|Barton Ii, Franklin|
Submitted to: Eastern Analytical Symposium
Publication Type: Abstract Only
Publication Acceptance Date: July 20, 1997
Publication Date: N/A
Technical Abstract: A near-infrared (NIR) reflectance model was previously developed, using partial least squares regression, to predict total dietary fiber content in a wide variety of cereal and grain products (standard error of cross validation =1.6%, multiple coefficient of determination =0.98). Soluble and insoluble fractions of dietary fiber have different human physiological leffects and their amount in foods is of interest to consumers, the medical community, and the cereal product industry. The current work explores the use of NIR reflectance spectroscopy for the prediction of soluble and insoluble dietary fiber content in cereal products. Insoluble and soluble dietary fiber in milled cereal samples were measured by the AOAC enzymatic gravimetric procedure (AOAC, 991.43) and ranged from 0.46 to 47.98% and 0.03 to 10.33%, respectively. The milled cereal samples (N=88) were scanned using an NIRSystems 6500 monochromator and ISI software. Calibrations were obtained for insoluble and soluble dietary fiber using partial least squares regression. The standard error of cross validation and multiple coefficient of determination were 1.33% and 0.99 and 0.89% and 0.80 for insoluble and soluble dietary fiber, respectively. The results indicate the potential of NIR reflectance spectroscopy for the prediction of dietary fiber fractions. Primary constituents influencing model development were investigated.