[Eeglablist] the formular to reconstruct data from ICs

[arno]

Dear xian,


> supose I only use some goodIC to reconstrust data
>  
> is this formula right?
> data=eeg.icawinv(:,goodIC)*eeg.icaact(goodIC,:)
Yes, this is correct.
> the result looks right but how do I use EEG.icasphere in this formula?
You don't since EEG.icawinv is the inverse matrix of 
EEG.icaweight*EEG.icasphere (and not the inverse of EEG.icaweight only).

Best,

Arno