Abstract
Purpose: To investigate the tradeoff between speed and accuracy of perceptual decisions we performed combined measurements of accuracy and response time (RT) in a motion discrimination task. To facilitate comparison with physiological experiments in our laboratory, we used the same random dot stimulus and the same eye movement response paradigm in humans and monkeys. Incorporating measurements of RT provides further constraints on models of temporal integration and allows for links to the underlying physiology.
Methods: Human observers made 2-alternative direction discriminations on stochastic moving dot displays. Difficulty was controlled by changing the percentage of dots moving coherently in a common direction. Subjects indicated their judgment by making an eye movement to one of two peripheral targets.
Results: Psychophysical performance was similar in monkeys and humans. At the highest motion strengths, accuracy was perfect and RT was fastest. At slightly lower motion strengths, accuracy remained near perfect but RT slowed. At motion strengths near discrimination threshold, both accuracy and RT was affected. At the lowest motion strengths, accuracy fell to chance and RT was slowest. These results are consistent with an additive process that accumulates evidence over time about the direction of motion. In contrast, the data are not consistent with the decision being determined by the first occurrence of a suprathreshold event (“probability summation”).
Conclusions: The constellation of accuracy and RT results can be explained by an additive accumulation of evidence bearing on the two alternatives. This accumulation may have a neural basis in the buildup of neural activity observed in the monkey lateral intraparietal area (LIP) during this task.
Acknowledgements: HHMI, NEI #EY11378, NCRR #RR00166