Purchase this article with an account.
Johannes Burge; Inferring the shape of decision variable distributions from psychometric functions. Journal of Vision 2018;18(10):209. doi: https://doi.org/10.1167/18.10.209.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The psychometric function is a mainstay of psychophysics and behavioral neuroscience research, and signal detection theory (SDT) is a widely used tool for analyzing and interpreting psychometric data. SDT is perhaps most important for inferring the statistical properties of decision variables that are posited by the theory to underlie behavioral performance. Decision variables are usually most conveniently modeled as being Gaussian distributed. In many cases, however, there are theoretical and empirical reasons to assume that decision variables are non-Gaussian. How do non-Gaussian decision variables impact the shape of the psychometric functions collected in behavioral experiments? And how does spatial (or temporal) uncertainty, an important factor in many target detection tasks, complicate the ability to determine the shape of the decision variable distribution from the shape of the psychometric function? Here, with a series of computational analyses, we demonstrate how non-Gaussian decision variable distributions determine the shape of psychometric functions in the presence (or absence) of arbitrary levels of uncertainty. We also show how well the shape of the psychometric function can be used to infer the shape of the decision variable distribution in the presence (or absence) of arbitrary levels of spatial or temporal uncertainty. The results provide a guide for interpreting the results of psychophysical experiments.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only