Modification of response time variability in a decision-making task.

Previous single-unit recordings in monkeys showed that essential information regarding a decision is available earlier to posterior parietal cortex than expected based on simultaneously measured behavioral response times (RTs). We propose that this observation reflects a tendency of the brain to reduce RT variability in repetitive response sequences. To test this, we studied the effects of experimentally introduced variability in trial duration on RTs in humans. We found that humans adapt timing based on the timing of surrounding trials, essentially reducing RT variability in trial sequences. This implies that RTs do not always reflect the minimal time it takes to make a decision. Implications for the interpretation of behavioral data are discussed.