Topology of learning and correction in dynamic balance

Abstract

"The discrepancy between [our personal experiences] and our idealizations of knowledge leads us into counterproductive strategies for learning and thinking" (Papert 1993). This discrepancy has driven most research towards understanding how people fail to accomplish a task, while few researchers have focused on looking at those instances where people have developed strategies for correcting their failure. This work has as its core belief that by identifying the strategies of correction we will develop better ways to assist in learning. However, to successfully assist in learning, we need to take into account both general rules of behavior and individual solutions. To explore this, we look at a concrete example from posture control, balance, to understand how learning modifies it. This research focuses on identifying an algorithm looking at the process of correction during dynamic balance. We outline an experiment whereby healthy subjects attempt to learn to balance on a two-degrees of freedom platform through external-focus feedback. The intent is to capture and analyze how the structure of old and new-learned body synergies for dynamic balance changes over time. The analysis we present offers a perspective of how subjects achieve kinetic coherence by building strategy maps.