External Email - Use Caution
Dear freesurfer community,
I have been trying to experiment a bit with the dmri_pathstats command that TRACULA uses to compute tract statistics. I read previously that one could change the default threshold value that stats are computed with by using the "--pthr .##" option. (I was curious about using this since I often run into the peculiar problem of some tracks, e.g. forceps major/minor, having mostly normal looking path.pd maps, with a few voxels having unusually high values, thus potentially biasing results. These tracks are often set at abnormally high thresholds when visualizing the tracks. e.g. a forceps major normally has a threshold [30,45], while some may be thresholded as high 100 or 150, making the tract look problematic).
I experimented by running the dmri_pathstats command with a few different values for "--pthr .##":
* --pthr .01 * --pthr .05 * --pthr .10 * --pthr .20 * --pthr .80 * --pthr 1.00
However, the outputted "pathstats.overall.txt" and "pathstats.byvoxel.txt" files are all identical. Suspecting a potential bug, I did a few more tests, with values such as:
* --pthr 2.0 * --pthr 100 * --pthr 50000000000000 * --pthr hi * --pthr goodbye * --pthr
Interestingly, all of the outputted "pathstats.overall.txt" and "pathstats.byvoxel.txt" files were all identical to the previous six tests. The program does not return an error message either. Is this a potential bug? Or was the "--pthr" functionality removed from dmri_pathstats?
Incidentally, I am using freesurfer version 5.3.0. Maybe this was updated in a further release?
Thank you for any clarification!
Sincerely,
Adam Schadler