<div dir="ltr"><div><div><div><div>Hi Romain (and all),<br><br>Thanks for giving me access to you data so that I could investigate the problem.<br><br></div>So as Donald said previously, there's been a number of issues arise regarding the <span style="font-family:courier new,monospace">dwi2response</span> command in the new MRtrix. This program is almost identical to the method described in <a href="http://www.sciencedirect.com/science/article/pii/S1053811913008367">this paper</a> (with a few tweaks), which is recommended reading for anyone interested in spherical deconvolution. We are looking into alternative ways to estimate the response function, but these are not likely to appear overnight.<br>
<br></div>I have tweaked a couple of the default parameter values in <span style="font-family:courier new,monospace">dwi2response</span>, which should prevent the program from running for long periods of time. I had disabled certain mechanisms for terminating the optimisation, having found them unnecessary for my own data, but clearly they are necessary in other circumstances.<br>
<br></div>However, it is important to consider the relationship between the data you pass to this process, and the parameters you select. Of particular note is the maximum harmonic degree used in the response function estimation. By convention, if you do not specify <i>lmax</i> explicitly, MRtrix will select the largest <i>lmax</i> that is supported by your image data, based on the number of unique diffusion sensitisation directions in the b-value shell being used. Exclusively, those who have reported problems with <span style="font-family:courier new,monospace">dwi2response</span> have been using data with a low b-value and a large number of directions. This poses two problems. Firstly, a large number of directions will result in automatic selection of a large <i>lmax</i>, which makes both the initial SH transformation and CSD considerably more computationally expensive. Secondly, at low b-values, there is simply <i>no power present</i> in these higher harmonic orders; the magnitude of diffusion sensitisation is not adequate to introduce such high-frequency components into the measured signal profile. I would highly recommend reading <a href="http://onlinelibrary.wiley.com/doi/10.1002/nbm.3017/abstract">this article</a> to see the evidence of why this is the case.<br>
<br></div>So, for anyone using the <span style="font-family:courier new,monospace">dwi2response</span> command, I have the following recommendations:<br><ul><li>Update your MRtrix3 install using "<span style="font-family:courier new,monospace">git pull; ./build</span>"; this will change a couple of the default parameters in <span style="font-family:courier new,monospace">dwi2response</span> and hopefully prevent very long execution times.<br>
</li><li>Use the <span style="font-family:courier new,monospace">-mask</span> option! You're not going to use any voxels outside the brain to determine the single-fibre response function, so no point in processing them.</li>
<li>Set the <span style="font-family:courier new,monospace">-lmax</span> option manually based on <i>both </i>the number of directions in the b-value shell you are using<i> and</i> the b-value of that shell. Particularly if you get the "ERROR: matrix must be positive definite" error: this means your <i>lmax</i> is <i>waaay</i> too high.</li>
</ul><p>Normally I'd say "happy tracking", but I guess in this case it should be "happy RF estimation"!</p><p>Rob<br></p></div><div class="gmail_extra"><br clear="all"><div><div dir="ltr"><br>--<br>
<br><span style="color:rgb(255,102,0)"><b>Robert Smith, Ph.D</b><br>Research Officer, Imaging Division</span><br><br>The Florey Institute of Neuroscience and Mental Health<br>Melbourne Brain Centre - Austin Campus<br>245 Burgundy Street<br>
Heidelberg Vic 3084<br>Ph: +61 3 9035 7128<br>Fax: +61 3 9035 7301<br><a href="http://www.florey.edu.au/" target="_blank">www.florey.edu.au</a><br></div></div>
<br><br><div class="gmail_quote">On Tue, Jun 17, 2014 at 3:15 AM, Donald Tournier <span dir="ltr"><<a href="mailto:jdtournier@gmail.com" target="_blank">jdtournier@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>
<p dir="ltr">Hi Romain,</p>
<p dir="ltr">Yes, we've had reports of people having trouble with the new dwi2response app. This is something we're going to be looking into in the near future...
</p>
<p dir="ltr">Cheers,<br>
Donald</p>
<p dir="ltr">--<br>
Dr J-Donald Tournier (PhD)</p>
<p dir="ltr">Senior Lecturer, Biomedical Engineering<br>
Division of Imaging Sciences & Biomedical Engineering<br>
King's College London</p>
<p dir="ltr">A: Department of Perinatal Imaging & Health, 1st Floor South Wing, St Thomas' Hospital, London. SE1 7EH<br>
T: <a href="tel:%2B44%20%280%2920%207188%207118%20ext%2053613" value="+442071887118" target="_blank">+44 (0)20 7188 7118 ext 53613</a><br>
W: <a href="http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering" target="_blank">http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering</a><br>
</p><div><div class="h5">
<div class="gmail_quote">On 16 Jun 2014 16:49, "romain valabregue" <<a href="mailto:romain.valabregue@upmc.fr" target="_blank">romain.valabregue@upmc.fr</a>> wrote:<br type="attribution">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hello<br>
<br>
I get an error on several dataset with de dwi2response :<br>
<br>
dwi2response -grad grad.b -mask mask_erode3.nii -sf sf_out.nii rfsl_eddycor.nii resp.txt<br>
<br>
dwi2response: initialising response function... 100%<br>
dwi2response: optimising response function.. . gsl: lib/math/cholesky.cpp:151: ERROR: matrix must be positive definite<br>
Default GSL error handler invoked.<br>
Abandon<br>
<br>
It takes quite a long time before getting the error (about 1/2 an to 1 hour)<br>
<br>
<br>
Although it works ok with connectome B3000 data, I have problem with other data set that was working before with mrtrix2<br>
<br>
Many thanks<br>
<br>
Romain<br>
<br>
______________________________<u></u>_________________<br>
Mrtrix-discussion mailing list<br>
<a href="mailto:Mrtrix-discussion@www.nitrc.org" target="_blank">Mrtrix-discussion@www.nitrc.<u></u>org</a><br>
<a href="http://www.nitrc.org/mailman/listinfo/mrtrix-discussion" target="_blank">http://www.nitrc.org/mailman/<u></u>listinfo/mrtrix-discussion</a><br>
</blockquote>
</div>
</div></div></div>
</blockquote></div><br></div>