Abstract
Using cortical microstimulation (µStim), Newsome et al. showed that direction-selective neurons in area MT represent the evidence upon which monkeys base their decision about direction in a stochastic random-dot motion task. Near threshold, µStim caused monkeys to favor the direction preferred by the stimulated neurons. Our goal is to understand the mechanism underlying the transformation of such sensory evidence to a decision. Insight into this process can be gained by measuring the amount of time it takes to reach a decision. We trained monkeys on a version of the random-dot motion task that allowed us to measure both sensitivity and reaction time (RT). In contrast with previous experiments, the monkey controlled viewing duration by making an eye movement to one of two choice targets when it was prepared to indicate its decision. The percentage of coherently moving dots was varied over a range that spanned psychophysical threshold and µStim was applied on half of the trials. Consistent with previous reports, µStim caused the monkey to bias its decisions in favor of the preferred direction of neurons near the stimulating electrode. Microstimulation also affected RT in a manner that depended on the direction of motion. When motion was in the neurons' preferred direction, µStim reduced RT for correct choices. When the motion was in the opposite (null) direction, µStim prolonged RT. These findings suggest that the monkey forms its decision by integrating motion information coded in area MT until the accumulated weight of evidenced is sufficient to commit to a behavioral response. The prolonged RT on null direction trials suggests that MT information is not only used as evidence in favor of the preferred direction, but also as evidence against the opposite direction. The effect of µStim on both the speed and accuracy of decisions lends new insight into the neural mechanisms responsible for reading the visual cortex.
Supported by DFG (DI 819/1-1), HHMI, NCRR (RR00166), NEI (EY11378), and Poncin.