A psychophysiological analysis of response-channel activation and outcome states in Eriksen's noise-compatibility task.

Eleven subjects were tested with a modified version of Eriksen's noise-compatibility task. Four noise conditions were realized--neutral, compatible, fully incompatible, and partially incompatible context. Target and flanking stimuli had an intrinsic association to either the left- or the right-hand side. As a response subjects pulled, with either their left or their right hand, an ordinary bicycle handbrake from a preload position towards a stop. The EMG from the forearm muscles and the movement trajectory were recorded continuously. Trials were categorized according to the outcome states, i.e., whether a full movement or only a rudimentary EMG activation had occurred in the correct and in the incorrect response channels, respectively, and whether the correct-channel activation was leading or lagging. Incompatible noise delayed information transfer and increased the likelihood of errors, while compatible noise had an opposite effect, i.e., it facilitated transfer and reduced the likelihood of errors. The effect of noise on transfer times was the same for all outcome states. Moreover, in all cases, noise had an effect on transfer times only up to the point of EMG onset, while it left movement-execution times unaffected. These findings are seen as contradictory to a strong continuous flow conception according to which any response-related change in the input channel is continuously and immediately transferred to the response-execution device. At least the final stage of movement control seems to be autonomous and not subject to input control, except for the case that a movement already initiated has to be aborted completely.