Prediction of Dietary Fiber in Corn-Based Raw Ingredients and Finished Products Using Visible-Near-Infrared Spectroscopy (Vis-NIRS)

Dietary fiber is an essential nutrient with well-established health benefits, making it necessary to include it in nutritional labels. There is a growing demand in the food industry for rapid, accurate methods for measuring nutrients. Fiber content was rapidly estimated using visible-near-infrared spectroscopy (Vis-NIRS) in reflectance mode, a method that requires minimal sample preparation. Super partial least-squares regression (sPLSR) was used to predict total dietary fiber (TDF) and insoluble dietary fiber (IDF) directly from spectra in corn-based raw ingredients and finished products. The Association of Official Analytical Chemists (AOAC) methods 991.43 and 985.29 were used as reference methods to calibrate the regression models. Three calibration models were developed. A specialized whole-grain total dietary fiber (WG TDF) model estimated TDF in raw whole-grain corn ingredients. A global dietary fiber (GDF) model similarly determined TDF of both raw ingredients and finished products, including corn chips. This model serves as a comprehensive screening tool and provides robust predictions across diverse processed matrices. An insoluble dietary fiber (IDF) model was established for raw corn ingredients to support the inclusion of IDF content in nutritional labeling. The models had average root-mean-square errors of validation (RMSEV) of 0.84 ± 0.01% for WG TDF, 1.116 ± 0.008% for GDF, and 1.05 ± 0.01% for IDF. This study demonstrates a chemometric approach using Vis-NIRS for the rapid estimation of dietary fiber in corn-based products, offering an alternative to time-consuming and expensive wet-chemical methods.