Computational Neuroscience
Computational Neuroscience

© Boström

Our Computational Neuroscience Lab conducts model-based research on various aspects of human movement and perception. In particular, we investigate how motor skills are learned and improved, and how actual and phantom pain emerges. First and foremost, we put forward the idea to combine internal forward and inverse models into one unified neural circuit that controls and predicts movement in a parallel fashion. Further current research is done on developing a 3D human model with bones, muscles, reflexes, and motor control units, called Myonardo.
Closely related basic theories comprise the reafference principle, internal models in the brain, and the representation of body segments. Other related areas of interest cover phantom pain, chronic back pain, postural control, whiplash injury, eye-hand coordination, and balancing.
We engage artificial neural networks, such as reservoir computing networks and Kohonen maps. For data evaluation, we use advanced statistical methods, such as general and generalized linear models. Model programming and data analysis is done using MATLAB, Simulink, and R.
We are cooperating with the Friedrich-Schiller-University Jena.
Authors whose research relates to our work are: D. Wolpert, M. Kawato, D. Sussillo, B. Hommel, W. Prinz, and V.S. Ramachandran.