Search

Abstract: Diffusion MRI can be used to study the structural connectivity within the brain. Brain connectivity is often represented by a binary network whose topology can be studied using graph theory. We present a framework for the construction of weighted structural brain networks, containing information about connectivity, which can be effectively analyzed using statistical methods. Network nodes are defined by segmentation of subcortical structures and by cortical parcellation. Connectivity is established using a minimum cost path (mcp) method with an anisotropic local cost function based directly on diffusion weighted images. We refer to this framework as Statistical Analysis of Minimum cost path based Structural Connectivity (SAMSCo) and the weighted structural connectivity networks as mcp-networks. In a proof of principle study we investigated the information contained in mcp-networks by predicting subject age based on the mcp-networks of a group of 974 middle-aged and elderly subjects. Using SAMSCo, age was predicted with an average error of 3.7years. This was significantly better than predictions based on fractional anisotropy or mean diffusivity averaged over the whole white matter or over the corpus callosum, which showed average prediction errors of at least 4.8years. Additionally, we classified subjects, based on the mcp-networks, into groups with low and high white matter lesion load, while correcting for age, sex and white matter atrophy. The SAMSCo classification outperformed the classification based on the diffusion measures with a classification accuracy of 76.0% versus 63.2%. We also performed a classification in groups with mild and severe atrophy, correcting for age, sex and white matter lesion load. In this case, mcp-networks and diffusion measures yielded similar classification accuracies of 68.3% and 67.8% respectively. The SAMSCo prediction and classification experiments indicate that the mcp-networks contain information regarding age, white matter lesion load and white matter atrophy, and that in case of age and white matter lesion load the mcp-network based models outperformed the predictions based on diffusion measures. [Copyright &y& Elsevier]

Abstract: The ability to study changes in brain morphometry in longitudinal studies majorly depends on the accuracy and reproducibility of the brain tissue quantification. We evaluate the accuracy and reproducibility of four previously proposed automatic brain tissue segmentation methods: FAST, SPM5, an automatically trained k-nearest neighbor (kNN) classifier, and a conventional kNN classifier based on a prior training set. The intensity nonuniformity correction and skull-stripping mask were the same for all methods. Evaluations were performed on MRI scans of elderly subjects derived from the general population. Accuracy was evaluated by comparison to two manual segmentations of MRI scans of six subjects (mean age 65.9±4.4years). Reproducibility was assessed by comparing the automatic segmentations of 30 subjects (mean age 57.0±3.7years) who were scanned twice within a short time interval. All methods showed good accuracy and reproducibility, with only small differences between methods. The conventional kNN classifier was the most accurate method with similarity indices of 0.82/0.90/0.94 for cerebrospinal fluid/gray matter/white matter, but it showed the lowest reproducibility. FAST yielded the most reproducible segmentation volumes with volume difference standard deviations of 0.55/0.49/0.38 (percentage of intracranial volume) respectively. The results of the reproducibility experiment can be used to calculate the required number of subjects in the design of a longitudinal study with sufficient power to detect changes over time in brain (tissue) volume. Example sample size calculations demonstrate a rather large effect of the choice of segmentation method on the required number of subjects. [Copyright &y& Elsevier]

Abstract: A fully automated brain tissue segmentation method is optimized and extended with white matter lesion segmentation. Cerebrospinal fluid (CSF), gray matter (GM) and white matter (WM) are segmented by an atlas-based k-nearest neighbor classifier on multi-modal magnetic resonance imaging data. This classifier is trained by registering brain atlases to the subject. The resulting GM segmentation is used to automatically find a white matter lesion (WML) threshold in a fluid-attenuated inversion recovery scan. False positive lesions are removed by ensuring that the lesions are within the white matter. The method was visually validated on a set of 209 subjects. No segmentation errors were found in 98% of the brain tissue segmentations and 97% of the WML segmentations. A quantitative evaluation using manual segmentations was performed on a subset of 6 subjects for CSF, GM and WM segmentation and an additional 14 for the WML segmentations. The results indicated that the automatic segmentation accuracy is close to the interobserver variability of manual segmentations. [Copyright &y& Elsevier]