1. We evaluated the hypothesis that the neural control of complex motor behaviors is simplified by building movement sequences from a series of simple neural "building blocks." In particular, we compared two reflex behaviors of the frog, flexion withdrawal and the hindlimb-hindlimb wipe reflex, to determine whether a single neural circuit that coordinates flexion withdrawal is incorporated as the first element in a sequence of neural circuits comprising the wipe. The neural organization of these two reflexes was compared using a quantitative analysis of movement kinematics and muscle activity patterns [electromyograms (EMGs)]. 2. The three-dimensional coordinates of the position of the foot over time and the angular excursion of hip, knee, and ankle joints were recorded using a WATSMART infrared emitter-detector system. These data were quantified using principal-components analysis to provide a measure of the shape (eigenvalues) and orientation (eigen-vector coefficients) of the movement trajectories. The latencies and magnitudes of EMGs of seven muscles acting at the hip, knee, and ankle were analyzed over the interval from EMG onset to movement onset, and EMG magnitudes during the initial flexion of the limb. These variables were compared during flexion withdrawal and the initial flexion movement of the limb during the hindlimb-hindlimb wipe reflex (before the onset of the frequently rhythmic portion when the stimulus is removed) when the two reflexes were elicited from comparable stimulus locations. 3. In both the flexion reflex and the initial movement segment of the wipe reflex, the foot moves along a relatively straight line. However, the foot is directed to a more rostral and lateral position during flexion than during wipe. All three joints flex during flexion withdrawal, whereas during the wipe, the knee and ankle joints flex but the angular excursion of the hip joint may vary. The different orientations of the movement trajectories are associated with EMG patterns that differ in both timing and magnitude between the two reflexes. 4. The differences in the kinematics and EMG patterns of the two reflexes during unrestrained movements make it unlikely that the neural circuit that coordinates flexion withdrawal is incorporated as the first element in the sequence of neural circuits underlying the wipe reflex. 5. Unlike the wipe reflex, during flexion withdrawal there is no apparent constraint on the accuracy of placement at the end of the movement, yet the animals nevertheless achieved consistent final positions of both the foot and of each joint. The implications of these findings with respect to the controlled variables are discussed.
Perturbation-specific responses by two neural circuits generating similar activity patterns
