Dear Maarten,

What exactly do you mean by 'supposed to be filtered'? If you want to take out a certain part of the variance (such as the mean), then make sure to select 'correct for XX', where XX is the F contrast of the regressors explaining variance that you want to include, and excluding regressors that you do not want to include. In case of mean centering, you would leave out the constant term.

Hope this helps,
Hanneke

On 19 January 2012 10:57, Maarten De Vos <[log in to unmask]> wrote:
dear SPM experts,

for the analysis of ROIs of face processing, I am trying to save individual blobs as ROI.   For this,

- I press Eigenvariate button, give it a name, and select under VOI definition 'cluster'.  It indeed saves the concerning blob into an image, but despite the fact that the time series is supposed to be filtered, the mean activity in the blob is centered around a non-zero value.  (see attached image)

Can somebody confirm that the time series of the roi is indeed filtered, how it is then filtered and why i still have this clear offset?

thanks a lot for your help,
maarten



--
Hanneke den Ouden, PhD

Center for Neural Science
New York University
4 Washington Place, room 873
New York NY 10003