System identification of tendon reflex dynamics

Patellar tendon reflexes were evaluated in 12 healthy adult subjects using several measures of the reflex responses and of the system input-output relationship. A hand-held instrumented hammer was used to tap the patellar tendon and to elicit the reflex response. Tendon reflex dynamics were estimated using the recorded tapping force (as input) and the quadriceps muscle electromyogram and knee joint extension torque signals (as output). A dome-shaped rubber pad was mounted onto the most sensitive spot on the patellar tendon, where it served as a tapping target, and helped to reduce the reflex variability significantly (p < 0.01). The input-output properties of the system relating the reflex torque to the tapping force were characterized using several measures: the tendon reflex gain (Gtr), contraction rate (Rc), and half-relaxation rate (Rhr). Reflex loop delay (t(d)) was estimated using the delay from the onset of tapping force to the onset of reflex torque. We determined that these system parameters provided significantly more repeatable and consistent characterization of tendon reflexes than did reflexive torque or EMG signals alone (p < 0.025). The input-output relationship relating the EMG signals of the stretched muscle to the tapping force was also identified to help characterize neuromuscular dynamics of tendon reflexes. The observed sensitivity and consistency of the reflex system measures suggest that with appropriate simplification of the instrumentation, these methods may prove useful in routine clinical practice, and may allow more precise quantification of the tendon jerk than is currently feasible with standard clinical tests.