Evidence of deterministic chaos in the myoelectric signal

Our aim was to study whether the myoelectric signals can be better modelled as outputs of a nonlinear dynamic system rather than as random stochastic signals. Both the nonlinear predictability and the dimensionality of the signals were studied using methods of nonlinear dynamics. The signals were measured from the biceps brachii muscle during both fatiguing and non-fatiguing isometric contractions at low load levels. The myoelectric signals were found to be nonlinear and to have a structure statistically distinguishable from random noise. The correlation dimension describing the dimensionality of the myoelectric signal decreased during local muscular fatigue. The results support the use of the theory of nonlinear dynamics for the modelling of the myoelectric signals.