COMPONENT SEPARATION IN NMR IMAGING AND MULTIDIMENSIONAL SPECTROSCOPYTHROUGH GLOBAL LEAST-SQUARES ANALYSIS, BASED ON PRIOR KNOWLEDGE

Use of prior knowledge with regard to the number of components in an i mage or NMR data set makes possible a full analysis and separation of correlated sets of such data, It is demonstrated that a diffusional NM R microscopy image set can readily be separated into its components, w ith the extra benefit of a global least-squares fit over the whole ima ge of the respective diffusional rates. As outlined, the computational approach (CORE processing) is also applicable to various multidimensi onal NMR data sets and is suggested as a potentially powerful tool in functional MRI. (C) 1998 Academic Press.