Evaluating the Accuracy of Diffusion MRI Models in White Matter

• DOI: 10.1371/journal.pone.0123272
• Publication Date: 2015-04-16T18:03:29Z
• Authors: Ariel Rokem; Jason D. Yeatman; Franco Pestilli; Kendrick N. Kay; Aviv Mezer; Stefan van der Walt; Brian A. Wandell
• Keywords: Data set

Models of diffusion MRI within a voxel are useful for making inferences about the properties tissue and inferring fiber orientation distribution used by tractography algorithms. A model must fit data accurately. However, evaluations model-accuracy commonly models have not been published before. Here, we evaluate two main classes models. The tensor (DTM) summarizes as 3-dimensional Gaussian distribution. Sparse fascicle (SFM) summarize signal sum signals originating from collection fascicles oriented in different directions. We use cross-validation to assess at gradient amplitudes (b-values) throughout white matter. Specifically, each all matter voxels one set then predict second, independent set. This is first evaluation these In most DTM predicts more accurately than test-retest reliability; SFM higher reliability also model-accuracy, particularly measurements with (a) b-value above 1000 locations containing crossings, (b) regions brain surrounding optic radiations. has better parameter-validity: it estimates function (fODF) voxel, which tracking.