This project will result in the deployment of a fully implanted myoelectric signal (MES) acquisition system in 3 veterans that is used to control advanced dexterous prosthetic hands and wrists for individuals with transradial (below elbow) amputations. A novel MES controller algorithm based on a time-delayed neural network and a probabilistic adaptive filter that selectively suppresses unintended movement commands will allow simultaneous and continuous control of the multiple motions provided by the advanced prosthesis. The implanted MES acquisition system will use a new family of devices that include an implanted battery, implanted amplifiers and analog-to-digital converters, and an implanted telemetry system. This implanted MES acquisition system will provide highly selective and highly repeatable MES recordings from any remaining muscle in the limb, even those located deep within the limb or not located directly underneath the socket. The implanted intramuscular electrodes can be sized for any muscle, will be firmly anchored within the muscles, and do not require any donning by the user. The implanted MES acquisition system will communicate wirelessly with a small external hub that is mounted inconspicuously on the prosthetic limb using the industry standard MedRadio protocol. The external hub receives the MES signals, implements the new MES controller algorithm, and then interfaces with the control bus of the prosthesis to command its various joint motions. The controller algorithm will be initially developed using temporary fine-wire EMG electrode recordings and a virtual prosthesis simulator in which 10 amputee participants will perform simulated functional tasks. The muscle signals used in these experiments will be varied to determine which muscles provide the best MES signals for control, how many muscles will be needed in the final system, and to estimate the expected performance of the final system. In parallel, the external hub will be realized using a small commercial device with built-in MedRadio telemetry and a microcontroller for implementing the advanced controller algorithm, and regulatory approvals (Investigational Device Exemption from the Food and Drug Administration and local Institutional Review Board) will be obtained. The permanent system (implanted MES system, external hub with advanced control algorithm, dexterous commercial upper limb myoelectric prosthesis) will then be integrated and realized in 3 individuals with transradial amputations. Extensive technical assessments will demonstrate the performance of the system components. An extensive battery of clinical assessments will compare the functional performance of the user's standard of care prosthesis control system (typically surface EMG-based recordings with a state controller) to that of the new implanted MES - advanced controller algorithm approach developed under this project. The proposed work is directly aligned with the strategic plan of the VA Rehabilitation Research and Development Service and is an excellent complement to the recent development of sophisticated and dexterous upper limb prostheses by the Defense Advanced Research Projects Agency (DARPA) and by several commercial prosthesis manufacturers. The proposed fully-implanted MES acquisition system and advanced controller algorithm will provide the much richer command information needed to realize the potential of these advanced upper limb prostheses. Although initially targeted to individuals with transradial amputations, this same basic approach (i.e., implanted MES with advanced controller) is broadly applicable to other veteran amputee populations in the future, including transhumeral, partial hand, shoulder disarticulation, and lower extremity amputees. Targeted muscle reinnervation (TMR), another advanced prosthesis control approach, could also be greatly enhanced by the use of the techniques proposed here. The proposed system thus has the potential to truly revolutionize the way that prosthetists and prosthesis users think about myoelectric control!
Reducing the functional consequences of limb loss has long been a major focus of the VA Rehabilitation Research and Development Service. Recently, more than 1000 soldiers have returned from current operations in Iraq and Afghanistan with amputations, and the VA RR&D Service has prioritized research to make prostheses capable of providing function approaching pre-injury levels. Advanced prostheses that nearly match the mechanical performance of natural human arms are beginning to emerge from federal research projects and from commercial manufacturers, but the way that users control these prostheses is still too awkward and unreliable. This project will develop a new approach that will allow amputees to control their prostheses in essentially the same way they would control their natural arm. This new system will be surgically implanted, so nothing other than the prosthesis will have to be put on during use, and the quality of control it provides will always be the same. This system could eventually work with almost any myoelectric or powered prosthesis.
