In paced finger-tapping tasks employing different types of period perturbations, a “perturbation context” is evident during resynchronization, biasing the error-correction mechanism. In this work, we show, both experimentally and theoretically, that responses to different perturbation types are dynamically incompatible when they are presented in separate experiments with different contexts. In other words, they cannot be described by the same underlying dynamical system. This finding may explain numerous seemingly contradictory results reported in previous studies, as well as the difficulty of reproducing responses to both perturbation types using a single mathematical model. On the other hand, when both perturbation types are presented randomly within the same experiment, the resulting responses are compatible and can be interpreted as produced by a single underlying nonlinear mechanism. We conclude that a single dynamical system can account for responses to perturbations of all types, signs, and magnitudes, which is nonetheless recalibrated according to the perturbation context.
