 |
 |
My experience is that the tractography results are dramatically different (worse) with 12x5 data vs 60x1 data. Peace, Matt. From: Colin Reveley <[log in to unmask]> Reply-To: FSL - FMRIB's Software Library <[log in to unmask]> Date: Saturday, August 3, 2013 6:59 PM To: <[log in to unmask]> Subject: Re: [FSL] The minimal required number of diffusion direction for tractrography study off the top of my head I'd start here http://radiographics.rsna.org/content/26/suppl_1/S205.short and look at what cites it. I believe there are still special circumstances (like getting very high spatial resolution very rapidly for e.g. comparison to susceptibility weighted maps at high field) in which 12 directions might be used. Maybe those people talk about it. It's atypical and/or well before my time. I doubt tractography is defensible. But if I was trying it my first idea would be to leverage the numex=6, and how repeatable the tracking is which actually would be well covered by behrens 2003 on the FSL wiki. best, Colin On 4 August 2013 00:30, Chou Paul <[log in to unmask]> wrote: > Dear Colin > > Could you provide me the literature about the number of direction to estimate > diffusion tensors ? I wish I can get some ideas from the literature, as you > said this study include a rare subject group (we also acquired this dataset > longitudinally) > > Best > > Paul > > > Date: Sat, 3 Aug 2013 20:01:04 +0100 > From: [log in to unmask] > Subject: Re: [FSL] The minimal required number of diffusion direction for > tractrography study > To: [log in to unmask] > > > Without contradicting matt (or your reviewer by the sound of it) you do at > least have the averaging and if very carefully justified possibly the > rotationally invariant (e.g. FA) stuff could still work out if the data has > enough other value (like a rare subject group). > > I wouldn't take my word for it, it just seems "possible". 30/32 is accepted as > necessary to estimate diffusion tensors even though one would think only 6. > There is literature. > > best, > > Colin > > > On 3 August 2013 08:12, Paul Chou <[log in to unmask]> wrote: >> Dear all FSL experts >> >> I have a DTI dataset which was acquired six years ago (patient vs control >> study). The acquisition parameter for this DTI dataset were as follows : b >> value = 1000; 13 directions; NEX=6 and voxel size= 2*2*2.2 mm. I performed >> the TBSS analysis and probabilistic tractography for this study. After long >> time of journal review, I got the feedback from one reviewer and he/she said >> the acquisition parameter for DTI direction is "extremely outdated" (only 13 >> directions in this dataset) and reject my research. Based on these "extremely >> outdated acquisition scheme" for DTI, he also suggested me to exclude any >> results and discussion of the tractography part which I performed in this >> study. Do any experts have any suggestion or recommendation on the minimal >> required number of diffusion direction for tractrography study? I wish I >> could still publish these results (especially for tractrography part) without >> further technical problem in other journal. >> >> >> Best >> >> Paul > >