Getting a Grip on the Impact of Incidental Feedback From Body-Powered and Myoelectric Prostheses.

Michael A Gonzalez,Jiyeon Kang,Deanna H Gates,R Brent Gillespie,Christina Lee

doi:10.1109/tnsre.2021.3111741

Abstract

Sensory feedback from body-powered and myoelectric prostheses are limited, but in different ways. Currently, there are no empirical studies on how incidental feedback differs between body-powered and myoelectric prostheses, or how these differences impact grasping. Thus, the purpose of this study was to quantify differences in grasping performance between body-powered and myoelectric prosthesis users when presented with different forms of feedback. Nine adults with upper limb loss and nine without (acting as controls) completed two tasks in a virtual environment. In the first task, participants used visual, vibration, or force feedback to assist in matching target grasp apertures. In the second task, participants used either visual or force feedback to identify the stiffness of a virtual object. Participants using either prosthesis type improved their accuracy and reduced their variability compared to the no feedback condition when provided with any form of feedback (p < 0.001). However, participants using body-powered prostheses were significantly more accurate and less variable at matching grasp apertures than those using myoelectric prostheses across all feedback conditions. When identifying stiffness, body-powered prosthesis users were more accurate using force feedback (64% compared to myoelectric users' 39%) while myoelectric users were more accurate using visual feedback (65% compared to body-powered users' 53%). This study supports previous findings that body-powered prosthesis users receive limited force and proprioceptive feedback, while myoelectric prosthesis users receive almost no force or proprioceptive feedback from their device. This work can inform future supplemental feedback that enhances rather than reproduces existing incidental feedback.

Highlights

A NALOGUES to the mechanoceptors and proprioceptors of the natural hand do not exist in conventional bodypowered or myoelectric prostheses
The study consisted of a grasp aperture matching task [8] in which we rendered force, vibration, or visual feedback at two levels to represent discrete contact events at the boundaries of a virtual object, and a stiffness identification task in which we provided graded force or visual feedback to simulate the properties of a physical object
Grasp aperture error was not affected by limb (p = 0.607) or feedback type (p = 0.067) (Fig. 2A), nor were there any significant interactions (p = 0.137)

Summary

Introduction

A NALOGUES to the mechanoceptors and proprioceptors of the natural hand do not exist in conventional bodypowered or myoelectric prostheses. This denies prosthesis users access to the reliable sensory feedback and subconscious feedback processing available to an anatomical limb when they perform tasks with their device. While native sensory feedback from the anatomical limb is absent, prosthesis users still have access to natural cues from their prosthesis. This incidental feedback includes visual, auditory, and socket-transmitted loads or vibrations [3]–[5]. The relative availability and utility of incidental feedback between body-powered and myoelectric prosthesis users has only been characterized through anecdotal evidence [6]

Objectives

Methods

Results

Discussion

Conclusion