Home Research BCCN I Project Details D1 - Motor learning with multiple internal models
Personal tools

D1 - Motor learning with multiple internal models

Michael Herrmann

Motor control is a behavior that necessarily incorporates constant adaptation in order to cope with changing surrounding. The input-output relations involved in motor control are commonly assumed to be represented in the brain in internal models (e.g. Bushan et al. 1999). From the ability of switching between different motor regimes, the multiplicity of senso-motoric models can be inferred, which act on different levels or under different task conditions. Evidence for, e.g. separate visuomotor and dynamic models comes from interference studies of human motor learning (Krakauer et al. 1999). It is, however, not clear how these multiple models are realized in the brain and how the switching is done between them. Neurophysiological studies have demonstrated that within the motor system, a hierarchical distinction between representations within visual, joint, or muscle coordinates does not seem to exist (Kakei et al. 1999). This may explain why it has been found that the learning of visuomotor and dynamic transformation can interfere with each other (Tong et al. 2002). Internal models have been shown to be necessary in adaptive control (Sontag 2003) and in autonomous robot behavior (Der et al. 2002).

Belongs to Group(s):