A comparison of equivalent noise methods in investigating local and global form and motion integration

Date
2022-11-15
Advisor
Supervisor
Co-Advisor
Co-Supervisor
Instructor
Source Title
Attention, Perception, & Psychophysics
Print ISSN
1943-3921
Electronic ISSN
Publisher
Springer New York LLC
Volume
83
Issue
1
Pages
152 - 165
Language
English
Type
Article
Journal Title
Journal ISSN
Volume Title
Series
Abstract

Static and dynamic cues within certain spatiotemporal proximity are used to evoke respective global percepts of form and motion. The limiting factors in this process are, first, internal noise, which indexes local orientation/direction detection, and, second, sampling efficiency, which relates to the processing and the representation of global orientation/direction. These parameters are quantified using the equivalent noise (EN) paradigm. EN has been implemented with just two levels: high and low noise. However, when using this simplified version, one must assume the shape of the overall noise dependence, as the intermediate points are missing. Here, we investigated whether two distinct EN methods, the 8-point and the simplified 2-point version, reveal comparable parameter estimates. This was performed for three different types of stimuli: random dot kinematograms, and static and dynamic translational Glass patterns, to investigate how constant internal noise estimates are, and how sampling efficiency might vary over tasks. The results indicated substantial compatibility between estimates over a wide range of external noise levels sampled with eight data points, and a simplified version producing two highly informative data points. Our findings support the use of a simplified procedure to estimate essential form-motion integration parameters, paving the way for rapid and critical applications to populations that cannot tolerate protracted measurements.

Course
Other identifiers
Book Title
Keywords
Internal noise, Sampling efficiency, Form-motion integration, Equivalent noise methods
Citation
Published Version (Please cite this version)