Thumbnail Image 1


Dynamic integration of sensory evidence and reward

In this project, we investigate how people trade off the cost and benefit of information when making decisions. Here, both the cost and benefit of information are linked with time: More time is associated with better information quality that helps performance. However, more time is also associated with bigger cost, as the amount of monetary reward subjects would receive decreases over time.

We call this type of choices – ‘Decisions about when’. Below we describe a task developed (jointly with Alireza Soltani) to answer this question.

Imagine there is a box with many red and green balls inside. On each trial, we sample from the box and “sequentially” present the outcomes on the screen. The white circle in the periphery represents the amount of available time and would start to decrease upon the presentation of the red and green dots (see figure 1).

What subjects have to do in this task is to choose which ball – red or green – has more in number in the box. The more information she or he collects, the more accurate the choice. Hence, the more time given to the subjects, the higher the probability of making a correct judgment.


There are two sessions included in this experiment. In the first session (Speed-Accuracy Tradeoff, SAT session), we aim to investigate how subjects’ performance (judgment accuracy) changes as a function of time. In order to do so, we implement five different time limits – 1s, 3s, 5s, 7s, and 9s,

Before the experiment, the subjects are given the following instruction: there were 100 of red and green balls in the box. On each trial, the ratio of red/green balls is either 55:45 or 45:55. The computer would sample (without replacement) 10 balls per second from the box and sequentially present – one ball at a time – on the screen.

Subjects have to indicate his or her decision within the time limit. The subjects would receive 60 points for making a correct judgment and would lose 60 points for making an incorrect judgment.







In the second session (decreasing-reward session), the amount of reward for making a correct decision decreases over time. This manipulation allows us to investigate how subjects trade off the cost of time – manifested as decreasing reward – against the benefit of spending more time to collect more information. Different from Session 1, the peripheral white circle now represents the amount of reward.





Back to Top