The prosthesis was taught to read thoughts
Researchers from North Carolina State University (North Carolina State University) and the University of North Carolina at Chapel Hill (University of North Carolina) have developed a new technology for decoding neuromuscular brain signals and controlling prosthetic hands.
The work is based on computer models that accurately simulate the behavior of natural structures in the forearm, wrist and hand. This technology can also be used to develop new computer interfaces for games and other applications.
Existing prostheses work on the principle of machine learning. This approach requires the patient to "teach" the device to recognize certain patterns of muscle activity and translate them into commands – for example, to bend or unbend the prosthetic hand. The learning process can be lengthy and time-consuming.
The authors decided to develop a simpler and more understandable, reliable and practical way to control the prosthesis. They developed a model of the musculoskeletal system. To do this, electromyographic sensors were placed on the forearms of six healthy volunteers. The researchers recorded neuromuscular signals that the hand received when performing various actions. These data were used to create a single model that translated the recorded signals into commands that control the prosthesis.
When a person loses an arm, the brain continues to send commands that control it. The new system uses sensors to read these commands and transmits them to a computer in a virtual model of the musculoskeletal system. It converts the received information and sends a control signal to the prosthesis. The prosthesis responds quickly with a coordinated movement close to how a human hand would move.
Lizhi Pan, NC State University
During the preliminary check, both healthy people and people who lost an arm were able to complete all the test tasks using a prosthesis with a new neural interface, even despite the lack of training.
Currently, the group is recruiting volunteers who have undergone transradial (below the elbow) arm amputation for further testing of their development in everyday life. In addition, the researchers intend to combine the musculoskeletal model with traditional machine learning. This will allow the program to study the daily needs and preferences of each person and better adapt to a particular user in the long term.
Article L. Pan et al. Myoelectric Control Based on A Generic Musculoskeletal Model: Towards A Multi-User Neural-Machine Interface is published in the IEEE Transactions on Neural Systems and Rehabilitation Engineering journal.
Aminat Adzhieva, portal "Eternal Youth" http://vechnayamolodost.ru based on NC State News: New Tech May Make Prosthetic Hands Easier for Patients to Use.