CNV and temporal uncertainty with 'ageing' and 'non-ageing' S1-S2 intervals.

OBJECTIVES Uncertainty about the timing of a known external event is an everyday phenomenon but has been rarely investigated with electrophysiological methods. We studied how the amplitude of the contingent negative variation (CNV) is affected by temporal variation of S2 presentation. Competing hypotheses about the development of CNV during the foreperiod until S2 presentation were that CNV either would follow a monotonic trend, be it increasing or decreasing, or alternatively that the time-course of CNV would be affected by the probability with which S2 was presented at each time-point in a given task. METHODS The interval between cueing stimulus and imperative stimulus was randomly chosen from 3 different values between 1.3 and 2.6 s, using 3 different probability distributions in separate blocks: an 'ageing', a 'non-ageing' and a 'Gaussian' distribution. RESULTS As previously shown, reaction times were determined by the probability of the imperative stimulus at the given length of the foreperiod. The same was found for CNV amplitude: the effects of temporal uncertainty on CNV mainly depended on the particular distribution of temporal probabilities used in a block. The relevant parameter was the a posteriori probability of event occurrence, very similar to the effects of this parameter on response times. In fact, the major part of the effect of a posteriori probability on CNV was common variation of CNV and response times. CONCLUSIONS Thus, under temporal uncertainty the amplitude of CNV reflects the subjective expectancies for the occurrence of a given event, with this variation being related to variations in response times.