EIGENVECTOR-BASED SPECTRAL ENHANCEMENT OF NUCLEAR-MAGNETIC-RESONANCE PROFILES OF SMALL VOLUMES FROM HUMAN BRAIN-TISSUE

Nuclear Magnetic Resonance (NMR) spectroscopy is a low-energy technique which suffers from poor inherent signal-to-noise ratio (SNR). In a clinical setting, it is often desirable to study small regions of tissue in patients to aid in the detection and diagnosis of disease states. Analysis of the smaller regions, however, degrades the SNR further and renders conventional spectral estimation techniques such as the discrete Fourier transform useless. We demonstrate the utility of two complex eigenvector-based algorithms, Multiple Signal Classification (MUSIC) and Minimum Norm, in the detection of resonances within small sample volumes. The results indicate that these methods are clearly superior to Fourier transform-based techniques currently available on clinical NMR scanners.