The configuration and relaxation of motor task sets

Often the performance of a task does not only require the processing of certain stimuli in certain ways, but also certain patterns of interlimb coordination. We studied shifts between different tasks involving different patterns of intermanual coupling by means of the timed-response procedure, which allows to trace state variables related to task sets. The tasks required the production of rapid bimanual reversal movements with symmetric or parallel directions. Symmetric movements are associated with symmetric coupling, as indicated by positive intermanual correlations between the directions of left-hand and right-hand movements, whereas parallel movements are associated with parallel coupling, as indicated by negative intermanual correlations. Task switches were associated with gradual changes of the intermanual correlations, which indicate the state of intermanual coupling as a major ingredient of a task set, in the course of action preparation. At short preparation intervals intermanual correlations were those appropriate for the preceding trial; with increasing preparation time they were replaced by those appropriate for the current trial, but the influence of the preceding trial did not disappear completely. In-between trials, intermanual coupling drifted toward a symmetric coupling, but not to uncoupled limbs. After a change of the task the specification of movement directions was slowed, but its initiation was not delayed. According to these results, task sets relax toward attractors which can be different from the absence of task sets. They are gradually configured during task preparation with a persistent influence of the preceding task, and the specification of response characteristics does not wait until the configuration of the new task set is completed. The research reported in this paper was supported by grant He 1187/14-1 of the Deutsche Forschungsgemeinschaft. We thank Barbara Herbst, Holger Küper, Kevin Schepers, and Elisei Rotaro for their support in running the experiment.