Pattern Recognition–based Myoelectric Control of Partial-Hand Prostheses
(Project D2, R4: Manipulation)
Levi Hargrove, PhD
Loss of part of the hand—e.g., the thumb or multiple digits—impairs the ability to make functional hand-grasps; this causes severe disability. Myoelectric prostheses use surface electromyographic (EMG) signals to control joint movements. Small, lightweight myoelectric prostheses such as motorized fingers have recently become available for partial-hand amputees. However, obtaining appropriate EMG control signals remains a challenge.
Controlling partial-hand prostheses using EMG signals from remaining intrinsic hand muscles (i.e., muscles within the hand) or touch-sensitive switches has shown promise. However, in practice, it is very difficult to get good EMG control signals from these muscles. The intrinsic muscles are small and sometimes damaged as a result of the injury or amputation. Also, because the palm of the hand is very mobile it is difficult to keep electrodes in place. It is also challenging to fit individuals with a socket that is comfortable, well suspended, and does not impair wrist motion. Thus few individuals can successfully use a myoelectric partial-hand prosthesis. EMG signals from extrinsic hand muscles that are located within the forearm remain largely intact in partial-hand amputees. As such, these muscles present an alternative option for prosthesis control. However, using conventional control methods is impractical because these muscles are so closely spaced, making it difficult to position electrodes over individual muscles. In addition, extrinsic hand muscles are in close proximity to muscles that control the wrist. This means that the user must keep the wrist still in order to control the hand, so the user is thus unable to preposition the hand, which limits function. Conventional myoelectric control strategies using either intrinsic or extrinsic muscles only allow the user to control one hand grasp. Thus neither approach is adequate to control newer devices that offer multiple functions. Pattern recognition–based myoelectric control methods have been used successfully to provide enhanced, intuitive control of upper limb prostheses. In this approach, movement-specific patterns of EMG signals are deciphered by a computer algorithm and used to determine user intent. The objective of this research project is to develop a pattern recognition–based myoelectric control strategy for partial-hand prostheses. The goal is to allow control of multiple hand grasps while retaining the ability to move the wrist, thus providing significant improvement in prosthesis control and function for partial-hand amputees.
The most frequent cause of partial-hand amputation is trauma, often a result of work- or leisure activity–related accidents or combat-related injury, thus the majority of partial-hand amputations occur in relatively young, active individuals. These individuals need prostheses that can effectively replace lost function, allowing them to maintain productive, independent lifestyles, pursue employment, and engage in active pursuits.