Fully Automatic Baseline Correction in Magnetic Resonance Spectroscopy

Abstract

Proton Magnetic Resonance Spectroscopy (1H MRS) in conjunction with Magnetic Resonance Imaging (MRI) has been a significant topic of research for quantitative assessment and early detection of neurodegenerative disorders for more than two decades. However, robust techniques for MRS data analysis are still being developed for wide clinical use. Many neurodegenerative diseases exhibit changes in concentrations of specific metabolites. One of the challenging problems in developing consistent quantitative estimation of metabolite concentration is proper correction of the MRS baseline due to the contributions from macromolecules and lipids. We have proposed a novel approach based on interpolation of minima in MR spectra and applied this technique to both in vitro and in vivo MRS data analysis. Our results demonstrate that the proposed method is fast, independent of tuning, and provides an accurate estimation of MRS baseline, leading to improved computational estimates for metabolic concentrations.