Cross-correlation analysis of multi-unit EMG recordings in man

Abstract

This paper discusses the measurement of synchrony between 2 multi-unit EMG recordings. The suitability of a number of previously described indices of synchrony is reviewed. A new index of synchronisation is proposed. This index is independent of the number of units contributing to the correlogram but is dependent upon their firing frequency and upon the bin width of the correlogram. The autospectral density function calculated from each multi-unit record is used to estimate motor unit firing rates. The relationship between the frequency of firing and the chosen index of synchronisation is reciprocal such that if the firing rate is doubled, the index is halved. This may be explained if the proportion of events that is synchronized remains constant and is independent of the firing rate. It is argued that this in turn indicates that the proportion of common and non-common sources driving the neurons also remains unchanged in these experiments. In these circumstances, and to take into account changes in unit firing that may be present in different recordings, it is convenient to normalize an experimentally determined index E/M at a frequency f to some standard frequency fs. The problem of superimposition of spikes that occurs as the contraction strength is increased is discussed.