posted Sep 10, 2009 10:26 AM by Fang-Cheng Yeh
[
updated Jan 5, 2012 11:53 AM by Fang-Cheng Yeh
]
Q: Every time I run the tracking, the tractography is not the same.
A:
DSI studio performs sub-voxel seeding. Once
the seeding area is assigned, the program will place the seeds "within"
each voxel RANDOMLY.
Therefore, the result of tracking will vary even though the seeding area is still the same.
Q:I am wondering whether DSI studio can provide any information about the ADC?
A: 1.open the src file
2.In the reconstruction window, select DTI as the reconstruction
method, resulting a fib file.
3.The generated fib file contents the ADC for each voxel.
To access the ADC from the fib file:
1.open the fib file by DSI Studio. [File]->[Export all voxels]->[Export to MAT]
2.open the generated mat file by matlab. There is a matrix named "md",
which is the ADC value for each voxel.
Alternatively, the fib file is also a MAT file. You may rename it as a
mat file as open it by Matlab to access the ADC.
Q: The tractography coordinate goes negative. is it correct?
A: Negative coordinate can be found because the coordinate system is voxel-centered. For example, (0,0,0) is at the center of the voxel, and part of the space in that voxel has negative coordinate, such as the point (-0.1,-0.1,-0.1) is within the voxel (0,0,0).
|
posted Sep 27, 2008 12:17 PM by Fang-Cheng Yeh
[
updated Jul 7, 2009 9:44 AM
]
Q:Is it possible to export the ODF values?
A: Yes. First of all, check the "record ODF" checkbox on the
reconstruction window to export the ODF values to .fib files. Then you
can access the ODF values by Matlab. The detail is in the following
link.
http://dsi-studio.labsolver.org/Manual/export-data-to-matlab
Should there be any problems or question, please send me an email.
Q: I fed DSI Studio with my own DSI images, but the reconstruction result looks like there is no meaningful fiber at all...
A: The imported DICOM image are supposed to be arranged in the same order as the b-table I provided on the website. It is highly possible that your b-table does not match the order of the DICOM images. If the b-table is correct, another way to figure the problem is to observe the generated ODFs. DSI studio provided an interface to illustrate the reconstructed ODF, and it always helps me figure out what's wrong with the reconstruction process. On the reconstruction page, you can check the "Record ODF" checkbox to export the reconstructed ODF, then the .fib file will contain the ODF information for inspection. Also, I suggest that you try a more robust reconstruction method "GQI+Local maximum", which is a dataset-independent q-space reconstruction method (only on the lastest DSI-Studio problem). If the method doesn't work, I am pretty sure that either the b-table or the dataset has something wrong. |
posted Sep 27, 2008 11:49 AM by Fang-Cheng Yeh
[
updated Dec 29, 2008 6:39 AM
]
Q:What is the direction table used in DSI datset? Is the software able to handle any number of directions?
A:According on a paper addressing this issue*, the optimum parameters for DSI dataset is that the maximum b-value=4000 for 203-point grid sampling scheme. The sample b-table is uploaded as the attachement file. For DSI Studio to process your DSI images, the images must be arranged in the same order as the one in the uploaded b-table. Different number of directions are also supported.
*"Optimization of diffusion spectrum imaging and q-ball imaging on clinical MRI system", Li-Wei Kuo, Jyh-Horng Chen, Van Jay Wedeen and Wen-Yih Isaac Tseng.
Q:Does your software is hard-coded to a specific slice number or matrix size? What are the fields in the DICOM files that the software is looking for ? Does the program support other images arrangement other than montage?
DSI Studio is not hard-coded to a specific slice number or matrix size. It will parse the DICOM headers to look for the following parameters:
image width (0x0028,0x0011); // the image width
image height (0x0028,0x0010); // the image height
matrix size (0x0018,0x1310); // for example "128 0 0 128"
For DICOM files obtained from SIEMENS machines, DSI Studio assume the images are montage, and the slice number is obtained by (image width/matrix size)*(image height/matrix size). For images from GE, the slice number are set to the number of the b0 images.
|
|
