Portal de Periódicos da CAPES

How to use MATLAB to fit the ex-Gaussian and other probability functions to a distribution of response times

2008; University of Ottawa; Volume: 4; Issue: 1 Linguagem: Inglês

10.20982/tqmp.04.1.p035

1913-4126

Yves Lacouture, Denis Cousineau,

Spectroscopy and Chemometric Analyses

This article discusses how to characterize response time (RT) frequency distributions in terms of probability functions and how to implement the necessary analysis tools using MATLAB.The first part of the paper discusses the general principles of maximum likelihood estimation.A detailed implementation that allows fitting the popular ex-Gaussian function is then presented followed by the results of a Monte Carlo study that shows the validity of the proposed approach.Although the main focus is the ex-Gaussian function, the general procedure described here can be used to estimate best fitting parameters of various probability functions.The proposed computational tools, written in MATLAB source code, are available through the Internet.In recent years there has been an upsurge of interest in the response times (RT) of cognitive processes.This can be attributed in part to the wide availability of computer programs that allow experiments to be conducted automatically and RT to be measured with millisecond precision.However, and perhaps more importantly, many researchers see RT as major constraints when testing models of cognitive processes.Although RT are now routinely measured and frequently reported, there is a lack of standard tools for characterizing RT distributions.The difficulty is that simple descriptive statistics usually do not provide an adequate characterization of the data, and