Magnetic-resonance morphometry: Image-analysis methodology development for affective disorder

In this article we review some important methodological issues for measurement of brain structures and lesions visualized on magnetic resonance (MR) imaging. Such issues include the method of sectioning, number of slices, shape of the structure, orientation of the slices, and fuzziness of the edges. Sources of error specific to morphometry obtained from MR scans are also considered, including issues of partial volume, contrast and itensity changes, and magnetic-field inhomogeneties. Each of these factors is important and should be borne in mind while undertaking MR studies for morphometry. In addition, a key element contributing to variability in the assessment of structures is the anatomical definition of specific brain regions. We briefly summarize the anatomical definitions that are useful in assessing many relevant brain areas and for evaluating signal hyperintensities that are frequently found in elderly patients with affective disorder. The reliability of assessment using these methods is noted. Studies of neuroanatomical changes associated with affective disorder will benefit from further improvement in the reliability and accuracy of MR morphometric methods. Depression 1:159–171 (1993). © 1993 Wiley-Liss, Inc. In this article we have reviewed some of the important methodological issues for measurement of brain structures and lesions visualized on MR. These issues should be borne in mind while undertaking MR studies for morphometry. In addition, a key element contributing to variability in the assessment of structures is the anatomical definition of specific brain regions. We have briefly summarized the anatomical definitions we have found useful in assessing many relevant brain areas and for evaluating signal hyperintensities that are frequently seen in elderly patients with affective disorder. The reliability of assessment using these methods is noted. In addition to these factors, image-acquisition methods can also affect measurement. However, these factors have not been well studied and are not addressed in this article. We have also not compared different image-analysis systems. This is an extremely important undertaking that has yet to be conducted. Future studies of neuroanatomical changes in affective disorder will benefit from further improvement in reliability and accuracy of MR morphometric methods. Deparession 1:159–171 (1993). © 1993 Wiley-Liss, Inc.