 |
 |
Hi, To start with, are you sure you were using the mm^3 volume output from fslstats and not the number of voxels? If you used the latter it would definitely result in much larger values for your 0.7mm images. It would be helpful to know the exact command you were running for this. If this isn't the problem, I would still not expect the use of the vtk to fix things. If there is an overestimate then you should look at the resulting images are decide what is causing the problem. It might be related to a registration problem or the boundary correction. It is difficult to diagnose such a problem blindly and I would really strongly recommend that you carefully inspect some examples where you think the volume is overestimated, and compare this to cases where you think the volume is OK. If you find something that is not working well (e.g. boundary correction, but inspecting the origseg outputs) and you are not sure how to best fix it then let us know the details and we'll try to help. All the best, Mark From: FSL - FMRIB's Software Library <[log in to unmask]<mailto:[log in to unmask]>> on behalf of Taha Abdullah <[log in to unmask]<mailto:[log in to unmask]>> Reply-To: FSL - FMRIB's Software Library <[log in to unmask]<mailto:[log in to unmask]>> Date: Wednesday, 29 June 2016 at 22:59 To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> Subject: [FSL] Running First with 0.7 mm isotropic voxels Dear FSL team, I have a fairly large data set and the majority of the images have voxel dimensions in 1mm isotropic, however, there is a large number in which the voxel dimensions are 0.7 mm isotropic. When running run_first_all on 0.7 isotropic dimensions the values generated by fslstats were significantly overestimaed i.e. unilateral hippocampal volumes > 5000 thalamic volumes > 10000. All subject had an overestimation of volume. I have read some documentation on FIRST and as well as other methods to 'fix' the partial volume effect occurring within these 0.7mm isotropic voxel images. Yet, all still fail to correct for this overestimation. My most fruitful attempt was using *vtk files and using first_utils --meshToVol with the baseline image as MNI_152 or subjid_all_fast_firstseg or original T1 image and when running fslstats on these images I have fairly accurate volumes for the 0.7mm isotropic images with this mehod. However, I attempted this same method on the 1mm isotropic data and the values were now underestimated i.e. when running fslstats subjid_all_fast_firstseg.nii.gz -l low thresh -u upper thresh -V; the outputs for ex in the L_Hipp are 3500 (a reasonable volume), however, when using the *vtk files and running first_utils --meshToVol the output was approximately 2200. Could anyone please advise on the best/accurate method to extract correct subcortical ROI volumes for the 0.7 mm isotropic data? Best, Taha