Machine Learning in Neural Engineering

Department Empirical Inference
Max Planck Institute for Intelligent Systems

Closed-Loop Robotic Control for Stroke Rehabilitation

Current rehabilitation methods for stroke patients are limited in providing significant long-term functional recovery. In this project, we develop an integrated rehabilitation strategy that combines robot-assisted physical therapy with closed-loop neural feedback by means of a brain-computer interface. This enables us to guide the patient in establishing cortical activation patterns that support functional recovery.

A crucial aspect of this project is the investigation of the neural basis of visuomotor learning and its relevance for post-stroke recovery. Obtained insights are then translated into clinical rehabilitation protocols in collaboration with the University Hospital Tübingen and the Deparment for Intelligent Autonomous Systems at Technische Universität Darmstadt.

More details on this project can be found in the following selected publications:

Predicting Motor Learning Performance from Electroencephalographic Data (Journal of NeuroEngineering and Rehabilitation 2013) Visuomotor learning performance can be predicted from pre-trial EEG data, providing novel insights into the neural basis of motor-learning.

Closing the sensorimotor loop: haptic feedback helps decoding of motor imagery (Journal of Neural Engineering, 2011) Closed-loop control of a robot arm by means of a BCI enhances decoding accuracy relative to no haptic feedback.

Using Brain-Computer Interfaces to Induce Neural Plasticity and Restore Function (Journal of Neural Engineering, 2011) A review article that discusses the current state-of-the-art and open problems in using BCIs for inducing cortical reorganization.

Epidural ECoG Online Decoding of Arm Movement Intention in Hemiparesis (First ICPR Workshop on Brain Decoding, 2010) Post-stroke movement intent can be decoded with high-accuracy and short feedback delay from epidural ECoG recordings.