help
help > RE: registration problem
Sep 28, 2016 11:09 AM | Pravesh Parekh - National Institute of Mental Health and Neurosciences
RE: registration problem
Dear Dr. Alfonso,
I was under the impression that the 'r' prefix is prepended only when reslicing step is explicitly being performed. Please correct me if the below is an incorrect interpretation:
The realignment step has three options (excluding the realign and unwarp):
Realign (estimate): which only outputs the rp_*.txt file (six motion parameters)
Realign (reslice): changes the voxel dimensions of the images to match the first image and outputs the 'r*' file
Realign (estimate and reslice): which does both the above
The coregistration step similarly has three options:
Coregister (estimate): modifies the header to match the source and template (no explicit output)
Coregister (reslice): matches the voxel dimensions to some template space (output is a r* file)
Coregister (estimate and reslice): coregisters two images and then reslices the second to match the first
So, in the present case, a coregister (estimate and reslice) step would actually create resliced functional volume which would be oriented in the same way as structural?
Yours sincerely
Pravesh
Originally posted by Alfonso Nieto-Castanon:
I was under the impression that the 'r' prefix is prepended only when reslicing step is explicitly being performed. Please correct me if the below is an incorrect interpretation:
The realignment step has three options (excluding the realign and unwarp):
Realign (estimate): which only outputs the rp_*.txt file (six motion parameters)
Realign (reslice): changes the voxel dimensions of the images to match the first image and outputs the 'r*' file
Realign (estimate and reslice): which does both the above
The coregistration step similarly has three options:
Coregister (estimate): modifies the header to match the source and template (no explicit output)
Coregister (reslice): matches the voxel dimensions to some template space (output is a r* file)
Coregister (estimate and reslice): coregisters two images and then reslices the second to match the first
So, in the present case, a coregister (estimate and reslice) step would actually create resliced functional volume which would be oriented in the same way as structural?
Yours sincerely
Pravesh
Originally posted by Alfonso Nieto-Castanon:
Hi Heidi &
Pravesh,
Sorry about the confusion, that was my mistake (I should have referred to swau* files instead of swar*, and au* instead of ar*). The 'r' prefix is prepended when one uses simply the "realignment" step, while the 'u' prefix is instead prepended when one uses the simultaneous "realign and unwarp" step (which is the one used in the default MNI pipeline in CONN). If you are using the default MNI-space preprocessing pipeline your fully preprocessed functional volumes should be named swau[original_files] (note that adding an intermediate 'functional coregistration' step will not change this, since coregistration does not produce additional files, it simply changes the header info in your functional volumes to indicate the associated change in the image-to-world mapping)
Let me know if this clarifies
Best
Alfonso
Originally posted by Pravesh Parekh:
Sorry about the confusion, that was my mistake (I should have referred to swau* files instead of swar*, and au* instead of ar*). The 'r' prefix is prepended when one uses simply the "realignment" step, while the 'u' prefix is instead prepended when one uses the simultaneous "realign and unwarp" step (which is the one used in the default MNI pipeline in CONN). If you are using the default MNI-space preprocessing pipeline your fully preprocessed functional volumes should be named swau[original_files] (note that adding an intermediate 'functional coregistration' step will not change this, since coregistration does not produce additional files, it simply changes the header info in your functional volumes to indicate the associated change in the image-to-world mapping)
Let me know if this clarifies
Best
Alfonso
Originally posted by Pravesh Parekh:
Hi
Heidi,
'r' in SPM file naming convention is usually added to file names as prefix when reslicing is performed. Given Dr. Alfonso's suggestion, I think you might want to run the 'coregister and reslice; option which should give you a r* file; however, the functional coregistration to structural step in Conn does not reslice [till 15h] which is perhaps why you are not seeing the r* file.
If indeed the above makes sense and is correct, you could just run the pipeline till slice timing correction. After that, use the "apply individual preprocessing step" option (functional tab; lower side of the screen, upon clicking the "functional tools" option). There, you can modify the coregistration step to coregister and reslice. Once this step is over, you can then continue running the rest of the MNI steps by going back to the main preprocessing tab. Alternatively, you can run that step in SPM outside of Conn and change the files in the functional tabs.
Of course, the above is assuming that I have interpreted Dr. Alfonso's suggestions correctly...but maybe he would be able to offer a more convenient way to get around the bend?
Regards
Pravesh
Originally posted by Heidi Jacobs:
'r' in SPM file naming convention is usually added to file names as prefix when reslicing is performed. Given Dr. Alfonso's suggestion, I think you might want to run the 'coregister and reslice; option which should give you a r* file; however, the functional coregistration to structural step in Conn does not reslice [till 15h] which is perhaps why you are not seeing the r* file.
If indeed the above makes sense and is correct, you could just run the pipeline till slice timing correction. After that, use the "apply individual preprocessing step" option (functional tab; lower side of the screen, upon clicking the "functional tools" option). There, you can modify the coregistration step to coregister and reslice. Once this step is over, you can then continue running the rest of the MNI steps by going back to the main preprocessing tab. Alternatively, you can run that step in SPM outside of Conn and change the files in the functional tabs.
Of course, the above is assuming that I have interpreted Dr. Alfonso's suggestions correctly...but maybe he would be able to offer a more convenient way to get around the bend?
Regards
Pravesh
Originally posted by Heidi Jacobs:
Dear Alfonso,
I have been trying to follow your suggestions. It took a while (59 7T datasets take some time :)), but now I noticed that I do not have swar files, but swau files.
I have been trying to figure out the spm prefix system and the R seems to indicate "registration".
Do you know why I don't have swar files?
I have chosen the default preprocessing with the extra "functional coregistration to structural" step.
I checked the folder with the data and there is also no swar file....
Many thanks
Heidi
I have been trying to follow your suggestions. It took a while (59 7T datasets take some time :)), but now I noticed that I do not have swar files, but swau files.
I have been trying to figure out the spm prefix system and the R seems to indicate "registration".
Do you know why I don't have swar files?
I have chosen the default preprocessing with the extra "functional coregistration to structural" step.
I checked the folder with the data and there is also no swar file....
Many thanks
Heidi
Threaded View
| Title | Author | Date |
|---|---|---|
| Heidi Jacobs | Sep 5, 2016 | |
| Alfonso Nieto-Castanon | Sep 8, 2016 | |
| Heidi Jacobs | Sep 14, 2016 | |
| Pravesh Parekh | Sep 23, 2016 | |
| Alfonso Nieto-Castanon | Sep 27, 2016 | |
| Heidi Jacobs | Sep 28, 2016 | |
| Alfonso Nieto-Castanon | Sep 28, 2016 | |
| Heidi Jacobs | Sep 28, 2016 | |
| Pravesh Parekh | Sep 28, 2016 | |
| Alfonso Nieto-Castanon | Sep 28, 2016 | |