In-vivo tissue characterization of brain by synthetic MR proton relaxation and statistical chisquares parameter maps

Proton spin density (N[H])- and relaxation time (T1 and T2)-weighted magnetic resonance (MR) images at different anatomical sections and/or orientations of the brain are routinely used for clinical MR diagnosis of various types of intracranial disorders and injuries. However, numerical information pertaining to the relaxation behavior of water in the brain is very difficult to be extracted and quantified form these conventional MR images. Using Carr-Purcell-Meiboom-Gill spin echo imaging sequences, multiple raw MR images of human and animal brains with selected values of repetition and echo times were acquired using a clinical 1.5-Tesla MR scanner. Using a non-interactive nonlinear regression algorithm, first-order (N[H], T1 and T2) and higher order (biexponential and distribution) T2 proton relaxation parameter maps, as well as a new set of statistical chi-square parameter maps of the brains were calculated from the raw MR images pixel-by-pixel. We propose that the use of calculated relaxation and chi-square maps may further improve the capability of MRI in clinical diagnosis and staging of intercranial disorders and injuries