Hi,

If you have run the denoising step in Conn, then the data would be band pass filtered, not prior to that. 

I am not in lab right now so can't run this and confirm, but I assume that "after regression" option would apply the band-pass filter _after_ the confounds have been regressed out (confounds would be realignment, time series from WM, CSF, and other confounds like main effect of task, if included), as opposed to filtering and regression steps happening together.

I think you can skip band-pass filtering by defining your filter to encompass the entire frequency range ([0 Inf] should do the trick, I believe?)

If you just want the time series, the easiest way would be to call REX (which is a part of the Conn toolbox). An example usage can be as follows (note that REX is quite powerful and you should explore other ways to call REX):

time_series = rex(functional volume, mask, 'summary_measure', 'mean', 'level', 'rois', 'scaling', 'roi');

An example of the "mask" above can be the brainmask.nii file shipped with conn (conn\utils\surf\mask.volume.brainmask.nii).

If you have done preprocessing and denoising within Conn, then you can save the time series from all ROIs as .mat files (select these in the Options pane in Setup tab). See here for a description and location of most of the files that Conn would create. 

Hope this helps


Best
Pravesh
Originally posted by Isadora Olive: