The application of neural networks to myoelectric signal analysis: a preliminary study

• Published in: IEEE transactions on biomedical engineering 1990-03, Vol.37 (3), p.221-230
• Main Authors: Kelly, M.F., Parker, P.A., Scott, R.N.
• Format: Article
• Language: English
• Subjects: Algorithms; Amputation - rehabilitation; Arm - physiology; Arm - surgery; Biological and medical sciences; Computational efficiency; Degradation; Electrodiagnosis. Electric activity recording; Electromyography - methods; Fundamental and applied biological sciences. Psychology; General aspects; Humans; Investigative techniques, diagnostic techniques (general aspects); Laser sintering; Least squares methods; Male; Mathematics in biology. Statistical analysis. Models. Metrology. Data processing in biology (general aspects); Medical sciences; Models, Neurological; Multilayer perceptrons; Muscle Contraction - physiology; Muscles; Nerve Net - physiology; Nervous system; Nervous System Physiological Phenomena; Neural networks; Neural prosthesis; Prosthetics; Signal analysis
• ISSN: 0018-9294; 1558-2531
• DOI: 10.1109/10.52324

Two neural network implementations are applied to myoelectric signal (MES) analysis tasks. The motivation behind this research is to explore more reliable methods of deriving control for multi-degree-of-freedom arm prostheses. A discrete Hopfield network is used to calculate the time series parameter for a moving average MES model. It is demonstrated that the Hopfield network is capable of generating the same time series parameters as those produced by the conventional sequential least-squares algorithm. Furthermore, it can be extended to applications utilizing larger amounts of data, and possibly to higher-order time series models, without significant degradation in computational efficiency. The second neural network implementation involves using a two-layer perceptron for classifying a single-site MES on the basis of two features, the first time series parameter and the signal power.< >