Electromyogram-based neural network control of transhumeral prostheses

Abstract

Upper-limb amputation can cause a great deal of functional impairment for patients, particularly for those with amputation at or above the elbow. Our long-term objective is to improve functional outcomes for patients with amputation by integrating a fully implanted electromyographic (EMG) recording system with a wireless telemetry system that communicates with the patient’s prosthesis. We believe that this should generate a scheme that will allow patients to robustly control multiple degrees of freedom simultaneously. The goal of this study is to evaluate the feasibility of predicting dynamic arm movements (both flexion/extension and pronation/supination) based on EMG signals from a set of muscles that would likely be intact in patients with transhumeral amputation. We recorded movement kinematics and EMG signals from seven muscles during a variety of movements with different complexities. Time-delayed artificial neural networks were then trained offline to predict the measured arm trajectories based on features extracted from the measured EMG signals. We evaluated the relative effectiveness of various muscle subsets. Predicted movement trajectories had average root-mean-square errors of approximately 15.7° and 24.9° and average R2 values of approximately 0.81 and 0.46 for elbow flexion/extension and forearm pronation/supination, respectively.

Within-socket myoelectric prediction of continuous ankle kinematics for control of a powered transtibial prosthesis

“…This emphasis on classification parallels current techniques used in upper limb prosthetic systems to compensate for the uncertainty in mapping a subset of EMG inputs to multiple degrees of freedom and types of movement (Kuiken et al 2005Yatsenko et al 2007, Artemiadis and Kyriakopoulos 2010, Bueno French et al 2011, Pulliam Lambrecht et al 2011, Akhtar Hargrove et al 2012, Hebert and Lewicke 2012, Jiang et al 2012, Jiang et al 2013, Li et al 2013. Multi-layer artificial neural networks and SVMs have been used extensively for this purpose in upper extremity prosthetic systems and have been shown to provide accurate discrimination across classes of limb movement, particularly when used in combination with neurofuzzy systems and auto-regressive models (Englehart and Hudgins 2003, Karlik et al 2003, Liu et al 2007, Au et al 2008.…”

Proceedings of the first workshop on Peripheral Machine Interfaces: going beyond traditional surface electromyography

“…Studies using computational nexting showed the ability to predict and anticipate the future position, motion, sEMG input signals, and contact forces of a myoelectrically controlled robotic limb (Pilarski et al, 2013a), to anticipate the control functions desired by a user (Pilarski et al, 2012), and also to predict the timing of a user’s control behavior (Edwards et al, 2013). This move towards more knowledgeable controllers supports and resonates with non-real-time PMI prediction learning work, e.g., the sEMG-driven predictions of upper-arm joint trajectories demonstrated by Pulliam et al (2011). Creating systems that acquire and maintain predictive temporally extended knowledge regarding human–machine interaction has been shown to be both possible and potentially virtuous.…”

Estimation of distal arm joint angles from EMG and shoulder orientation for transhumeral prostheses

“…After acquisition, the data were filtered with a 5th-order Butterworth high-pass filter with a cutoff frequency of 10 Hz to remove movement artifacts. The EMG data were windowed at 200 ms with an overlap of 75 ms to make an effective timestep of 125 ms. Four time-domain features were extracted from each channel: mean absolute value, waveform length, number of zero crossings, and number of slope sign changes [6, 18]. …”

“…Pulliam et al [6] used EMG recordings from the upper arm and chest to predict the angles of the elbow and forearm simultaneously. Specifically, they implemented a time-delayed adaptive neural network (TDANN) to predict the angles of elbow flexion/extension ( E FE ) and forearm pronation/ supination ( F PS ) [7, 8].…”