Quantification of human chewing-cycle kinematics.

This study introduces new methods of quantifying and evaluating the human chewing cycle. These methods were validated on a sample of 26 young adults (11 women and 15 men) between 20-35 years of age. Movements of the mandibular central incisors were recorded (100 Hz) using an optoelectric computer system while the participants chewed gum. A subsample of 10 cycles was automatically selected, based on multiple objective criteria to ensure representative cycles for each individual. Once representative cycles had been identified, multilevel statistical models were used to evaluate and describe the sample's kinematic patterns. The multilevel procedures allow for missing observations, they do not assume equal intervals, and variation can be partitioned hierarchically. Two-level models showed significantly shorter cycle duration for males (835 msec) than females (973 msec). Inferior-superior (IS) cycle range was 2.6 mm larger and maximum IS velocity was 19.6 mm/sec faster in males than females. There were no significant differences in medial-lateral (ML) and anteroposterior (AP) excursive ranges or velocities. With the exception of cycle duration and ML ranges of motion, random variation was three to five times larger between individuals than between cycles. The three-level models showed that eighth-order polynomials were necessary to describe IS, AP, and ML chewing movements of the entire cycle. The models identified highly significant sex differences in cycle kinematics (excursions, velocities, accelerations, etc.) for each aspect of movement (AP, IS, and ML). It is concluded that this approach provides several important advantages over existing methods, including (a) its objectivity, (b) a more complete description of kinematic patterns, (c) a hierarchical description of variation, and (d) its ability to test hypotheses statistically.