Join the seminar online

Abstract

Eye tracking can be a fantastic method to gauge visual attention: humans tend to look where they attend, so tracking gaze is highly informative. Unfortunately, decent-quality data requires specialised equipment that can only be deployed in lab settings. Where other research has scaled through web-based data collection, its reliance on dedicated hardware prevents eye-tracking studies to do the same. One potential avenue has been webcam-based eye tracking, but this has proved prone to poor data quality and exceptionally high attrition (losing around 50% of a sample is not uncommon). Fortunately, there exists a more inclusive alternative: mouse-viewing. This technique blurs a stimulus screen outside of a cursor-locked aperture, in an attempt to mimic the high-resolution fovea and its blurry periphery. In this talk, I will highlight a variety of studies into this approach that show which aspects of visual attention can and cannot be captured. I will also quantify how the reliabilities of measures derived from mouse and gaze compare.

Biography

Edwin Dalmaijer is a Senior Lecturer at the University of Bristol. Previously, he read for a DPhil in Experimental Psychology at the University of Oxford and conducted postdoctoral work at the MRC Cognition and Brain Sciences Unit at the University of Cambridge. His research interests are best summarised as the quantitative exploration of development, both within individuals and in populations. Broadly, he investigates how affective and cognitive faculties impact each other, and how they are affected by the environment. Edwin triangulates problems with narrowly focused experiments aided by computational models of behaviour, machine learning to find complex patterns in large datasets, and agent-based population simulations.