Manipulation of P3 latency: speed vs. accuracy instructions.

Twelve young female subjects were presented with a series of horizontal line-pairs of same or different length in a two-alternative, forced-choice RT task, with 60 of each type pair in each block of trials. In one block (Easy) lines differed by 30%, in another block (Difficult) lines differed by 7%. Subjects were first given 60 practice trials with the Easy discrimination and with the instruction that speed and accuracy should be emphasized equally. For the next block of trials, accuracy was emphasized with a monetary bonus for accurate performance. Finally, in the last block of trials, speed was emphasized with a monetary bonus for speedy performance. Additionally, a penalty was incurred for RTs that exceeded a criterion level based on each individual subject's performance. The order of Easy and Difficult discrimination blocks was maintained within a subject but balanced across subjects. From the latency-adjusted P3s recorded from Pz, we obtained P3 latencies, amplitudes and single-trial P3 latency/RT correlations. RT to correct and incorrect trials and error data were also collected. P3 was considerably larger during the Speed than Accuracy conditions. The single-trial P3 latency/RT correlation was higher in Speed than in Accuracy runs. RT was 235 msec faster and P3 was 40 msec earlier during the Speed than during the Accuracy runs. On the other hand, discrimination difficulty delayed P3 and RT about equally, 28 and 43 msec respectively. This pattern suggests that speed instructions and discrimination difficulty affect stimulus processing time and response production time differently.