DisplayTitle

Prediction of Cooked Rice Texture Quality Using Near-Infrared Reflectance Analysis of Whole-Grain Milled Samples

¹USDA/ARS, Richard B. Russell Research Center, Athens, GA. The mention of firm names or trade products does not imply that they are endorsed or recommended by the U.S. Department of Agriculture over other firms or similar products not mentioned. ²Corresponding author. E-mail: bobw@Athens.net ³USDA/ARS, Southern Regional Research Center, 1100 Robert E. Lee Blvd., New Orleans, LA 70124. ⁴USDA/ARS, Rice Quality Laboratory, Beaumont, TX. ⁵University of Arkansas Rice Research and Extension Center, Stuttgart, AR. ⁶Louisiana State University, Rice Research Station, Crowley, LA. ⁷California Cooperative Rice Research Foundation, Biggs, CA.

Rice quality is based on chemical and physical properties affecting its appearance, flavor, and texture characteristics. Sensory quality can be assessed by a combination of descriptive sensory and physicochemical property evaluations. The purpose of the present study was to assess the potential of near-infrared reflectance spectroscopy (NIRS) and NIRS in combination with other physicochemical measurements for the determination of sensory texture attributes in whole-grain milled rice samples. Six rice samples representing combinations of variety and growing locations received treatments of two degrees of milling and five drying conditions to achieve final moisture levels of 12 or 15% (n = 120). Quality measurements of the cooked rice included sensory and instrumental texture analyses. Quality measurements of the uncooked rice included amylose and protein (chemical reference), whiteness, transparency, and degree of milling (appearance units of milled rice), and NIRS analyses. Partial least squares (PLS) regression was used to reveal the relationships between the different types of measurements. The sensory texture attributes: manual adhesiveness (MADHES), visual adhesiveness (VADHES), and stickiness to lips (STICKI) were related to deep-milled samples and positively correlated to amylose, whiteness, and milling degree. The attribute roughness (ROUGH) was related to light-milled samples and positively correlated to protein and negatively correlated to amylose. The main variation in sensory attributes was a result of amylose and protein contents of the rices. A noise-compensation value, relative ability of prediction (RAP), was used to express the degree of prediction (1.0 = best possible prediction). NIRS gave the best prediction results for the texture attributes: MADHES, VADHES, and STICKI with an RAP of 0.57, 0.54, and 0.56, respectively. NIRS is best at predicting texture characteristics of cooked rice perceived in the visual, tactile, and initial oral phases of sensory evaluation. The calibration of NIRS plus physicochemical variables did not improve the predictability of sensory texture over NIRS alone. The prediction of sensory texture in rice by NIR needs to be further investigated on a large number of samples with different varieties, growing locations, cultivation methods, harvesting methods, and processing after harvesting.