Figure 1.
Representative diffusion-weighted images with the 30 directions encoding scheme.
The images were from the same subject and slice location, with forward phase-encoding blips (left panel), and with reversed phase-encoding blips (right panel). NC30 had an artifactual signal pileup around the frontal base of the skull, which was corrected not in EC30 with trilinear or spline but in ET30. The EC30 with trilinear interpolation were blurred, compared with EC30 with spline interpolation.
Figure 2.
Representative diffusion-weighted images with the 60 directions encoding scheme.
The images were from the same subject and slice location. The NC60 with forward phase-encoding blips had an artifactual signal pileup around the temporal base of the skull, which was not corrected in EC60, but corrected in ET60. Again, the EC30 with trilinear were blurred, compared with EC30 with spline interpolation.
Figure 3.
Comparisons between NC and EC images with trilinear interpolation.
The white matter skeletons with the higher FA for NC30 or NC60 were shown in red/yellow, and those with higher FA for EC30 or EC60 with trilinear interpolation were shown in blue/lightblue. The upper and lower row showed paired comparisons between NC30 and EC30 with trilinear, NC60 and NC60 with trilinear interpolation, respectively. NC30 and NC60 had higher FA values in most white matter skeletons, compared with EC30 and EC60 with trilinear interpolation, except in the posterior limb of the right internal capsule in the upper row. These data were overlaid onto the MNI152_T1_1 mm template, with the mean FA skeletons shown in green. The significance level was set at a P value of <0.05 with FWE correction.
Figure 4.
Efficiency of correction schemes acquired with 30 directions diffusion encoding.
Tripled paired group comparisons with TBSS were conducted between NC30 and EC30 with spline interpolation in the upper, NC30 and ET30 in the middle, and EC30 with spline interpolation and ET30 in the lower row. The FA values for ET30 were significantly higher than those for NC30 or EC30 in most white matter skeletons (blue/lightblue). The FA values for EC30 with spline interpolation were significantly higher than those for NC30 (blue/lightblue), and lower than those for ET30 in most white matter skeletons. These data were overlaid onto the MNI152_T1_1 mm template, with the mean FA skeletons shown in green. The significance level was set at a P value of <0.05 with FWE correction.
Figure 5.
Efficiency of correction schemes acquired with 60 directions diffusion encoding.
Tripled paired group comparisons with TBSS were conducted between NC60 and EC60 with spline interpolation in the upper, NC60 and ET60 in the middle, and EC60 with spline and ET60 in the lower row. The FA values for ET60 were significantly higher than those for NC60 or EC60 in most of the white matter skeleton (blue/lightblue). In contrast, white matter skeleton with significantly higher (blue/lightblue) and lower (red/yellow) FA values for EC60 with spline interpolation than those for NC60 were observed in various areas. These data were overlaid onto the MNI152_T1 template, and the mean FA skeleton is shown in green. The significance level was set at a P value of <0.05 with FWE correction.
Figure 6.
Comparison of ET30 and ET60 images.
The FA values for ET60 were significantly higher than those for ET30 in most of the white matter, with slight left hemisphere predominance. These data were overlaid onto the MNI152_T1 template, and the mean FA skeleton is shown in green. The significance level was set at a P value of <0.05 with FWE correction.