Utilizing Redundancy in Motor Learning

Open Access
Ranganathan, Rajiv
Graduate Program:
Doctor of Philosophy
Document Type:
Date of Defense:
August 04, 1982
Committee Members:
  • Karl Maxim Newell, Dissertation Advisor
  • Karl Maxim Newell, Committee Chair
  • Mark Latash, Committee Member
  • Dagmar Sternad, Committee Member
  • Joseph Paul Cusumano, Committee Member
  • redundancy
  • variability
  • motor learning
  • movement path
  • interception
The issue of redundancy (Bernstein, 1967) is one of the central issues in motor control. In almost every motor task, there are infinite solutions to accomplish the task goal. We addressed the questions of how learning influences the utilization of redundancy and whether practice schedules that facilitate the use of redundant solutions enhance learning. We used a virtual interception task and examined the redundancy at the level of the movement path of the end effector. In Experiment 1, we found that there was evidence for utilizing path redundancy even after extensive learning. Spatial variability in the middle of the path was greater than the variability near the target. With learning, there was a decrease in path variability throughout the movement path, but with extended learning there was a selective decrease in path variability only near the target. The structure of the movement variability also indicated that different paths were being used to hit the target. In Experiment 2, we addressed the question of whether the ability to flexibly use redundant solutions was influenced by the degree to which redundancy was utilized during practice. Contrary to expectations, we found that the group with lowest utilization of redundancy during practice was better able to utilize multiple paths to hit the target, indicating that flexibility may be emergent from learning the target location. Additionally, these results also highlight the importance of the task context in drawing inferences based on analysis of movement variability. In Experiment 3, we examined the influence of introducing variability at the task goal level and the execution redundancy level on retention and generalization. Variability at the task goal level was induced by changing the target location whereas variability at the execution redundancy level was induced by requiring participants to use different movement paths to achieve the same target location. The results showed that introducing variability at the task goal level led to better performance when the target was in different locations. However, the manipulation of movement variability at the execution redundancy level did not show any positive effects on retention or generalization. Overall, the results support the view that skilled performance does not involve consistent repetition of a single movement pattern; instead, multiple solutions are always used to achieve the task goal. However, an important caveat is that the ability to utilize redundant solutions may be an emergent feature of learning a particular task-relevant parameter. As a result, practice schedules that explicitly constrain the participant to use different ways to achieve the task goal may not necessarily facilitate learning.