How differences in rice texture and oral physiology affect sensory preference and eating behavior of consumers: A generalized additive models (GAMs) approach

Abstract

Both food texture and human oral physiology impact consumer preferences and eating behavior. However, the complex relationships between these factors are difficult to explain using simple mathematical models. For this reason, the potential of a generalized additive model (GAM) to relate food texture to oral physiology during the development of healthier foods was investigated. This semiparametric model combines the advantages of both parametric and nonparametric models. The oral behavior of a modified Japonica rice (JM), a low glycemic index (GI) food rich in resistant starch, was compared to that of regular Japonica rice (JR). Both JM and JR were cooked at three water-to-rice (W/R) ratios (low, recommended, high) to obtain samples with different textural properties. The oral physiological parameters of seventeen participants were measured, including biting force and saliva properties. Furthermore, time-intensity (TI) analysis, video recording and electromyography (EMG) were used to provide multiple dynamic sensory analyses. JM required more chewing cycles and longer chewing durations than JR, due to its firmer textural attributes. Linear GAM analysis demonstrated that high hardness and low stickiness of rice led to more chews, chewing duration, and muscle activities, raising the oral processing load and lowering consumer overall liking scores. Additionally, nonlinear GAM analysis indicated that average saliva flow rate and mucin concentration were negatively correlated with both chewing frequency and duration. A simulated gastrointestinal model showed that JM was digested more slowly than JR. In summary, this study provided valuable insights for optimizing the sensory attributes of high-resistant starch rice to enhance its palatability and consumer acceptance.