Abstract
We often need to make decisions based on visual information we get from continuously changing environments. We here explore how the uncertainty of objects’ speed affects decision making in a go/no-go task by analogy to crossing a busy street. Participants stood in front of a large screen, and saw in each trial two groups of three targets approaching the midline of the screen (one group moving leftwards and the other moving rightwards). In the bottom-centre of the screen there was a square that responded to the participants’ go/no-go response. Participants had to decide whether to make the square cross the screen in 300 ms without being hit by the moving targets (go) or not (no-go), by pressing one of two keys of an input device accordingly. We manipulated the environment’s uncertainty by having each group of targets with no, medium or high variability between the targets’ speeds, resulting in 6 conditions (no-no, no-medium, no-high, medium-high, medium-medium, high-high). The average time it took for the groups of objects to reach the midline of the screen (TTM) was 0.6, 0.74, 0.9, 1.10, or 1.34 sec. The conditions and TTM were randomly interleaved across trials. If participants succeeded in making the square cross the screen they won 50 points, but if the square was hit by the moving targets participants lost 200 points. No-go responses were neither rewarded nor penalized. We fit psychometric curves to participants’ go responses and compared them with what would be an optimal decision based on the probability of successful go responses across TTM and the reward. Results show that as soon as variability is introduced in a group of targets, the amount of no-go responses increases and decisions deviate from optimal (i.e. under-confident decisions). However, the variability does not influence the number of successful crossings.
Acknowledgement: Funding was provided by the Catalan Government (2017SGR-48) and grant PSI2017-83493-R (AEI/Feder, UE).