Predicting foodstuff from jaw dynamics during masticatory crushing in man

In previous in vitro experiments using an Instron instrument, each test food was found to have characteristic textural properties. In vivo experiments were now made (1) to determine the degree to which variations in the vertical jaw movements during the crushing phase of mastication can be explained by the inherent properties of the foodstuff being chewed, and (2) to establish the degree to which the foodstuff being chewed can be identified by certain features of the jaw dynamics. Five adults were used for chewing tasks with standardized pieces of beef, carrot or peanut. Each subject made two trials with each foodstuff. The movement of the lower incisal point was monitored; features of movement associated with jaw closing in the first chewing cycle were considered. Five of these features were not suitable to categorize the various test foods. Each of the remaining 4, however, was able to distinguish either one food from the 2 others (2 cases), or one from another (2 cases). Pattern recognition techniques based upon principal component analysis could differentiate jaw closing patterns associated with chewing beef from those involving peanut or carrot. The extent to which peanut could be distinguished from carrot was not as predictable as the categorization of peanut or carrot versus beef. Cross-correlation of in vitro force-time breakage characteristics and the jaw movement data showed that on average 52% of the variation in the vertical jaw movement during crushing of food could be explained by the inherent properties of the food. Because the masticatory apparatus acts as a low-pass filter, the high-frequency components of food breakage characteristics found in vitro were attenuated in vivo. It appears that the inherent filter properties of the masticatory apparatus limit the range of sensory and motor function of the jaw.