Toward improved sensorimotor integration and learning using upper-limb prosthetic devices

Abstract

To harness the increased dexterity and sensing capabilities in advanced prosthetic device designs, amputees will require interfaces supported by novel forms of sensory feedback and novel control paradigms. We are using a motorized elbow brace to feed back grasp forces to the user in the form of extension torques about the elbow. This force display complements myoelectric control of grip closure in which EMG signals are drawn from the biceps muscle. We expect that the action/reaction coupling experienced by the biceps muscle will produce an intuitive paradigm for object manipulation, and we hope to uncover neural correlates to support this hypothesis. In this paper we present results from an experiment in which 7 able-bodied persons attempted to distinguish three objects by stiffness while grasping them under myoelectric control and feeling reaction forces displayed to their elbow. In four conditions (with and without force display, and using biceps myoelectric signals ipsilateral and contralateral to the force display,) ability to correctly identify objects was significantly increased with sensory feedback.

Design and Integration of an Inexpensive Wearable Mechanotactile Feedback System for Myoelectric Prostheses

“…Future studies should investigate the use of a SingleTact sensor with a larger range, as some prosthetic applications may require measurements above the 10 N sensor evaluated in this study. It may be beneficial to compare the performance of this sensor to others such as strain gauges [9] , [16] , [33] – [37] or sensors specifically designed for prosthetic applications [38] , [54] , [55] . Long-term performance of these sensors should be quantified prior to any take-home studies, including investigation into coatings or covers, such as those employed by Rosenbaum-Chou et al [23] and Clemente et al .…”

“…Many studies have utilized force sensitive resistors (FSR) to measure applied force [22] – [25] , [29] – [32] , due to their low cost and thin profile. Other sensors include subminiature load cells [18] , strain gauges [9] , [16] , [33] – [37] , and impedance-based sensors [38] . In a 2011 review paper, Chappell concluded that the FSR was the optimal sensor for use in prosthetic applications, given the existing state of sensor technology.…”

Neural activation differences in amputees during imitation of intact versus amputee movements

“…Amputees cite lack of clear understanding of how to use their devices as well as dissatisfaction with the challenge of performing tasks of daily living (Kejlaa, 1993; Dudkiewicz et al, 2004). Amputees also encounter unique challenges in incorporating their prosthesis into tool use tasks due to altered sensory feedback (Ridding and Rothwell, 1995, 1997; Irlbacher et al, 2002; Reilly et al, 2008; Rosen et al, 2009; Gillespie et al, 2010; Stepp and Matsuoka, 2010; Rossini et al, 2011) and sense of agency (Ehrsson et al, 2008; Cipriani et al, 2009; Rosen et al, 2009). In addition to these difficulties, amputees report an uncomfortable foreignness when operating their prostheses (Smurr et al, 2008).…”

EMG-Controlled Non-Anthropomorphic Hand Teleoperation Using a Continuous Teleoperation Subspace

“…EMG teleoperation of a one DOF gripper usually uses proportional control to determine the gripper aperture [6] [7]. Using EMG, it is possible to estimate the aperture of a gripper in conjunction with the position of the elbow and the wrist [8].…”