NMRES - AN ARTIFICIAL-INTELLIGENCE EXPERT-SYSTEM FOR QUANTIFICATION OF CARDIAC METABOLITES FROM 31PHOSPHORUS NUCLEAR-MAGNETIC-RESONANCE SPECTROSCOPY

The application of high-resolution Phosphorus-31 Nuclear Magnetic Reso nance (P-31 NMR) Spectroscopy in biology and medicine has provided new insights into biochemical processes and also a unique assessment of m etabolites. However, accurate quantification of biological NMR spectra is frequently complicated by: (a) non-Lorentzian form of peak line-sh apes, (b) contamination of peak signals by neighboring peaks, (c) pres ence of broad resonances, (d) low signal-to-noise ratios, and (e) poor ly defined sloping baselines. Our objectives were to develop an expert system that captures and formalizes P-31 NMR spectroscopists' expert knowledge, and to provide a reliable, efficient, and automated system for the interpretation of biological spectra. The NMR Expert System (N MRES) was written in the C and OPS5 programming languages and implemen ted on a Unix-based (Ultrix) mainframe system with XWindows bit-map gr aphics display. Expert knowledge was acquired from NMR spectroscopists and represented as production rules in the knowledge base. A heuristi c weights method was employed to determine the confidence levels of po tential peaks. Statistical and numerical methods were used to facilita te processing decisions. NMR spectra obtained from studies of ischemic neonatal and immature hearts were used to assess the performance of t he expert system. The expert system performed signal extraction, noise treatment, resonance assignment, intracellular pH determination, and metabolite intensity quantitation in about 10 s per 4 KB (kilobyte) sp ectrum. The peak identification success rate was 98.2%. Peak areas and pH estimated by the expert system compared favorably with those deter mined by human experts. We conclude that the expert system has provide d a framework for reliable and efficient quantification of complex bio logical P-31 NMR spectra.