help
help > RE: Removing confound regressors
Sep 26, 2014 02:09 AM | Alfonso Nieto-Castanon - Boston University
RE: Removing confound regressors
Hi Conny,
Regarding the problem reading the confound-corrected timeseries, could you please let me know what error message you are getting when trying to import those files directly into SPM? (if the files are >3Gb and you are getting a "file too small" error then this is an issue with Large File Support in SPM which we are hoping to see fixed in the near future; in the meantime breaking the singl .nii file into multiple .img files using spm_file_split should work)
Regarding the timeseries that are removed in the Denoising step, those can be found in the files:
conn*/data/ROI_Subject#_Session#.mat (e.g. White/CSF areas timeseries)
conn*/data/COV_Subject#_Session#.mat (e.g. motion, scrubbing, etc.)
In those files the variable "data" will contain the timeseries of interest for each ROI/covariate (note that denoising is performed separately for each session/run). Just for reference, the files ROI_Subject#_Condition#.mat that you mention are the ROI timeseries after the denoising step.
Hope this helps
Alfonso
Originally posted by Conny McCormick:
Regarding the problem reading the confound-corrected timeseries, could you please let me know what error message you are getting when trying to import those files directly into SPM? (if the files are >3Gb and you are getting a "file too small" error then this is an issue with Large File Support in SPM which we are hoping to see fixed in the near future; in the meantime breaking the singl .nii file into multiple .img files using spm_file_split should work)
Regarding the timeseries that are removed in the Denoising step, those can be found in the files:
conn*/data/ROI_Subject#_Session#.mat (e.g. White/CSF areas timeseries)
conn*/data/COV_Subject#_Session#.mat (e.g. motion, scrubbing, etc.)
In those files the variable "data" will contain the timeseries of interest for each ROI/covariate (note that denoising is performed separately for each session/run). Just for reference, the files ROI_Subject#_Condition#.mat that you mention are the ROI timeseries after the denoising step.
Hope this helps
Alfonso
Originally posted by Conny McCormick:
Hello,
I would like to use conn to remove WM, CSF, and motion regressors from my fMRI timeseries, to then further analyse the denoised timeseries with SPM.
I checked the box "create confound corrected time-series" and conn produced a large .nii file that contains the denoised files. However, neither SPM nor conn can work with this file. If I try to select this .nii file, there is an error saying that conn cannot view this image. I tried to split the data with fslsplit. This results in the right amount of images, right dimensions, and reasonable voxel intensities. Do you have any idea what could go wrong with reading these files?
The other question is where I can find the actual confound regressors that get removed in the end? Would it be the file ROI_Subject001_Condition001.mat? However, there are three possible locations (d1data, d2data, and data). They all have information regarding the regressors. Do you mind explaining how these are different? Or point me to a manual that explains the output of conn in such detail?
Any help would be appreciated.
Thanks,
Conny
I would like to use conn to remove WM, CSF, and motion regressors from my fMRI timeseries, to then further analyse the denoised timeseries with SPM.
I checked the box "create confound corrected time-series" and conn produced a large .nii file that contains the denoised files. However, neither SPM nor conn can work with this file. If I try to select this .nii file, there is an error saying that conn cannot view this image. I tried to split the data with fslsplit. This results in the right amount of images, right dimensions, and reasonable voxel intensities. Do you have any idea what could go wrong with reading these files?
The other question is where I can find the actual confound regressors that get removed in the end? Would it be the file ROI_Subject001_Condition001.mat? However, there are three possible locations (d1data, d2data, and data). They all have information regarding the regressors. Do you mind explaining how these are different? Or point me to a manual that explains the output of conn in such detail?
Any help would be appreciated.
Thanks,
Conny
Threaded View
| Title | Author | Date |
|---|---|---|
| Conny McCormick | Sep 25, 2014 | |
| Alfonso Nieto-Castanon | Sep 26, 2014 | |
| sam krimmel | Oct 22, 2016 | |
| Shady El Damaty | Oct 22, 2016 | |