| T O P I C R E V I E W |
| vvs4 |
Posted - Apr 17 2010 : 12:18:25 PM
Hi,
I want to use the Diffusion II toolbox in SPM8 to process DTI, and then I want to batch it. I've already been able to do it in FSL, and am now moving to SPM because my lab is using it to process fMRI.
I don't think that I need really specific instructions, just a generalized idea of the order of operations of things. The DTI data that I analyzed in FSL was actually not from Duke - and what I basically did is use MRIcron with the dicoms to spit out the bvec, etc files, and then I followed the exact steps outlined in FSL (eddy correction, brain extraction, etc)
So perhaps a good place to start is understanding the output from the scanner - a good description of what exactly is in the series008 (images), series800(average DC), series801(FA), and series802(isotropic image) folders. I've only gone as far as to do a dicom import into SPM, and my 1088 dicom images in the series008 folder jumped down to 32, and I only have a couple of files for output for the others. It's not clear to me if the scanner does some of the preprocessing, resulting in all the folders mentioned above.
The only instructions that I could find were for SPM2:
http://alliance.seas.upenn.edu/~pauly2/wiki/index.php?n=Main.Camino-SPM
Scroll down to step by step processing.
I know that mostly everyone uses FSL, but if anyone can point me towards a good person, a good resource, or give some insight as to the scanner output, it would be greatly appreciated!
Best,
Vanessa |
| 3 L A T E S T R E P L I E S (Newest First) |
| petty |
Posted - May 06 2010 : 3:12:24 PM
there is a <bvalues> tag in the bxh header that contains all your bvals in order.
also I've never used SPM for dti, but i would assume you should only be processing the original DTI series ( series00# ), since the rest could be generated from that. |
| tharsh |
Posted - Apr 18 2010 : 9:26:24 PM
Hi Vanessa...
I've not used SPM to analyze DTI images, so I won't answer any questions there. And it seems like you've actually figured out everything else! Those are the gradient directions (in the order acquired). Your b-factor is probably 1000, I think that should be in the bxh header somewhere (maybe near the bottom).
The directories are what you said they are- the first contains the images (number of slices * number of directions, including the b0), the others are an ADC map, an FA map, and an average image, all created on the scanner (most DTI procesing software will also create these maps from the base images).
Not sure what else you need to know. If you're looking for uses for the FA and ADC maps, well, there will be some literature out there, and there are some clinical uses...
todd
|
| vvs4 |
Posted - Apr 18 2010 : 1:59:19 PM
Here is what I've figured out so far:
* In the main DTI (series900) images folder, I have 1088 files. When I import the dicoms into SPM, I get 16. I think this is because we have 68 slices in 15 directions (68 X 15) = 1020, and then add the 68 slices for the b0 images, and we get a total of 1088.
* I was reading on an old posting from 2007 that the "typical BIAC DTI with 15 directions" has a b value of 800. But I am looking at our scan parameters, and I see a B value of 1000. Can anyone confirm this for BIAC5?
* I then looked at the series009 bxh file, and found these little lovelies:
<datapoints>
<value>0 0 0</value>
<value>1 0 0</value>
<value>0.643 0.766 0</value>
<value>0.258 0.307 0.916</value>
<value>0.745 -0.594 0.303</value>
<value>0.164 -0.507 0.846</value>
<value>-0.796 -0.321 0.513</value>
<value>0.761 0.427 0.489</value>
<value>-0.506 0.833 0.224</value>
<value>0.667 -0.158 0.728</value>
<value>0.128 -0.959 0.254</value>
<value>-0.178 -0.898 -0.403</value>
<value>0.255 -0.59 -0.767</value>
<value>-0.34 -0.736 0.585</value>
<value>-0.801 0.329 0.501</value>
<value>0.336 0.043 -0.941</value>
</datapoints>
These must be the gradient directions!
In SPM with the Diffusion toolbox:
* I did dicom to nifti import
* I then did Re(set)DTI Information, and input the 16 images, the b values, and gradient directions.
I don't feel like I know enough to move forward beyond this, I still don't quite understand what is in the other folders from the scanner. I'm going to try and learn a little more and then continue.
Best,
Vanessa |
|
|
