Prediction of Total Dietary Fiber by Near-Infrared Reflectance Spectroscopy in High-Fat- and High-Sugar-Containing Cereal Products

Abstract

The near-infrared (NIR) spectral properties of cereal products containing high fat or high sugar can differ substantially from the spectral properties of other cereal products. An existing NIR model, using preprocessed reflectance spectra and partial least-squares analysis, for the prediction of total dietary fiber in cereal products was expanded to two new models called (1) the “fat-expanded” model and (2) the “fat- and sugar-expanded” model. The fat-expanded model enlarges the existing model with high-fat-content products as calibration samples, and the “fat- and sugar- expanded” model also includes products with high sugar and high crystalline sugar content. The dry milled cereal and grain products were analyzed in the laboratory according to AOAC method 991.43 for the determination of total dietary fiber, and NIR reflectance spectra were collected with a scanning monochromator. Data analysis and selection of representative high-fat and high-sugar samples were performed with a commercial analysis program. The two expanded models had standard errors of cross-validation and R2 similar to those of the existing model, with acceptable standard error of performance and r2 when tested with independent validation samples. The existing model was, thus, expanded to include high-fat, high-sugar, and high crystalline sugar cereal products while maintaining prediction accuracy. Keywords: Dietary fiber; near-infrared; chemometrics; cereal