No abstract

No abstract

In diffusion weighted MRI, subject motion and brain pulsation lead both to signal drop-outs and image misalignment. Unsedated neonates, with their higher heart rate and propensity for motion are particularly prone to degraded scan quality that impairs diffusion tensor estimation. Retrospective registration and robust estimators are two methods that have previously been demonstrated to address motion and intensity outliers, respectively, in diffusion data. However, when taken together, the resampling of images to correct for misalignment can have the effect of averaging outlier voxels with uncorrupted voxels, thereby making outliers more difficult to detect. This article presents a method to remove outliers prior to resampling while taking misalignment into account so that this averaging of outliers with good data can be avoided. The proposed method is compared to other processing pipelines using simulations and data from unsedated preterm neonates. These results demonstrate advantages to the proposed method, particularly in subjects with high motion. Magn Reson Med 66:92-101,

The influence of diffusion time (⌬), gradient duration (␦), and TE on the appearance of high b-value q-space diffusion MR images of excised rat spinal cord (SC) was evaluated. The water signal decays in the white (WM) and gray matter (GM) were analyzed when the diffusion was measured perpendicular (Ќ) and parallel (||) to the fibers of the SC, using three different approaches: single-component q-space analysis, the biexponential model, and the bi-Gaussian fit of the displacement distribution profile. Probability and displacement contrast and anisotropy indices were calculated for the WM and GM. It was found that WM/GM contrast increases as the diffusion time is increased when diffusion is measured perpendicular to the long axis of the SC. At a diffusion time of 50 ms, when diffusion was measured parallel to the fibers of the SC, the displacement was found to be higher for GM as compared to WM. For this direction the WM/GM contrast increased when diffusion time was increased, although here the changes were much less pronounced than for the perpendicular direction. The WM/GM displacement contrast nearly disappears for a diffusion time of 150 ms, when diffusion is measured parallel to the fibers of the SC. As expected, the anisotropy indices were found to be higher in WM than in GM, and increased with the increase in diffusion time. Both ␦ and TE affected the extracted parameters. It was found that long ␦ and long TE overemphasizes the apparent slowdiffusing water component of the SC, which is also the more restricted one. It is demonstrated that the single-component q-space analysis best describes diffusion in WM when diffusion is measured perpendicular to the fibers of the SC. In other cases, a more complete description is obtained by using twocomponent models. Magn Reson Med 54:96 -104, 2005.