Multivariate Image Analysis of Magnetic Resonance Images with the Direct Exponential Curve Resolution Algorithm (DECRA): Part 1: Algorithm and Model Study

Abstract

Antalek and Windig recently presented a fast method to resolve a series of NMR mixture spectra, where the contribution of the components varies with a decaying exponential [B. Antalek and W. Windig,J. Am. Chem. Soc.118, 10,331–10,332 (1996); W. Windig and B. Antalek,Chemom. Intell. Lab. Syst.37, 241–254 (1997)]. The method was called DECRA (direct exponential curve resolution algorithm). In this paper DECRA will be applied to two series of magnetic resonance images. The signal of one series is based uponT2relaxation, and the other is based uponT1relaxation. In order to evaluate the technique, the magnetic resonance images of a phantom where used. A transformation is introduced to enable the application of DECRA to aT1series of magnetic resonance images. A separate paper in this issue will describe the application of the techniques to magnetic resonance images of the human brain.