[Anonymous]

1. What are the supported (input) file formats?
   • BDP works best with input image volume in NIfTI format. It also provides some limited support for DICOM images. See BDP file-format for more details.
2. What are system requirements for BDP?
   • See “System Requirements” section in BDP technical details. In addition note that system requirement for running tractography is independent from that of running BDP and can change depending on diffusion model and dataset used.
3. Where can I download sample datasets?
   • Only diffusion data – click here
   • Diffusion data along with extracted surfaces and SVReg labels – click here
   • <small>Note about the above dataset: All the datasets were produced from the Beijing Enhanced data, Beijing Normal University, State Key Laboratory of Cognitive Neuroscience and Learning Enhanced Sample, which is available under a Creative Commons Attribution – Non-Commercial license (CC-BY). For more details please see: http://fcon_1000.projects.nitrc.org/indi/retro/BeijingEnhanced.html.</small>
4. What types of diffusion MRI data can be used with BDP? What is the convention of diffusion gradient direction (.bvec file)?
   • See BDP technical details for detailed description of supported diffusion MRI datasets and diffusion gradient directions.
5. Where are the flag description?
   • Here
6. What does “RAS” in file names mean?
   • BDP saves most image volumes in NIfTI-1 format with “RAS” convention i.e. the image data is stored such that the most rapidly increasing voxel index increases along Right side of subject, followed by Anterior side and then by Superior side of the subject. More precisely, the image is saved in an orientation which is closest to the ‘canonical space‘ and the (s-form) rotation matrix in NIfTI header is set as close to identity matrix as possible (ignoring scaling).
7. What are the unit of reported quantitative diffusion parameters in output images (MD etc.) and statistics files?
   • BDP reports diffusivity (MD, axial, radial, L2 and L3) in a unit which is reciprocal inverse of the units of input b-value specified in .bval file (or dicoms).
8. What are different types of distortion correction supported by BDP?
   • There are three different modes of distortion correction (1) Registration based distortion correction, (2) Fieldmap based distortion correction, and (3) No distortion correction. See Co-registration page for more detailed description.
9. Does BDP Co-registration correct the input diffusion data for eddy-current or motion artifacts?
   • As of now, BDP does not attempt to correct the input data for any eddy-current or motion artifact. If your dataset is affected by these artifacts then you may have to first correct your dataset for these artifacts outside BrainSuite. We plan to develop and include these correction methods in future.
10. Can I use diffusion MRI data split over multiple runs/acquisitions with BDP?
    • If you are using DICOM images as input you can use -d flag to specify a series of directories containing scans from multiple runs (each directory must contain images from a single unique run). See usage and flag description page for more details.
    • If you are using NIfTI-1 images as input, you have to first combine images (and corresponding gradient and b-value files) from multiple scans into one single 4D NIfTI-1 file. mergeDWIs from BDP Add-ons can be used for this purpose.

[JonatanSnir]

With respect to item #9, can I import data files after preprocessing with FSL – this would include BET, Eddy current corrections, registration to standard space, TBSS, DTIFIT, BEDPOSTX etc.. I hear great things about the brainsuite and would very much like to use it. If there is a proper way to prepare the output data from FSL preprocessing pipeline to fit the use of BDP could you please direct me to the instructions?
Thanks kindly,
Jon

[sychoi]

Sorry for the late response.

It depends on which components you want to you FSL for and which components you want to you BDP for.

Of the things you named, I would suggest only running the eddy current correction and then you can use that output from FSL as your input for BDP.

[micaceous]

Hello,

Following the co-reg of diffusion space to T1 space, are the bvectors reoriented in proportion to the degree of deformation?

Thank you,
Micah

[Anonymous]

Following the co-reg of diffusion space to T1 space, are the bvectors reoriented in proportion to the degree of deformation?

When transferring diffusion data to T1-space, b-vectors are reoriented using the computed rigid mapping between the spaces. The non-rigid EPI deformation (used for distortion-correction) of the diffusion data does not affect diffusion encoding directions, hence is not used for reorientation of b-vectors.

[micaceous]

Perfect, thank you!

[micaceous]

I just realized that Bhushan et al. (2012) is your manuscript, very cool, I enjoyed reading it! I’m curious to know, because I don’t quite understand everything in it…given that, in my current situation, the voxel size transformation going from larger DWI (1x1x2mm) to smaller T1 (1x1x1mm) practically doubles the in-plane resolution of DWI, does the interpolation essentially create new diffusion data that didn’t exist before? If so, how does it do that? (sorry I know that’s a vague question, but perhaps there’s a paper you could point me to?). I’m guessing it doesn’t duplicate the information in a single 2mm DWI slice into the two 1mm T1 slices right?

I’m wondering this because I’ve always previously performed tractography and analysis in native diffusion space, and so I’m curious to know how doing it in native T1 space will affect it. Perhaps there are some specific disadvantages or trade-offs? But even if so, I have such bad distortion in this dwi data (and I have no field map or double acquisition), that I think BrainSuite’s method will confer substantial improvement.

[jottino]

hi!

is it OK if I correct my dwi images for eddy currents in FSL, rotate my bvecs with fdt_rotate_bvecs.sh (FSL script) and then use that output (eddy corrected dwi volumes and rotated b-vecs) in BS-BDP for further processing?

I read a response in this same thread that says that bvecs are rotated (again) when registered to the T1-weighted image (.bfc.nii.gz).

Would my pipeline (FSL->BS) of preprocessing suppose any kind of problem? Thanks!!

[Author]

Posts