Chapter 11 Learning a dynamic limb synergy

Abstract

Dynamic and non-dynamic bimanual synergies were studied during pursuit tracking of a common visual target by both hands. Subjects were required to track a target with two response cursors simultaneously. The target moved in the centre of a computer screen, while the response cursors were located at either side of the target and controlled via a joystick with either hand. The target moved irregularly in amplitude and speed, with a frequency bandwidth up to 2Hz. The tracking system was contrived in two ways so as to produce either dynamic or non-dynamic synergies between the hands. Six subjects practised tracking where the joystick-cursor relation was a straighforward, non-dynamic, scalar relation, identical for the left and right joysticks. This meant that, in order to track the target, the left and right hands simply had to be moved in synergy in an identical fashion, with only one virtual degree of freedom of movement. This was a scalar, non-dynamic synergy. Another six subjects practised a novel task where the right joystick input was first linearly filtered (first-order, low-pass) before driving the response cursor. This mean that, in order to track the target together, the left and right hands had to be moved with different amplitudes and out of phase with each other, but still in synergy. The relative amplitude and timing of the limbs in this synergy were required to vary with the frequency of the movement, so that this was a dynamic synergy. The results showed that the subjects accomplished both non-dynamic and dynamic inter-limb synergies. In the former case, the hands were coupled as expected in a simple, one-to-one relation. In the latter case, the hands were coupled in a dynamic, linear relation, although the specific characteristics of this relation deviated from those required for full compensation for the effect of the filter on the joystick. The degree of inter-limb coupling did not differ between the non-dynamic and dynamic synergies and was high (87–91%) for both, indicating little independent activity between the limbs in either case. Furthermore, the demands of forming the inter-limb synergies, either non-dynamic or dynamic, interfered only minimally with tracking performance, since only minor differences were observed between single-handed and binanual tracking.