The objective of this study is to investigate the use of pattern classification methods for distinguishing different types of brain tumors, such as primary gliomas from metastases, and also for grading of gliomas. The availability of an automated computer analysis tool that is more objective than human readers can potentially lead to more reliable and reproducible brain tumor diagnostic procedures. A computer-assisted classification method combining conventional MRI and perfusion MRI is developed and used for differential diagnosis. The proposed scheme consists of several steps including region-of-interest definition, feature extraction, feature selection, and classification. The extracted features include tumor shape and intensity characteristics, as well as rotation invariant texture features. Feature subset selection is performed using support vector machines with recursive feature elimination. The method was applied on a population of 102 brain tumors histologically diagnosed as metastasis (24), meningiomas (4), gliomas World Health Organization grade II (22), gliomas World Health Organization grade III (18), and glioblastomas (34). The binary support vector machine classification accuracy, sensitivity, and specificity, assessed by leave-one-out cross-validation, were, respectively, 85%, 87%, and 79% for discrimination of metastases from gliomas and 88%, 85%, and 96% for discrimination of high-grade (grades III and IV) from low-grade (grade II) neoplasms. Multiclass classification was also performed via a onevs-all voting scheme.

Diffusion-tensor fiber tracking was used to identify the cores of several long-association fibers, including the anterior (ATR) and posterior (PTR) thalamic radiations, and the uncinate (UNC), superior longitudinal (SLF), inferior longitudinal (ILF), and inferior fronto-occipital (IFO) fasciculi. Tracking results were compared to existing anatomical knowledge, and showed good qualitative agreement. Guidelines were developed to reproducibly track these fibers in vivo. The interindividual variability of these reconstructions was assessed in a common spatial reference frame (Talairach space) using probabilistic mapping. Complex cognitive and behavioral processes that involve different functional areas of the brain are mediated by neural networks. The ability to identify and characterize the axonal fiber bundles comprising these networks is important for understanding normal as well as pathological processes affecting higher cerebral functions. Most studies of axonal fiber tracts to date have relied on invasive in vivo techniques in animals or ex vivo postmortem human tissue analyses (1,2). In addition, noninvasive radiological techniques, such as conventional MRI, have been able to parcellate only small portions of the white matter into specific tracts in restricted brain regions (3,4). Consequently, knowledge concerning these pathways in vivo is based mainly on primate-human extrapolations (2). In the last decade, diffusion imaging has been shown to be directionally dependent in the white matter (anisotropic diffusion) (5-7), which has been attributed to the organization of axonal fibers and their myelin sheaths. The magnitude and orientation of this anisotropy can be assessed using diffusion-tensor imaging (DTI) (8 -11), which has shown that brain regions with a high density of axonal fibers (e.g., deep white matter) have a high anisotropy. Using DTI and newly developed data processing (tract-tracing) techniques, we have recently shown the capability to perform delineation and 3D reconstruction of the cores of some axonal fiber bundles in animals (12,13), resulting in MR images that agree with standard anatomical data. Further application of this methodology to the study of healthy volunteers has confirmed the feasibility of this emerging technique for human studies (14 -19). However, several crucial steps are still necessary to allow a comprehensive use of this technique for the assessment of the anatomy of human connectivity. These steps include the establishment of procedures to reconstruct reproducibly the fiber trajectories in different individuals, and the validation of their origin through comparison with existing neuroanatomical knowledge. Although the latter validation can only be qualitative for human studies, it remains essential. This work describes our first efforts towards this goal, in which we illustrate the capability of the 3D tracking technique to identify the cores of several long-association fibers. The interindividual variability of these reconstructions is subsequently assessed in a common ...

Diffusion kurtosis imaging provides independent and complementary information to that acquired with traditional diffusion techniques. The additional information is thought to indicate the complexity of the microstructural environment of the imaged tissue and may lead to broad-reaching applications in all aspects of neuroradiology.

Spatial normalization of diffusion tensor images plays a key role in voxel-based analysis of white matter (WM) group differences. Currently, it has been achieved using low-dimensional registration methods in the large majority of clinical studies. This paper aims to motivate the use of high-dimensional normalization approaches by generating evidence of their impact on the findings of such studies. Using an ongoing amyotrophic lateral sclerosis (ALS) study, we evaluated three normalization methods representing the current range of available approaches: low-dimensional normalization using the fractional anisotropy (FA), high-dimensional normalization using the FA, and high-dimensional normalization using full tensor information. Each method was assessed in terms of its ability to detect significant differences between ALS patients and controls. Our findings suggest that inadequate normalization with low-dimensional approaches can result in insufficient removal of shape differences which in turn can confound FA differences in a complex manner, and that utilizing high-dimensional normalization can both significantly minimize the confounding effect of shape differences to FA differences and provide a more complete description of WM differences in terms of both size and tissue architecture differences. We also found that high-dimensional approaches, by leveraging full tensor features instead of tensor-derived indices, can further improve the alignment of WM tracts.