Figure 9
From: Distortion correction using topup algorithm by single k-space (TASK) for echo planar imaging
Quantitative comparisons of the corrected images in all subjects and comparisons of the estimated field maps in a representative subject. (a) Dice coefficient (five center slices), (b) Hausdorff Distance (five center slices), (c) CBF (center slice), (d) perfusion SNR (center slice) and (e) perfusion tSNR (center slice) were measured in the EPI images processed by TASK (TASK, light gray), double-echo GRE method (double-echo GRE, white), and in the linear EPI images (Linear, dark gray). The gray matter mask was used to calculate the average values of the gray matter area. The ground truth brain boundary from GRE image was the reference of the Dice coefficient and Hausdorff distance. For fair comparison, the linear EPI images were not processed through distortion correction under the assumption of no additional B0 field mapping. Two-tailed t-test verified the statistical significance. (f) The field maps obtained by the TASK, double-echo GRE method, and conventional topup (Topup) of subject #2 are shown. (g) The scatter plot of the voxel distribution of the estimated field maps between the center slices of “Topup” and “TASK” from all subjects shows a strong correlation (\(r =0.80\pm 0.05\)).
