Unsupervised, automated segmentation of the normal brain using a multispectral relaxometric magnetic resonance approach

Abstract

The purpose of this study was the development and testing of a method for unsupervised, automated brain segmentation. Two spin-echo sequences were used to obtain relaxation rates and proton-density maps from 1.5 T MR studies, with two axial data sets including the entire brain. Fifty normal subjects (age range, 16 to 76 years) were studied. A Three-dimensional (3D) spectrum of the tissue voxels was used for automatic segmentation of gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) and for calculation of their volumes. Accuracy and reproducibility were tested with a three-compartment phantom simulating GM, WM, and CSF. In the normal subjects, a significant decrease of GM fractional volume and increased CSF volume with age were observed (P < 0.0001), with no significant changes in WM. This multispectral segmentation method permits reproducible, operator-independent volumetric measurements.

MRI-based volumetry of head compartments: Normative values of healthy adults

“…An evaluation of the ICV estimation procedure yielded an error for the IC volume of less than 1.5% (Hentschel and Kruggel, 2004). Phantom-based validation studies (Alfano et al, 1997) found a variance of 2.6 -3.2% between phantom volumes and segmented compartments. From retest experiments using a segmentation approach similar to the one applied here, the mean and the variance of the difference between volumes from different scans were 0.4% resp.…”

Automated Determination of Brain Parenchymal Fraction in Multiple Sclerosis

“…[10][11][12] Rather than using images, this new technique defines the tissue types of the brain as specific combinations of these 3 parameters and can thereby synthesize tissue maps, 13 similar to the Alfano method. [14][15][16] In this way, the major components of the ICV (gray matter, WM, and CSF) can be defined by physical properties rather than by relative image characteristics. The use of such tissue maps makes it possible to perform fully automated calculation of BPF without the need for manual postprocessing.…”

Quantitative analysis in clinical applications of brain MRI using independent component analysis coupled with support vector machine

“…This is often the case in MR image classification where many unknown brain tissue clusters may be present. Nevertheless, unsupervised methods are generally preferred because they appear to be successful in clinical environments for classification of normal brain data and often have clear characteristic knowledge (22). But, difficulties also arise in the pathological cases which contain unknown complex knowledge (23,24).…”