 |
 |
Hello, we typically apply AROMA to each run individually, as this allows to correct for potential run-specific motion. In your case the over speech runs might require a different type of correction, putting all runs together might 'average' that out. This is also how we applied AROMA in the Pruim et al. papers, ie. to each run separately. The only thing to keep in mind when applying AROMA on SPM data is the difference in bounding box size between the two programs. Not correcting for this will impact proper mask placement and feature extraction. hth, maarten On Fri, Aug 19, 2016 at 10:29 PM, Hamann, Stephan <[log in to unmask]> wrote: > Dear FSL experts, > > I’m trying to use ICA-AROMA on some fMRI data that were preprocessed with > SPM12 and wanted to check whether the steps I’m using are correct, as I’m a > relative newbie with FSL. > > Each subject has 4 functional runs, two of which involve overt speech > responses, which is the reason for trying ICA-AROMA as it seems to do an > excellent job of reducing motion artifacts relative to some other methods. > > 1. In SPM for each individual subject I’ve (a) realigned the NIfTI images, > (b) normalized them to MNI space, (c) smoothed the images, and (d) > converted the 3D images across all four functional runs into a single 4D > file (one for each subject). > > 2. Next, I ran ICA-AROMA on each subject’s 4D file individually with the > required motion parameter files, mask image, etc. > > From the Pruim et al. papers on ICA-AROMA and the manual it’s still a > little unclear whether the processing should be done for each subject > individually (across all runs). The task-based fMRI example they use has > multiple runs per subject but it wasn’t clear to me how the data from > different runs were entered into ICA-AROMA. > > Some other ICA-based programs such as GIFT involve entering all the group > data to compute components, so a possible alternative would be to enter all > of the subjects at once (?). Or alternately, ICA-AROMA could be run on the > data from each individual run, but it seems preferable to concatenate data > across all of a subject's runs for ICA. > > Some tips on the best workflow would be much appreciated! > > Thanks, > > —Stephan > > > > > > > > > > > > ________________________________ > > This e-mail message (including any attachments) is for the sole use of > the intended recipient(s) and may contain confidential and privileged > information. If the reader of this message is not the intended > recipient, you are hereby notified that any dissemination, distribution > or copying of this message (including any attachments) is strictly > prohibited. > > If you have received this message in error, please contact > the sender by reply e-mail message and destroy all copies of the > original message (including attachments). > -- Maarten Mennes, Ph.D. Senior Researcher Donders Institute for Brain, Cognition and Behaviour Radboud University Nijmegen Nijmegen The Netherlands Google Scholar Author Link <http://scholar.google.com/citations?user=pLlSTVgAAAAJ&hl=en>