Pattern recognition approaches in biomedical and clinical magnetic resonance spectroscopy: a review

In recent years considerable effort has been devoted to applying pattern recognition techniques to the complex task of data analysis in magnetic resonance spectroscopy. It may be argued that such techniques will facilitate putting MRS technology to practical clinical use. This paper reviews approaches of pattern recognition commonly used in the analysis of MR spectra for biomedical applications. It briefly introduces the mathematical and algorithmic formulation of each of the techniques, noting their developmental background and their relationship to each other, and discusses their strengths and limitations. It then reviews how these techniques have been implemented in MRS applications. In doing so the paper also highlights a number of problems related to the design and testing of MRS/pattern recognition applications which currently prevent these techniques from being in wide practical clinical use, and suggests ways to avoid those pitfalls.