Article révisé par les pairs
Résumé : The aim of the present study was to determine whether the negative relationship usually found between reaction time and foreperiod duration in the variable foreperiod paradigm is entirely due to sequential foreperiod effects. It has been shown that when a particular foreperiod has been preceded by a longer one on the previous trial, reaction time is longer than when the preceding foreperiod was equal or shorter. This may be sufficient to explain the increase in reaction time observed for short foreperiods in variable foreperiod conditions. The present results show that, when sequential effects were controlled by the elimination of all trials where the foreperiod was shorter than the preceding one, the negative slope of the reaction time-foreperiod function diminished but did not disappear. The results suggest an interpretation of the role of conditional probability of stimulus arrival in terms of variation in the tendency to reprepare when the moment initially chosen for preparation appears to fall short of the moment at which the stimulus is actually presented.