When time is money! Dynamic integration of sensory evidence and diminishing rewards in perceptual decision making

碩士 === 國立陽明大學 === 神經科學研究所 === 101 === Why do we need time? Because we want to collect information and it takes time. Why do we want more information? Because it could help making better decisions. Taking more time before making a response could be beneficial because it often allows more information...

Full description

Bibliographic Details
Main Authors: Ya-Hsuan Liu, 劉雅瑄
Other Authors: Shih-Wei Wu
Format: Others
Language:en_US
Published: 2013
Online Access:http://ndltd.ncl.edu.tw/handle/11900831354122497715
Description
Summary:碩士 === 國立陽明大學 === 神經科學研究所 === 101 === Why do we need time? Because we want to collect information and it takes time. Why do we want more information? Because it could help making better decisions. Taking more time before making a response could be beneficial because it often allows more information to be gathered which in turn could improve performance. Taking more time could also be costly because time often acts as a limited resource that competes against others. Research on perceptual decision making (PCM) has made tremendous progress to the understanding of the neurobiological foundation of decision making in the context of information accumulation. However, the tradeoff between information gathering and time cost remains little known. In this study, we investigated how humans perform this tradeoff in a random-dot motion (RDM) task, a well-established paradigm for studying information accumulation. Critically, we designed the task such that the reward for making correct judgment would decrease over time. This decreasing reward schedule was designed to resemble time cost. To perform well in this task, the subjects needed to choose a response time that appropriately balanced information gathering and the cost of time. To evaluate performance, an optimal model of response time was developed. We found that subjects (n=22) changed response time in the direction consistent with the model, suggesting that integration computations takes place in this context. However, we also found that subjects tended to be slower than they should when required to make very fast response, whereas the pattern was opposite when there was less pressure for timely response. Two models were proposed to explain these two distinct suboptimal patterns.