Submitted to: Journal of Texture Studies
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/26/1996
Publication Date: N/A
Citation: N/A Interpretive Summary: Texture of a food product relates to how the food behaves when it breaks down. Instrumental measurements of how food breaks down when forces are applied are sought in order to find those measurements that relate to the same way that the human perceives texture when food is chewed. A new apparatus (B.I.T.E, bi-cyclical instrument for texture evaluation), that uses dentures and a three directional movement to simulate chewing, was evaluated against a standard compression device to perform the Texture Profile Analysis (TPA) test. Mathematical models were sought to explain and predict single sensory parameters from instrumental measurement(s). Hardness and springiness could be predicted reasonably well using non- linear models that first required some transformations of the data. However, several measurements were needed in more complicated non-linear models to predict cohesiveness and chewiness. This study aids food scientists and engineers who develop or test food products or texture instruments.
Technical Abstract: Generally, the analysis of a product's texture involves sensory evaluation of a set of attributes by a trained panel along with an instrumental evaluation of similar attributes to generate a correlation value. The most popular instrumental imitative test is the Texture Profile Analysis (T.P.A.), a double compression test performed with flat plates attached to an Instron Universal Testing Machine (I.U.T.M.). From the force/ deformation curve generated by the test, the parameters of hardness, cohesiveness, springiness, and chewiness are extracted. A new bi-cyclical instrument for texture evaluation (B.I.T.E) uses the I.U.T.M. to generate its motion, but substitutes a set of artificial dentures for the flat plates and replaces the unidimensional motion of the upper plate with a tri-dimensional movement simulating a chewing motion. Five parameters were extracted from the force/deformation curve of the B.I.T.E. to evaluate their potential use for the prediction of four sensory attributes (i.e., hardness, cohesiveness, springiness and chewiness). Satisfactory univariate non-linear models (i.e. Fechner and Stevens models) were found for both hardness and springiness, but not for cohesiveness and chewiness. The use of several parameters in non-linear models showed potential use as predictors of sensory cohesiveness and chewiness.