| Abstract No.: | A-G1200 |
| Country: | Canada |
| | |
| Title: | KINEMATIC EFFECTS OF DECISION-MAKING IN MONKEYS |
| | |
| Authors/Affiliations: | 1 Alex Pastor-Bernier*; 1 Paul Cisek;
1 University of Montreal, QC, Canada.
|
| | |
| Content: | The objective of this project was to test the hypothesis that competitive interactions between representations of alternative actions depend on spatial parameters related to the options (spatial position) and biasing factors to each option (size of reward), and whether they influence movement kinematics. To test this hypothesis a monkey was trained to perform a reaching decision task which involves moving a cursor from a central circle to one of eight possible target locations. At the beginning of each trial the monkey places the cursor in the center and two targets appear. Each target was associated with different rewards (1-3 pellets) indicated by different target border styles. The monkey has to keep the cursor in the center until the targets change color (GO signal). Then, it must move to one of the targets and hold the cursor there to get the reward. In one variant of the task the monkey is presented with only one target (1-T) whereas in a second variant two targets are presented (2-T), and the monkey is free to move to either of them after GO (free trials). In a third variant two targets are presented, but one disappears after GO leaving the monkey with only the remaining target to go to (forced trials). In 1-T trials, as the reward size increased we did not observe a significant decrease in reaction time (RT) but found a significant decrease in movement time (MT) (p<0.01), indicating that the monkey was more motivated and moved more quickly when the target was more valuable. In 2-T forced trials, MTs were also significantly shorter for trajectories towards high valued targets. In some forced 2-T trials the monkey was forced to move to the less valuable target, and in many of these, we observed that the initial movement was in the direction of the high valued target and then curved toward the remaining target. This effect was most pronounced in the trials with the shortest RTs. This could be explained in terms of neural activity as follows: When the two targets appear, two populations of cells in dorsal premotor cortex (PMd) are activated simultaneously. Just before GO, the population tuned to the high value target is the strongest, biasing the directional activity in PMd towards that target. If the high value target vanishes after GO, the activity of this population will gradually decrease. Therefore, early movements (low RT) will be more biased by that target than later movements (high RT) initiated after the activity tuned to that direction has decreased. This is consistent with the hypothesis that the competition between two options is performed by a tuned neuronal population and takes place in the same region that implements the action. |
| | |
| Back |
|
