Thanks for your answer! We are actually using the targets as waypoints, in order to evaluate the connectivity distributions. Your approach seems sound, but I'm not sure I understand how to convert the data to a probability or percent indexes? You mean the -thrp/P operations?
Would you say that is better or safer than stating an absolute value for the threshold? (We are currently testing with -thr 100, but mainly just to make sure the commands are doing what we are expecting)
On Thu, 3 May 2012 10:46:40 -0400, Cunningham, Dustin <[log in to unmask]> wrote:
>Are you using your target as a waypoint or classification target, or in other words are you visualizing the connectivity distributions (fibers) or classifying your seed? Either way I think you may want to convert everything to a probability index (or percent). You could then apply a threshold, i.e. 1% (to remove spurious fibers), to each individual subject then binarize the tracts. You could then average all the binary tracts from all the subjects and keep only those overlapping to a certain threshold i.e. only those tracts that were found in >75% of the subjects were kept.
>Just thinking out loud. Hope this helps.
>From: FSL - FMRIB's Software Library [[log in to unmask]] On Behalf Of Daniel Preciado [[log in to unmask]]
>Sent: Thursday, May 03, 2012 9:31 AM
>To: [log in to unmask]
>Subject: [FSL] FSLmaths add - Display range threshold
>We are doing a probabilistic tractography study with over 170 subjects, and we need to add together all individual probabilistic tractography results in one average file. We know how to do that with fslmaths, however we are not sure which is the optimal threshold to use as display range. We would rather be more conservative than liberal, at least for now to get a feeling of how the tracts look like. We are currently tracking fibres from the caudate n. as seed, and the dlPFC as targets, in case that helps.
>Do you have any suggestions you can give us about this matter?
>Thanks in advance!