Diffusion processing included geometric distortion correction, eddy current correction, Gibbs ringing removal, denoising, and registration of the diffusion and structural images. To improve geometric distortion correction and registration, an “imitation” T2-weighted image was constructed from each participant's T1-weighted image to better match the contrast of the b0 image. The imitation T2 is simply the T1 image (dark gray matter, bright white matter, and dark cerebrospinal fluid [CSF]), with the contrast adjusted to match that of a T2 image (bright gray matter, dark white matter, and bright CSF). See Supplementary Figure S1 for examples of images at each MRI processing step. Structural and diffusion preprocessing was carried out in AFNI and TORTOISE, respectively (Cox, 1996; Irfanoglu et al., 2018; Pierpaoli et al., 2010).

Diffusion processing included geometric distortion correction, eddy current correction, Gibbs ringing removal, denoising, and registration of the diffusion and structural images. To improve geometric distortion correction and registration, an “imitation” T2-weighted image was constructed from each participant's T1-weighted image to better match the contrast of the b0 image. Geometric distortion correction involved aligning the b0 image to the imitation T2, then correcting distortions using nonuniform B-spline grid sampling (Irfanoglu et al., 2011). Gradient vectors were rotated according to the eddy correction and registration. Before reconstruction, the quality of each data set was assessed by calculating the correlation between neighboring diffusion directions/volumes. If more than 10% of the data were excluded due to poor correlation (r < 0.9), the data set was excluded. All data were visually inspected after preprocessing to ensure adequate alignment.

Spin distribution functions for each voxel were obtained using GQI (Yeh et al., 2010). GQI was chosen due to the sensitivity to crossing fibers and because it is one of the few higher order diffusion models that can be applied to any diffusion sampling scheme that is balanced, that is, the isotropic voxels are reconstructed as an isotropic spin distribution function (SDF), which is checked during reconstruction (Yeh et al., 2010). Q-space diffeomorphic reconstruction (QSDR), an extension of GQI, involves aligning the subject's quantitative anisotropy map (QA; derived from SDFs) to a QA template in Montreal Neurological Institute space (HCP-842) using diffeomorphic mapping to allow both linear and nonlinear alignment, applying the inverse Jacobian to the subject space SDFs, and adjusting for scaling differences between the subject and template, conserving the amount of diffusion spins in the original data (Yeh and Tseng, 2011). Alignment quality was assessed by the correlation between subject and template SDFs. Default QSDR settings were used except for the number of fibers resolved, which was reduced to three due to the limited directions in the current data set.

Deterministic tractography (curvature limit = 45°, min length threshold = 30 mm, max length threshold = 300 mm, one round of topology-informed tract trimming) (Yeh et al., 2018), based on QA, was performed between cortical parcels. The QA threshold was automatically determined by 0.6*Otsu's threshold, which has been shown to be optimal for resolving true connections while limiting false positives (Maier-Hein et al., 2017; Otsu, 1979). Resulting adjacency matrices were binarized. Diffusion reconstruction and tracking were performed in DSI Studio.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.