help
help > RE: using ART with slice timing corrected files
Sep 18, 2015 03:09 AM | Alfonso Nieto-Castanon - Boston University
RE: using ART with slice timing corrected files
Hi Catherine,
Regarding (1), yes, if possible using the subject-specific tissue masks is preferable over using template masks (if provides better identification of white/csf regions for aCompCor; see this post for additional details http://www.nitrc.org/forum/message.php?m...)
Regarding (2), you are right, the problem you are encountering is likely the result of your realignment rp_*.txt files having different number of scans/timepoints as your functional data after removing the first few scans. You are also right that if you remove a few functional scans you should probably adjust the conditions onsets (subtract TR*N from all conditions onsets, where N is the number of scans that you removed) in order to have them still point to the appropriate times in your functional timeseries. Because of all of this I think that perhaps the simplest approach is to simply flag the first few scans as outliers and have the scrubbing procedure remove them automatically during denoising (that avoids all of these issues/considerations). I will see if I can squeeze that additional option (to have the option in ART to flag the first N scans for removal) into the next release of the toolbox which is just about to go out.
Regarding (3), just to make sure, the contrast to look for between-group differences would be [1 -1] (not sure what you meant by the [1 0;0 1] part, that would look at the main connectivity effects in any of the two groups, instead of looking at the differences in connectivity between the two groups). When you display the results in SPM you can simply select the contrast named 'connectivity results' and that will already be the defined as the same contrast that you entered in CONN (no need to manually re-enter it, but of course you can also manually enter a different contrast if you prefer).
Hope this helps
Alfonso
Originally posted by Catherine Hubbard:
Regarding (1), yes, if possible using the subject-specific tissue masks is preferable over using template masks (if provides better identification of white/csf regions for aCompCor; see this post for additional details http://www.nitrc.org/forum/message.php?m...)
Regarding (2), you are right, the problem you are encountering is likely the result of your realignment rp_*.txt files having different number of scans/timepoints as your functional data after removing the first few scans. You are also right that if you remove a few functional scans you should probably adjust the conditions onsets (subtract TR*N from all conditions onsets, where N is the number of scans that you removed) in order to have them still point to the appropriate times in your functional timeseries. Because of all of this I think that perhaps the simplest approach is to simply flag the first few scans as outliers and have the scrubbing procedure remove them automatically during denoising (that avoids all of these issues/considerations). I will see if I can squeeze that additional option (to have the option in ART to flag the first N scans for removal) into the next release of the toolbox which is just about to go out.
Regarding (3), just to make sure, the contrast to look for between-group differences would be [1 -1] (not sure what you meant by the [1 0;0 1] part, that would look at the main connectivity effects in any of the two groups, instead of looking at the differences in connectivity between the two groups). When you display the results in SPM you can simply select the contrast named 'connectivity results' and that will already be the defined as the same contrast that you entered in CONN (no need to manually re-enter it, but of course you can also manually enter a different contrast if you prefer).
Hope this helps
Alfonso
Originally posted by Catherine Hubbard:
Thanks Alfonso for your reply, this has been
extremely helpful. I entered my rp_*.txt files as a first-level
covariate named 'motion' and the ART output .mat files named
'outliers' which I specified for each subject in the setup tab for
the Covariates 1st-level. All files were located in the same
directory. This seems to be working well.
1) The only thing I did differently from my description in my previous post was use the MNI brain template (as well as the grey matter, white matter and csf templates) instead of each individual's wT1 (and segmented GM, WM, and CSF files) – would it be better just to do the latter?
2) I did try specifying 'remove initial scans', for the first four volumes by running this option in the preprocessing step, but received an error. I'm wondering if this was because my rp_*.txt and .mat files had the original number of volumes and if I should have removed the first four before inputting these files at the 1st level. Also, if I do opt to use 'removal of initial scans' from the preprocessing tab, do I need to change the 'duration' time under the conditions tab to account for removal of these four volumes?
3) I went ahead a conducted a second level analysis, to look at between subjects effects in CONN. I specified the contrast 1-1 (1 0; 01) in the between-subject box (to look at differences between patients and controls) and then clicked on the results explorer tab to view the results. My question is: how do I view the results for this contrast in SPM? I tried using the 'spm display', do I simply need to respecify the contrast in the contrast manager window? Is there a way for CONN to implement this without having to respecify a new contrast in SPM?
Thanks so much for your help!!!
1) The only thing I did differently from my description in my previous post was use the MNI brain template (as well as the grey matter, white matter and csf templates) instead of each individual's wT1 (and segmented GM, WM, and CSF files) – would it be better just to do the latter?
2) I did try specifying 'remove initial scans', for the first four volumes by running this option in the preprocessing step, but received an error. I'm wondering if this was because my rp_*.txt and .mat files had the original number of volumes and if I should have removed the first four before inputting these files at the 1st level. Also, if I do opt to use 'removal of initial scans' from the preprocessing tab, do I need to change the 'duration' time under the conditions tab to account for removal of these four volumes?
3) I went ahead a conducted a second level analysis, to look at between subjects effects in CONN. I specified the contrast 1-1 (1 0; 01) in the between-subject box (to look at differences between patients and controls) and then clicked on the results explorer tab to view the results. My question is: how do I view the results for this contrast in SPM? I tried using the 'spm display', do I simply need to respecify the contrast in the contrast manager window? Is there a way for CONN to implement this without having to respecify a new contrast in SPM?
Thanks so much for your help!!!
Threaded View
| Title | Author | Date |
|---|---|---|
| Catherine Hubbard | Sep 10, 2015 | |
| Alfonso Nieto-Castanon | Sep 14, 2015 | |
| Melanie Morrison | Feb 5, 2020 | |
| Alfonso Nieto-Castanon | Feb 5, 2020 | |
| Catherine Hubbard | Sep 16, 2015 | |
| Alfonso Nieto-Castanon | Sep 18, 2015 | |
| Catherine Hubbard | Sep 18, 2015 | |
| Catherine Hubbard | Sep 18, 2015 | |
| Alfonso Nieto-Castanon | Sep 26, 2015 | |