Statistical Analysis of Magnetic Resonance Imaging Data in the Normal Brain (Data, Screening, Normality, Discrimination, Variability), & Implications for Expert Statistical Programming for ESS™ (The Expert Statistical System)

SYNOPTIC ABSTRACT

In Magnetic Resonance Imaging, a region of interest is subjected to a strong pulsed magnetic field in a scanner similar to a CAT scanner (which, however, uses X-rays instead of magnetic fields). Under this stimulation, the molecules emit signals which are analyzed into several “parameters”, each with a theoretical justification under some strong chemistry assumptions. Other, possibly more informative, parameters might be able to be constructed. This study concentrates on the traditional parameters T₁, T₂, and ¹Hdensity. T₁ and T₂ are times relating to exponential signal decay (T₁) and maximum signal (T₂), while ¹Hdensity is thought to indicate relative amount of Hydrogen ions present. The data is presented (and screened and reduced), then analyzed to answer questions appropriate to NMR research (including tumor detection, tissue discrimination, and discriminant analysis). Implications for automated expert statistical system analysis are drawn, and the ESS system emphasized.

Key Words and Phrases:

• magnetic resonance imaging
• brain scans
• expert systems
• statistical analysis
• tumor detection
• tissue discrimination
• discriminant analysis
• NMR
• MRI
• normality testing
• data screening
• outliers
• plotting
• contours