WinVis — a novel approach to designing software for psychophysical experiments

Abstract

 

Dramatic improvements in the graphics performance of desktop computers offers a unique opportunity for vision scientists to avoid costly duplication of effort by adopting a platform that promotes the sharing of software and minimizes hardware costs. To this end we have been developing WinVis, a software platform that uses object-oriented techniques to simplify and expedite the design of psychophysical and physiological experiments. WinVis retains flexibility through multiple levels of user access, from experiments designed over the WEB using standard Internet browsers to lower level programming using Matlab extensions. To promote sharing of ideas and effort, the system provides WEB access to a database of stimulus classes for users to incorporate into experiments. Stimulus classes are easily combined and extended to create new stimulus classes which themselves become database entries. The stimulus classes inherit functionality from one another, a feature of object oriented design. The result is an expanding database, the utility of which increases with use. Trial and experiment classes are also contained in the database for online experiment design. The user provides design criteria (on the WEB), a remote program generator creates a fully executable program that is delivered to the users local computer. The goal is to make it easy for researchers to exchange stimulus/experiment classes via the database to validate and extend each other's research findings. With regular database updates, researchers will be able to view stimuli described in the latest vision science journals. WinVis includes a Matlab extension for performing psychophysical experiments. This extension is integrated with the WEB based stimulus/trial class database system to enhance its functionality. The application of an object oriented design strategy in defining visual stimuli provides an efficient way to build upon existing stimulus objects.

 

Footnotes

 Supported by: NIH grant EY12339