SIMULATION OF ELECTROMYOGRAPHIC SIGNALS

A technique for simulating electromyographic (EMG) signals is reported . The four step method is physiologically based and begins with the mo delling of a cross-section of a muscle. Within this cross-section moto r unit territories of various sizes are randomly distributed and withi n a detection area at the centre of the cross-section individual muscl e fibres are modelled and randomly assigned to appropriate motor units . For the motor units with fibres in the detection area, recruitment a nd firing time behaviours, as a function of an assumed level of contra ction, are then simulated. For each active motor unit with fibres in t he detection area, motor unit action potentials (MUAPs) are created us ing a line source volume conductor model. MUAPs can be created for var ious types of detecting electrodes including concentric and monopolar needle electrodes. Finally, the individual motor unit firing time beha viours and MUAPs are combined to create a complete EMG signal. The rou tines are interfaced through a series of user-friendly menus and forms , are implemented in C and can be run on any IBM compatible machine. E xemplary data are presented along with results from attempts to evalua te the model. Suggested uses of the simulation package, especially wit h respect to EMG signal decomposition, are discussed.