help
help > RE: roi MRCroGL and brain extraction
Mar 1, 2022 10:03 PM | bionerd
RE: roi MRCroGL and brain extraction
Originally posted by Chris Rorden:
Thank you for your reply. Since you mentioned fsl, I was wondering if you'd have any idea about a few queries I stumbled upon,
1) I have multi-shell DTI data and I had read that FSL's DTIFIT tool isn't ideal for processing this kind of data that has multiple b-values (please see attached image and site: https://fsl.fmrib.ox.ac.uk/fslcourse/201... for source info). However, when I ran it on my data, it somehow did produce FA maps, etc. All I need to know is whether anybody has a good detailed understanding about the background process that FSL performs in order to generate these results and how it exactly did that for my data???
2) Secondly, with respect to the same data, when I tried performing topup on it, it gave me the error: 'Topup: msg=topup_clp::topup_clp: Mismatch between b0_all and acqparams.txt' and I checked my acquisition paramters for the original dti nifti and it said that the phase enconding direction was: j- which I assume means it is A>>P encoded, correct?? The total readout time was 0.038625 so my acqparams looks like this: 0 -1 0.038625
3) Point-2 brings me to my third question: for topup, we merge A>>P (b0 image) and P>>A; how does one get the P>>A image? Is it supposed to be in the raw data I received from the scanner itself or is it something one generates (if so, how)?
I also came across a prebaked topup tool but am confused as to whether this is a file that needs to be downloaded first.
4) I needed to plot a graph between the signal intensity natural logarithm and the corresponding b values of 1000. Does anyone have any suggestions on how to do this?
What I tried previously was averaging the signal intensities over time (received this data from the time series plot using FSL) and then plotting it vs. b values from the bvals file. The problem being that dti data has different gradient directions and I think that is not taken into account when averaging signal intensities for say a common b value like 1000?
Any and all help would be greatly appreciated. Thanks for your time.
You can draw a region of interest, the wiki
provides details
https://www.nitrc.org/plugins/mwiki/index.php/mricrogl:MainPage#Drawing_Regions_of_Interest
The View/ExtractBrain menu item will run BET to create a Brain Extraction. On some operating systems you may need to install FSL to get BET.
The FSL BET is pretty simple, and may not generate a good result. I tend to use SPM's segmentation for this - my SPM script is here
https://github.com/rordenlab/spmScripts/blob/master/nii_nii2stl.m
https://www.nitrc.org/plugins/mwiki/index.php/mricrogl:MainPage#Drawing_Regions_of_Interest
The View/ExtractBrain menu item will run BET to create a Brain Extraction. On some operating systems you may need to install FSL to get BET.
The FSL BET is pretty simple, and may not generate a good result. I tend to use SPM's segmentation for this - my SPM script is here
https://github.com/rordenlab/spmScripts/blob/master/nii_nii2stl.m
Thank you for your reply. Since you mentioned fsl, I was wondering if you'd have any idea about a few queries I stumbled upon,
1) I have multi-shell DTI data and I had read that FSL's DTIFIT tool isn't ideal for processing this kind of data that has multiple b-values (please see attached image and site: https://fsl.fmrib.ox.ac.uk/fslcourse/201... for source info). However, when I ran it on my data, it somehow did produce FA maps, etc. All I need to know is whether anybody has a good detailed understanding about the background process that FSL performs in order to generate these results and how it exactly did that for my data???
2) Secondly, with respect to the same data, when I tried performing topup on it, it gave me the error: 'Topup: msg=topup_clp::topup_clp: Mismatch between b0_all and acqparams.txt' and I checked my acquisition paramters for the original dti nifti and it said that the phase enconding direction was: j- which I assume means it is A>>P encoded, correct?? The total readout time was 0.038625 so my acqparams looks like this: 0 -1 0.038625
3) Point-2 brings me to my third question: for topup, we merge A>>P (b0 image) and P>>A; how does one get the P>>A image? Is it supposed to be in the raw data I received from the scanner itself or is it something one generates (if so, how)?
I also came across a prebaked topup tool but am confused as to whether this is a file that needs to be downloaded first.
4) I needed to plot a graph between the signal intensity natural logarithm and the corresponding b values of 1000. Does anyone have any suggestions on how to do this?
What I tried previously was averaging the signal intensities over time (received this data from the time series plot using FSL) and then plotting it vs. b values from the bvals file. The problem being that dti data has different gradient directions and I think that is not taken into account when averaging signal intensities for say a common b value like 1000?
Any and all help would be greatly appreciated. Thanks for your time.
Threaded View
| Title | Author | Date |
|---|---|---|
| bionerd | Mar 1, 2022 | |
| Chris Rorden | Mar 1, 2022 | |
| bionerd | Mar 1, 2022 | |
| Chris Rorden | Mar 1, 2022 | |
| bionerd | Mar 1, 2022 | |