Wipe and flexion reflexes of the frog. II. Response to perturbations

1. To evaluate the hypothesis that the neural control of sensorimotor transformations may be simplified by using a single control variable, we compared the movement kinematics and muscle activity patterns [electromyograms (EMGs)] of the frog during flexion withdrawal and the hind limb-hind limb wipe reflex before and after adding an external load. In addition, the flexibility of spinal cord circuitry underlying the hind limb- hind limb wipe reflex was evaluated by comparing wipes before and after removal of one of the contributing muscles by cutting a muscle nerve. 2. The kinematics of the movements were recorded using a WATSMART infrared emitter- detector system and quantified using principal-components analysis to provide a measure of the shape (eigenvalues) and orientation (eigenvector coefficients) of the movement trajectories. The neural pattern coordinating the movements was characterized by the latencies and magnitudes of EMGs of seven muscles acting at the hip, knee, and ankle. These variables were compared 1) during flexion withdrawal and the initial movement segment of the limb during the hind limb-hind limb wipe reflex in both unrestrained movements and in movements executed when a load equal to ~10% of the animal's body weight was attached to a distal limb segment and 2) during the initial movement segment of the wipe reflex before and after cutting the nerve to the knee flexor-hip extensor, iliofibularis. 3. Addition of the load had no discernible effect on the endpoint position of the foot during either reflex. However, during the loaded flexion reflex, the ankle joint did not move until after the hip and knee joints had moved to their normal positions. This delayed flexion of the ankle was accompanied by large increases in the magnitude of EMG activity in two ankle muscles that exceeded the levels found during unrestrained movements. Significant changes in the temporal organization of the EMG pattern accompanied the change in joint angle relations during flexion withdrawal. 4. Despite the addition of an external load, all animals successfully and reliably removed the stimulus during the wipe reflex, and the relative timing of both the EMG pattern and joint angle motion was preserved. 5. Immediately after section of the nerve to a single muscle (iliofibularis), all animals successfully and reliably removed the stimulus during the wipe reflex. The relative timing of muscle activation was preserved, accompanied by a reduction in the activity level of gluteus magnus, a muscle with action reciprocal to iliofibularis. 6. Although there is no apparent constraint on the accuracy of foot placement at the end of flexion withdrawal, animals nevertheless achieved consistent final positions of both the foot and of each joint even after addition of the load. This suggests that position may be controlled in this reflex, with feedback based on some parameter of an intrinsic coordinate system specifying final position. The timing of motion at different joints was disrupted by the load, however, suggesting that in the flexion reflex there is no circuitry controlling the timing of relative joint motion. 7. The wipe reflex is a rhythmic, target-directed movement, traditionally described as the output of a central pattern generator. The accuracy, temporal coherence of the neural pattern, and kinematics in the initial movement segment of the hind limb- hind limb wipe reflex were preserved in the face of both an external load and removal of one of the muscles. Although the initial movement segment is not rhythmic, it is the first part of a rhythmically organized movement. It may be this rhythmic organization, rather than the fact that the wipe reflex is target directed, that distinguishes it from a temporally discrete movement such as flexion withdrawal. Possible neural circuitry underlying the two reflexes is discussed in this context.