Use of probabilistic weights to enhance linear regression myoelectric control

Abstract

Objective Clinically available prostheses for transradial amputees do not allow simultaneous myoelectric control of degrees of freedom (DOFs). Linear regression methods can provide simultaneous myoelectric control, but frequently also result in difficulty with isolating individual DOFs when desired. This study evaluated the potential of using probabilistic estimates of categories of gross prosthesis movement, which are commonly used in classification-based myoelectric control, to enhance linear regression myoelectric control. Approach Gaussian models were fit to EMG feature distributions for three movement classes at each DOF (no movement, or movement in either direction) and used to weight the output of linear regression models by the probability that the user intended the movement. Eight able-bodied and two transradial amputee subjects worked in a virtual Fitts’ Law task to evaluate differences in controllability between linear regression and probability-weighted regression for an intramuscular EMG-based three-DOF wrist and hand system. Main Results Real-time and offline analyses in able-bodied subjects demonstrated that probability weighting improved performance during single-DOF tasks (p<0.05) by preventing extraneous movement at additional DOFs. Similar results were seen in experiments with two transradial amputees. Though goodness-of-fit evaluations suggested that the EMG feature distributions showed some deviations from the Gaussian, equal-covariance assumptions used in this experiment, the assumptions were sufficiently met to provide improved performance compared to linear regression control. Significance Use of probability weights can improve the ability to isolate individual during linear regression myoelectric control, while maintaining the ability to simultaneously control multiple DOFs.

Multiday Evaluation of Techniques for EMG-Based Classification of Hand Motions

“…There is an extensive discussion in the literature about 59 performance of classifiers, with each having variable 60 number of amputees (trans-radial [43] or trans-humeral 61 [44], feature selection methods [45,46,47] (Figure 4), previous studies reported low performance up to 30% classification error [55]. In general, the performance of each classifier was similar to previously reported results [53,56]. Combined EMG was significantly better (P<0.05) than the surface and intramuscular EMG as a combined feature set improved the information level from muscles containing both local and global content.…”

“…Intramuscular signals provide independent control sites that can enable simultaneous and proportional control of multiple DOF's [56]. The downside of this simultaneous and proportional control is past pointing, isolating 1 DOF targets and ballistic nature of movements during positioning [56,57].…”

“…Intramuscular signals provide independent control sites that can enable simultaneous and proportional control of multiple DOF's [56]. The downside of this simultaneous and proportional control is past pointing, isolating 1 DOF targets and ballistic nature of movements during positioning [56,57]. Since both acquisition types (surface and intramuscular) and their control schemes (sequential and simultaneous) have limitations, a control scheme based on both surface (isolate single DOF) and intramuscular (provide simultaneous and proportional control of multiple DOF's) recordings could be devised for providing faster, intuitive and natural control.…”

The future of upper extremity rehabilitation robotics: research and practice

“…To match the natural dexterity of the human hand, several groups have proposed using regression‐based algorithms or continuous methods . Unlike pattern recognition, regression algorithms can use the rate of change in neurological activity to continuously and simultaneously estimate multiple control signals.…”

“…Unlike pattern recognition, regression algorithms can use the rate of change in neurological activity to continuously and simultaneously estimate multiple control signals. Both peripheral nerve and brain interfaces have had success in implementing these algorithms for continuous control of multiple DoFs . Although these algorithms show promise in accelerating the progress toward natural arm and hand control, there still remain some limitations.…”