Integration of simulated multipotential signals: the role of integration window width and of the number of spikes.

Electrical signals recorded from nerves/muscles represent the fundamentals for experimental data analysis including an assessment of respiratory motor output. The present work, based on theoretical model, is focused on the linearity and variability of rectified and integrated electroneurogram (ENG)/electromyogram (EMG) signals in relation to the frequency of spike incidence and moving average window width used for processing of signals. Our simulations of multipotential signals (multiunit action potentials) originating from an overlapping of four single units with phase shifts firing at two frequencies demonstrates that (1) integrated ENG/EMG signals are only approximately linearly proportional to the frequency of action potentials in the superposition - multipotential and (2) the width of the moving average window strongly influences the range (dispersion) of integrated values. Better quality of EMG recordings, a higher number of action potentials within the multipotential signals, and a wider width of the moving average window increase the accuracy of integrated ENG/EMG values during processing of motor output signals.