Item

Abstract

The so-called previous trial effect has been reported from perception to motor output in human behavior. For example, the speed from previously experienced moving targets biases movements on the current trial in coincident timing tasks. The precise mechanism underlying the previous trial effect is still under debate. In the present study, we tested the hypothesis that previous encounters with moving targets leave a neural trace of the information experienced, i.e. an implicit short-term memory mechanism. This information could be stored in either premotor or visual areas associated with the task, and influence neural processing of the current trial. We assume that injecting noise by means of transcranial magnetic stimulation (TMS) would interfere and degrade such short-term memory and hence reduce or abolish the previous trial effect if TMS is applied over neural sites involved in the mechanism. To test this hypothesis and to identify brain regions involved in mediating the previous trial effect, we asked healthy participants (n = 20) to perform a coincident timing task and applied a burst of TMS pulses (10 Hz) 500 ms into the inter-trial interval and 3500 ms before the next trial to disrupt the activity in right hV5/MT+, in left dorsal premotor cortex, and in the control position Vertex. Trial speed was counterbalanced in a way that allowed every speed to be equally often preceded by all speeds. As expected, participants presented a bias towards the speed of previous trial when intercepting moving targets without receiving TMS pulses. TMS applied over dorsal premotor cortex decreased the previous trial effect in comparison to Vertex stimulation. TMS applied over hV5/MT+ decreased the temporal bias only mildly, reaching significance when compared to performing the task without TMS, but only a trend when compared to Vertex stimulation. These results provide causal evidence that the previous trial effect is mediated to a large extent by inter-trial interval activity in the left dorsal premotor cortex, and to a lesser extent by right hV5/MT+, in a visuomotor integration task with moving objects. Absence of difference in overall timing error between TMS sites indicates that the TMS pulses did not affect processing on the current trial. This suggests that an implicit short-term memory mechanism keeps information from one trial to the next, and that this information — motor or visual — is blended with current trial information so that it biases behavior.