Correlation between instrumental and sensory ratings by evaluation of some texture reference scales

Abstract

The objective of this study was to investigate instrumental-sensory relationships of some texture scales using argentine foods as references. Textural characteristics of these foods were instrumentally investigated by the texture profile analysis technique. Principal component analysis (PCA) was used to describe the main attributes of the food samples. High Pearson's correlation coefficients were found between hardness and fracturability (r=0.94; P<0.0001), hardness and gumminess (r=0.71; P<0.0001) and springiness and cohesiveness (r=0.85; P<0.0001). PCA identified two significant principal components, which accounted for 81.2% of the variance in the instrumental data. Additionally, a trained panel described the texture characteristics of the food samples according to the standard reference scales. The correlation curves showed nonlinear relationships (R2 between 85.6% and 99.9%) which were used to predict sensory attributes of other food samples. Some texture attributes like hardness and fracturability were accurately predicted by mechanical properties, while others like cohesiveness and adhesiveness were less representative.