A Comparison of Surface and Intramuscular Myoelectric Signal Classification

Abstract

The surface myoelectric signal (MES) has been used as an input to controllers for powered prostheses for many years. As a result of recent technological advances it is reasonable to assume that there will soon be implantable myoelectric sensors which will enable the internal MES to be used as input to these controllers. An internal MES measurement should have less muscular crosstalk allowing for more independent control sites. However, it remains unclear if this benefit outweighs the loss of the more global information contained in the surface MES. This paper compares the classification accuracy of six pattern recognition-based myoelectric controllers which use multi-channel surface MES as inputs to the same controllers which use multi-channel intramuscular MES as inputs. An experiment was designed during which surface and intramuscular MES were collected simultaneously for 10 different classes of isometric contraction. There was no significant difference in classification accuracy as a result of using the intramuscular MES measurement technique when compared to the surface MES measurement technique. Impressive classification accuracy (97%) could be achieved by optimally selecting only three channels of surface MES.

Electromyogram pattern recognition for control of powered upper-limb prostheses: State of the art and challenges for clinical use

“…For transradial amputees, with electrodes placed above the remnant muscles of the forearm, dramatic improvement in accuracy results when the number of channels is increased up to four or five; beyond this, there are diminishing benefits. Th is effect has been shown in some detail by Hargrove et al [16]. Six individuals with intact upper limbs performed 10 types of motion (forearm pronation, forearm supination, wrist flexion, wrist extension, radial deviation, ulnar deviation, key grip, chuck grip, hand open, and a rest state).…”

“…Joint time-frequency methods have been shown to effectively represent transient EMG patterns resulting from dynamic contractions [25][26]. A comparison of feature sets has shown that for slowly varying EMG patterns, a concatenated TD/AR (TDAR) feature set outperform s all others [16,24] but the slight improvement in performance over simple TD fe atures incurs considerable processing overhead.…”

Myoelectric forearm prostheses: State of the art from a user-centered perspective

“…Studies by Tenore et al [35] and Sebelius et al [21,29] incorporated flexion and extension of the separate fingers and thumb in classification. In recent studies, more focus is found on functional grasps, such as the cylindrical, tripod, and lateral grasps [28,[36][37][38].…”

Classification of Human Hand Movements Using Surface EMG for Myoelectric Control

“…Fig. 1: Hand movements of Exercise B [8] 3 Data processing and feature extraction sEMG signals are non-stationary, having complicated time, frequency, and timefrequency features; therefore, the raw data cannot be used as the input for classifier training; an appropriate data pre-processing is often more influential than the choice of classifiers [2,3].…”