|Ph.D Student||Yosef Victor|
|Subject||Control of Bimanual Periodic Movements: Adaptation for|
|Department||Department of Electrical and Computer Engineering||Supervisors||PROFESSOR EMERITUS Gideon Inbar (Deceased)|
|DR. Karniel Amir|
|Full Thesis text|
This study presents evidence supporting the existence of an internal model in bimanual coordinated movements, exploring, in particular, the internal model in bimanual coordinated movements performed when altered visual-proprioceptive feedback is present.
Experiments were carried out with five subjects who performed periodic bimanual coordinated movements by rotating the handles asymmetrically with time-varying visual feedback different for each hand. The subjects were instructed to rotate the handles at maximum angular velocity, while synchronizing the positions of two red LED circles.
Each experimental session included three phases: Familiarization, Learning and Testing. During the Familiarization phase, the subjects became familiar with the experimental setup by rotating the handles; during this phase, the visual feedback ratio was 1:1. During the Learning Phase, the subjects were trained to maintain a steady angular velocity ratio of 3:2 between the dominant hand and the non-dominant hand by means of a 2:3 visual feedback for the dominant hand. The Testing Phase included a change in the dynamics of the dominant hand’s visual-feedback - dynamics of the plant - from a 2:3 ratio to a 1:1 ratio. The reaction of the subject to this change was recorded.
The experimental results indicated three performance types, which the third type consisting of one subject, exhibited both learning and internal model evolution processes as well as evident aftereffect phenomena. The main phenomena exhibited by subject of the third type were related to the rotation velocity. The rotation velocity increased during the Learning Phase up to four times faster than the initial value. This increased rotation velocity was maintained during the Testing Phase, indicating the evolution of an internal model during the Learning Phase and its utilization during the Testing Phase. The internal model developed caused a structure change from a closed loop structure to a feed forward structure, resulting in distinct performance benefits such as increased rotation velocities and shorter propagation delays. In addition, the model emphasizes the “penalties” involved in such an internal model if a change in dynamics occurs so that the plant dynamics no longer match the evolved internal model. In this case, aftereffect phenomena are exhibited, resulting in a prolonged settling time.
The second part of the study developed simulation models for the Testing Phase performance. The motivation for the development of the simulation models was to determine a minimal configuration, capable of mimicking the phenomena exhibited in the experimental data.