value. Therefore I have to pass the ERD % to the application. This should be easy to do in the normalizer if I add another AdaptationTypes and check for it during Update(). Let me know if there's a better way to go about getting the event-related amplitude change (in % baseline) to the application.

I guess that this isn't a change you want to include in the BCI2000 source because it would break all the standard applications to have the control signal vary between -100 and +100, but I thought I'd ask anyway in case it is another highly requested feature.

Re: Normalizing by mean instead of variance

Posted: 02 Nov 2012, 11:34
by gschalk
Chad,

For most applications, z-scoring is probably quite some better than doing % change, but what you suggest definitely makes sense for your lab.

Gerv