bci2000.org BBS

In some classification algorithms, we should calculate some thresholds. So before the algorithm was applied in online BCI system,Should we follow the online data processing for offline processing? For example, in offline program, if we want the data normalized , how many points we use? per blocksize or per trial?
thanks

I'm not sure whether I understand what you are trying to do. If your offline algorithm does not do normalization by itself, then why would you want to do it online? Or, in reverse, why would you want to add normalization to an offline algorithm that does not require it?