Virtual reality based Machine Learning Arm-Hand Function Evaluation and Support (VAFES)
In this project, a standardized test environment in Virtual Reality (VR) will be developed. Motion trackers and VR gloves cover a broad spectrum of relevant motion parameters. In particular, synchronous electroencephalographic (EEG) recordings supplement the motion data with neuronal signals. This combination makes it possible for the first time to use modern Machine Learning (ML) algorithms, such as deep learning, in the context of diagnosis and therapy of neurological diseases with hand and arm dysfunctions.
The easy-to-use functional test can be used in particular for complex extrapyramidal and/or cerebellar movement disorders to objectively classify the movement deficits and compare them with comparative data. Due to the increased sensitivity of the test and a generative model of arm movements, different stages of the disease can be better distinguished and the course of the disease more clearly documented. The early detection of diseases with a gradual course should also be improved. Therapeutically, the insights gained in tremor treatment – here utilizing a hand exoskeleton to be developed in the project – will be used to establish VR-supported neurofeedback therapy approaches for extrapyramidal and cerebellar movement disorders and fine calibration for deep brain stimulation for Parkinson’s treatment.
Movement and correlated EEG data from both healthy volunteers and patients will be used to build a publicly accessible database. The database will be made available as a reference for the development and validation of models and methods for the scientific community and companies. The developed modular hardware and software components will be used clinically / scientifically (VR test environment) as well as economically (test instrument, decoder, hand exoskeleton).
Smart upper extremity rehabilitation with an intelligent soft exoskeleton (REXO)
The Key component is a bio mechanically designed, an adaptive exoskeleton for the upper extremities, which will be developed in this project and used exploratory on patients. The exoskeleton considers the individual conditions of the disease and compensates as far as necessary for the dysfunction, which prevents the execution of necessary movements or supports the rehabilitation training by antagonistic activation. Due to intelligent sensors and actuators linkage, the system always provides exactly as much support or correction as is necessary for the respective patient situation.
With the exoskeleton, a holistic rehabilitation system is developed. The system includes the design and implementation of motion tasks in virtual reality, a feedback system based on bio signals, and a generic decoder for invasive and non-invasive brain-computer interfaces. In the technical implementation, the soft exoskeleton combines modern, very light, resilient materials with an intelligent, adaptive control system that does not require any adjustments from the wearer.
The result is new perspectives for improved rehabilitation of arm and hand functions. This can significantly improve patient care.
Motor-parietal cortical neuroprosthesis with somatosensory feedback for restoring hand and arm functions in tetraplegic patients
A neuroprosthesis is a system that allows a severely disabled person to control an extracorporeal robotic device with his or her thoughts. Recent developments in neuroprosthetics have great potential to increase the quality of life and autonomy of paralyzed patients. Although a couple of clinical studies for upper limb cortical prosthesis have been started in the USA many aspects of neuroprosthetic systems remain open to research. It is not clear, for example, which areas of the brain are providing the best control signals. Most studies so far focused on the motor cortex and its low-level motor commands. In a recent human, study my colleagues and I at Caltech could demonstrate that high-level cognitive signals which we derived from the posterior parietal cortex (PPC) can be used to drive a neuroprosthesis as well. The proposed project intends to combine the signals from both cortical areas – the motor cortex and PPC – to provide improved performance.
Furthermore, by utilizing intracortical microstimulation in the somatosensory cortex it is intended to elicit tactile sensations which are completely lost to tetraplegics who are paralyzed from the neck down. The implementation of effective somatosensory feedback is also likely to improve performance, especially in fine motor tasks, and has the potential to substantially improve quality of life. Another essential part of the project will be to use immersive virtual reality to study the influence of perspective, visual presentation, and multiple control scenarios before transitioning to a physical robotic limb. Virtual reality provides us with the opportunity to experiment with multiple perspectives, control schemes, and scenarios in a safe and rapid way.
Finally, it is intended to use a hybrid control system in which computer assistance is combined with cortical control signals. This combined system would be more robust against long-term signal degradation and could provide optimal assistance in activities of daily living for tetraplegic patients.