Other Applications

The problem of spatial normalization does not appear only in the analysis of structural and functional images of the brain. It appears, more generally, in applications in which image data from different subjects must be combined and analyzed together. In particular, in order to merge and directly compare images from different individuals, morphologic variability must first be removed. More specifically, images from different subjects must be transformed spatially so that they all reside in the same coordinate system, with anatomically corresponding regions being in similar locations. Another representative case in which spatial normalization is necessary is in studying the relationship between spinal damage and associated pain or other clinical symptoms. Figure 13 shows two spine images of different individuals, as well as the elastic warping of one of the images that brings the spinal region into registration with the other. If this procedure is applied to a number of patients, then associations between the location of a lesion in the spinal region and clinical symptoms can be precisely quantified, and it can be used in surgical planning.

FIGURE 12 A spatially normalized positron emission tomography (PET) image, overlaid on the digitized atlas image used as the template. The spatial transformation was determined from a higher resolution anatomical image of the same subject, and was subsequently applied to the lower resolution PET image. The PET and the anatomic images were first brought into alignment by correcting for position and orientation differences. See also Plate 25.

FIGURE 12 A spatially normalized positron emission tomography (PET) image, overlaid on the digitized atlas image used as the template. The spatial transformation was determined from a higher resolution anatomical image of the same subject, and was subsequently applied to the lower resolution PET image. The PET and the anatomic images were first brought into alignment by correcting for position and orientation differences. See also Plate 25.

0 0

Post a comment