Hi Susumu<br>I apologize for using your direct mail but I had some problems logging into the forum.<br>I have the following questions:<br>
<p class="MsoNormal"><br></p><p class="MsoNormal"><b>LDDMM</b></p>
<p><span><span>1)<span style="font:7.0pt "Times New Roman""> </span></span></span> We are first time users of LDDMM and and are encountering a problem when we use the single channel registration: "file not found: process.md5". Further, we are using the LDDMM to correct for distortions
in muscle DTI images, these images are acquired rather anisotropically (1.4mm
in plane and 5 mm thickness), so the requirement for isotropic voxel in LDDMM
is a stretch.<span> </span>Is there a LDDMM algorithm
as a 2D version available where we correct on slice by slice basis as the
maximum distortions are along the in-plane phase encode.<span> </span>Our T2 and b0 slices are matched acquisitions.<span> </span>If a 2D version is not available, would it be
OK to reslice the T2 and DTI volumes to be isotropic despite a highly
anisotropic acquisition?</p><b>Fiber Tracking</b>
<p class="MsoListParagraphCxSpLast">1)<span style><span style></span></span> When performing tracking in DTIStudio, is
there a way to obtain a track through a selected voxel.<span style> </span>The reason I ask is that when selecting even
a single voxel ROI, the number of tracks through it more than one; this is from
the brute force aspect of the algorithm, I guess.<span style> </span>We are trying to get muscle fiber pennation angles using
seed points starting near the aponeurosis.<span style>
</span>It will help us in trying to get the fiber that originates from the seed
point rather than the ROI at that seed point (or is there a non-brute force
available).<span style> <br></span></p><p class="MsoListParagraphCxSpLast"><span style>2) This feature (getting the fiber coordinates from a seed voxel) is also important for our brain fiber atlases where we are trying to evaluate the quality of tracking in atlases generated by different methods. </span><span style="font-size:11pt;line-height:115%;font-family:"Calibri","sans-serif"">One
method suggested is the sum of distances of equally spaced points along the fiber
tract starting from the seed point in atlas and subjects. The other method to
compare is the track end point divergence. When several tracks go through a
single voxel ROI, it makes comparisons difficult.<span style> <br></span></span></p><p class="MsoListParagraphCxSpLast"><br><span style="font-size:11pt;line-height:115%;font-family:"Calibri","sans-serif""><span style></span></span></p>
<p class="MsoListParagraphCxSpLast"><span style="font-size:11pt;line-height:115%;font-family:"Calibri","sans-serif""><span style>Thanks</span></span></p><p class="MsoListParagraphCxSpLast"><span style="font-size:11.0pt;line-height:115%;font-family:"Calibri","sans-serif""><span style>Usha<br>
</span></span></p><p class="MsoListParagraphCxSpLast"><span style><br></span></p>
<br><br><br><div class="gmail_quote">On Fri, Mar 15, 2013 at 4:27 PM, susumu mori <span dir="ltr"><<a href="mailto:susumu@mri.jhu.edu" target="_blank">susumu@mri.jhu.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
OK, Deepa,<div><br></div><div>I believe you did a typical DSI study. Because it would be of interest for many DtiStudio users, allow me to share my response to you, together with your images.</div><div><br></div><div>There are two important differences between DTI and DSI.</div>
<div>First, of course, DSI requires many more gradient orientations. In your case, you used 257 directions. In general, 100 orientation measurement takes more than 10 min. So, 257 orientations would take about 30 min. You repeated it twice. So, it must take 1 hour, if I'm not mistaken.</div>
<div>I once asked Van Wedeen about how many orientations would be preferable. He said, "even 300 is very sparse, if you map them on the surface of a sphere. So, the more the better" was his reply.</div><div>On the other hand, if you use a model like spherical decomvolution, there are papers demonstrating directions as few as 60 would give you decent estimation of crossing fibers. I don't have an answer to it.</div>
<div><br></div><div>Second, high-angular resolution imaging like DSI requires higher b values. This actually have more profound effects on your experiment design and image quality. DTI typically uses b = 700-1,000. The higher you go, the more signal you loose, more motion artifacts you get, and more eddy current distortion you suffer. On top of it, because each image is so noisy, detection and correction of artifacts become more difficult. DSI requires b = 3,000 - 30,000, which is much higher than DTI. </div>
<div><br></div><div>Some people believe DSI data naturally includes DTI data. Therefore, once you get DSI data, you wouldn't loose anything. However, that is not the case, because you suffer from SNR loss and more artifacts. This is why it is difficult to adopt DSI protocols for routine research and clinical studies.</div>
<div><br></div><div>I attached a nice paper by Derek discussing this issue. </div><div><br></div><div>Now, let's go back to your images. The b0 looks good. The DWIs are what you would expect from high b-value DWIs. Your images look very similar to Fig. 3 in Derek's paper with b=2,000-3,000. So your data are fine. They don't look like typical DTI data because I bet that the b-values you used are b = 2,000-3,000.</div>
<div><br></div><div>You have slight ghosting in the background, but it seems calibration error for your EPI phase correction; not a major problem, I think.</div><div><br></div><div>So, the bottom line is, your DICOM files were correctly read. The appearance of DWIs should be different between DSI and DTI protocols and your images are what you expect.</div>
<div><br></div><div>Hope this will help.</div><div><br></div><div>Susumu<br><br><div class="gmail_quote">On Fri, Mar 15, 2013 at 6:31 PM, Deepa Krishnaswamy <span dir="ltr"><<a href="mailto:deepa.krishnaswamy1@gmail.com" target="_blank">deepa.krishnaswamy1@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<div>I have attached a power point with a few images. </div><div>Yes I did read the data as Mosaic. There are 515 gradient directions in total, which is comprised of repeated 257 directions plus 1 b0. </div>
<div>I do know that DTIStudio does not support DSI data analysis, but isn't it still possible to fit the DTI model to DSI data? </div>
<div>I also just wanted to run through the LDDMM pipeline to get our data registered to the same space as the JHU MNI WM atlas, and then use our own tractography algorithms. </div><div>I hope this is possible, thanks. </div>
<div><br></div><div>Deepa <br><br><div class="gmail_quote">On Thu, Mar 14, 2013 at 10:20 AM, susumu mori <span dir="ltr"><<a href="mailto:susumu@mri.jhu.edu" target="_blank">susumu@mri.jhu.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Deepa,<div><br></div><div>Did you read the data as Mosaic?</div><div>Please send me the actual images in PowerPoint.</div>
<div><br></div><div>If you are using non off-the-shelf image protocol, it is possible that DtiStudio can't read the user-defined data format.</div>
<div>If you get a custom data acquisition method from somebody, I encourage you to talk with them about how to read the data. DtiStudio doesn't support DSI data analysis anyway.</div><div><br></div><div>Susumu<br><br>
<div class="gmail_quote"><div><div><div><div>On Tue, Mar 12, 2013 at 9:50 PM, Deepa Krishnaswamy <span dir="ltr"><<a href="mailto:deepa.krishnaswamy1@gmail.com" target="_blank">deepa.krishnaswamy1@gmail.com</a>></span> wrote:<br>
</div></div></div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div><div><div>
<div style="color:rgb(34,34,34);font-size:13px;font-family:arial,sans-serif">Hi, </div><div style="color:rgb(34,34,34);font-size:13px;font-family:arial,sans-serif">
I have a DSI dataset in MOSAIC format, along with the corresponding b btable. I noticed that after loading in the data, some of it appears extremely blurry/shifted. I know that the data actually does not look like this, as I have used other programs to view the raw data. Has anybody else encountered this problem? </div>
<div style="color:rgb(34,34,34);font-size:13px;font-family:arial,sans-serif">Thanks, </div><span><font color="#888888"><div style="color:rgb(34,34,34);font-size:13px;font-family:arial,sans-serif">
<br></div><div style="color:rgb(34,34,34);font-size:13px;font-family:arial,sans-serif">Deepa </div>
</font></span><br></div></div></div></div>_______________________________________________<br>
mristudio-users mailing list<br>
<a href="mailto:mristudio-users@mristudio.org" target="_blank">mristudio-users@mristudio.org</a><br>
<a href="http://lists.mristudio.org/mailman/listinfo/" target="_blank">http://lists.mristudio.org/mailman/listinfo/</a><br>
Unsubscribe, send a blank email to: <a href="mailto:mristudio-users-unsubscribe@mristudio.org" target="_blank">mristudio-users-unsubscribe@mristudio.org</a><br>
<br></blockquote></div><br></div>
</blockquote></div><br></div>
</blockquote></div><br></div>
<br>_______________________________________________<br>
mristudio-users mailing list<br>
<a href="mailto:mristudio-users@mristudio.org" target="_blank">mristudio-users@mristudio.org</a><br>
<a href="http://lists.mristudio.org/mailman/listinfo/" target="_blank">http://lists.mristudio.org/mailman/listinfo/</a><br>
Unsubscribe, send a blank email to: <a href="mailto:mristudio-users-unsubscribe@mristudio.org" target="_blank">mristudio-users-unsubscribe@mristudio.org</a><br>
<br></blockquote></div><br>