Transformed Image Standardization

There are several points to consider in standardizing results of spatial normalization field of view, orientation of the x-y-z axes, and location of the AC. Standard output can be achieved by storing all results in a fixed-size 3D image array (fixed field of view and spacing for x, y, and z directions) with the AC at a standard array location. The advantage of this approach is that subsequent display and analysis is simplified since all images are in the same format. An alternative method is to store only the information needed to reconstruct a transformed image. The former method is simpler to use, but it requires that copies of source and transformed images be maintained. The later method is more general and does not require storage of transformed images, but it requires special software to create transformed images when needed. For this reason most laboratories use a standard 3D image array for spatially normalized images.

A separate standard is needed for high-resolution MR images. General practice is to use x-y-z spacing of 2 mm for low-resolution (functional) images and 1 mm for high-resolution (anatomical) images. The FOV and mm location of the AC can remain the same, but the 3D arrays will be 8 x larger for the high-resolution MR images. One recommendation is that the x-y-z dimension be 256 x 170 x 150 mm to ensure that all brains fit into a standard physical volume [35]. It was also suggested and that the origin be 128 mm from the left, 87 mm from the top, and 107 mm from the front in order to center the brain in the standard volume. Several other locations and array sizes/ image volumes have been used [37].

0 0

Post a comment