This should just be a difference between inclusive and exclusive thresholds ( if your image has an integer data-type you could confirm this by off-setting the intensity thresholds by 1 ).
Hope this helps,
Dr Matthew Webster
John Radcliffe Hospital
University of Oxford
> On 12 Aug 2017, at 17:07, Antonios Konstantinos <[log in to unmask]> wrote:
> Hello fsl users,
> I am trying to figure out why I get a discreapancy in the result of fslstats thresholding and
> the fslmaths thresholding.
> Here is my example:
> fslstats image.nii.gz -P 5 -P 95 # Give me the 5 and 95 percentiles of the image (for the non zero voxels)
> 6.000000 1220.000000
> fslstats image.nii.gz -M -l 6 -u 1220 -M # Use these percentiles to threshold the image and calculate the new -M thresholded value and see how it differs from the non thresholded one.
> 492.315100 473.398263
> fslmaths image.nii.gz -thr 6.000000 -uthr 1220.000000 image_thresholded.nii.gz # Do now the same with fslmaths
> fslstats image_thresholded -M #Calculate the new -M value
> 439.634478 # I was expecting that this value would be 473.398263 as fslstats gave.
> What am I missing here?
> Thank you