Anatomic validation of spatial normalization methods for PET

Spatial normalization methods, which are indispensable for intersubject ana lysis in current PET studies, have been improved in many aspects. These met hods have not necessarily been evaluated as anatomic normalization methods because PET images are functional images, However, in view of the close rel ation between brain function and morphology, it is very intriguing how prec isely normalized brains coincide with each other. In this report, the anato mic precision of spatial normalization is validated with three different me thods. Methods: Four PET centers in Japan participated in this study. In ea ch center, six normal subjects were recruited for both (H2O)-O-15-PET and h igh-resolution MRI studies. Variations in the location of the anterior comm issure (AC) and size and contours of the brain and the courses of major sur d were measured in spatially normalized MR images for each method. Spatial normalization was performed as follows. (a) Linear: The AC-posterior commis sure and midsagittal plane were identified on MRI and the size of the brain was adjusted to the Talairach space in each axis using linear parameters. (b) Human brain atlas (HBA): Atlas structures were manually adjusted to MRI to determine linear and nonlinear transformation parameters and then MRI w as transformed with the inverse of these parameters. (c) Statistical parame tric mapping (SPM) 95: PET images were transformed into the template PET im age with linear and nonlinear parameters in a least-squares manner. Then, c oregistered MR images were transformed with the same parameters used for th e PET transformation. Results: The AC was well registered in all methods. T he size of the brain normalized with SPM95 varied to a greater extent than with other approaches. Larger variance in contours was observed with the li near method. Only SPM95 showed significant superiority to the linear method when the courses of major sulci were compared. Conclusion: The results of this study indicate that SPM95 is as effective a spatial normalization as H BA, although it does not use anatomic images. Large variance in structures other than the AC and size of the brain in the linear method suggests the n ecessity of nonlinear transformations for effective spatial normalization. Operator dependency of HBA also must be considered.