Application of DSP technique to nuclear magnetic resonance spectroscopy

The application of block processing, adaptive and eigenvector based methods to spectral estimation in nuclear magnetic resonance (NMR) spectroscopy is considered. These techniques have been used to enhance the spectra of complex in vivo NMR data records with varying signal-to-noise ratios. This has resulted in a substantial improvement of absorption spectrum estimates over that obtained by conventional Fourier methods that are currently available on clinical NMR instruments. Furthermore, these techniques have been effective in reducing baseline distortion. The algorithms discussed are available in Khoros, a multidimensional signal processing and visualization environment developed at the University of New Mexico.<>