Purchase this article with an account.
Alexander C. Huk, Michael N. Shadlen; Temporal integration of visual motion in macaque parietal cortex. Journal of Vision 2004;4(8):110. doi: https://doi.org/10.1167/4.8.110.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Discrimination of the direction of dynamic random dot motion depends on the accumulation of evidence favoring one direction over the other. Neurons in the lateral intraparietal area (LIP) of the macaque appear to represent the underlying decision variable. We hypothesized that LIP neurons represent the time integral of the noisy directional signals that are present in visual cortex. To test this, we recorded from LIP neurons while monkeys performed a 2-alternative direction-discrimination task in which we systematically perturbed motion strength over the course of each trial. A dynamic moving dot display, which varied randomly in direction and strength across trials, was superimposed on a dim dynamic random texture background. On 2/3 of trials, a 100 ms-long motion pulse was added to this background texture in the same or opposite direction of the dots. Monkeys were rewarded for choosing the correct direction of dot motion, and were thus encouraged to ignore the background motion pulses. Nonetheless, motion pulses biased judgments and response times. As previously shown for unperturbed dot displays, LIP responses ramped linearly upwards for choices that corresponded to an eye movement response into the neuron's response field, and ramped downwards for opposite choices, with slopes dependent on the strength of dot motion. The background motion pulses caused a brief change in firing rate, deflecting responses upwards or downwards, depending on pulse direction. This perturbation persisted as a fixed offset in the spike rate, often apparent for ∼500 ms after offset of the background motion pulse. Pulses inserted at different times during dot viewing exerted similar effects, suggesting that the integration process weights motion equally over the course of the trial. These results suggest that both decisions and LIP responses reflect a process of temporal integration with a time constant that is long relative to the time used to make decisions in this task.
This PDF is available to Subscribers Only