Course content / topics
In our daily live, we are able to perform complex movements and to adapt our movement behavior continuously to changing environments.
For doing so, our motor control system performs continuously various control and adaptation processes on different controls levels. In order to be able to understand and to model these control and adaptation processes, basic knowledge on the interaction between neural control and biomechanics is necessary.
Goals of this lecture are on the one hand to get an understanding of motor behavior as an interaction of neural control and biomechanics and (changing) environments and on the other hand to learn methods for the development of models of motor control and motor learning processes. An important part of the lecture will cover the description and modeling of motor adaptation. Human adaptive motor behavior will be presented in various experimental setups. The effects of different learning strategies, impairments in motor adaptation in neurological patients and the use of motor learning principles in neurorehabilitation will be presented.
In the first lectures we will introduce basic principles and methods on motor control and will review topics on biomechanics and control theory in order to be able to understand models of motor control and motor adaptation.
In the following, the concept of internal models will be introduced in the respect of predictive control and motor adaptation. Within this topic, students will be familiarized with concepts like the efference copy and experimental setups like visuomotor adaptation and force field adaptation. Students will learn how to model these adaptation processes. Furthermore effects of impaired internal models on motor learning as well as the use of motor adaptation paradigms in neurorehabilitation will be introduced and discussed.
In the lecture “Reinforcement Learning”, students should learn the basic principles of this leaning method including: fitness function, value function and temporal difference learning as well as the application of this concepts on motor learning.
In a following lecture, the interaction of different learning strategies and the dependence of learning strategies on the type of feedback information are discussed.
In addition, students will be familiarised with Bayesian statistical methods and how to apply them to sensorimotor problems. In particular, they will learn about Bayesian models of motor learning and adaptation. One session will be dedicated to understand generalisation of learned motor behaviours to novel situations. In another session, ideal actor models will be introduced as a normative class of models to understand sensorimotor behavior and their relation to economic models will be explained.
In particular, optimal control methods will be discussed as a class of models that allows understanding dynamic behavior.
Students will learn how optimal control models and Bayesian models are connected in motor tasks with uncertainty.
Basic knowledge in algorithms and machine learning
Shadmehr, R., M. A. Smith, et al. (2010). "Error Correction, Sensory Prediction, and Adaptation in Motor Control." Annu Rev Neurosci.
Franklin DW, Wolpert DM. Computational mechanisms of sensorimotor control. Neuron 2011, 3;72(3):425-42.
Sutton, R. and Barto, A. (1998). Reinforcement Learning – an introduction. MIT Press