What about analysing your task-based data in the usual way (ie
specifying a design matrix containing everything you know about your
experimental design) then using the "Write residuals" option when
estimating your model? You will be effectively regressing out task (and
high pass filtering and whitening the data) that you can then feed in
your functional connectivity analysis.
On 31/03/17 13:52, Maaly Nassar wrote:
> Thanks Donald.
> I need to regress out the fitted response from ROIs time series (as described in Fair et al., 2007*) and check if the functional connectivity of specific ROIs (after regressing task-based response as confounding factor) is similar to their functional connectivity from continuous resting state data. Simply, I need to replicate Fair's study and emulate resting state from task-based data. In CONN, the "deionization step" can consider the task-based fitted responses as confounding factors and regress them out, so I was wondering if I can do the same in SPM using the multiple regression section in fMRI specification???
> *Fair DA, Schlaggar BL, Cohen AL, Miezin FM, Dosenbach NUF, Wenger KK, et al. A method for using blocked and event-related fMRI data to study “resting state” functional connectivity. Neuroimage. 2007 Mar;35(1):396–405.
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG