Purchase this article with an account.
Jeff Pelz, Roxanne Canosa, Jason Babcock; Perceptual strategies in complex, extended tasks. Journal of Vision 2002;2(10):127. doi: https://doi.org/10.1167/2.10.127.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The dynamics and characteristics of eye movements in brief tasks performed under reduced laboratory conditions are relatively well-known. Less is understood about oculomotor behavior in natural tasks. In the real world, vision is a powerful tool for extracting information and testing hypotheses in real time; in the laboratory vision is often the instructed task. The most direct method for studying vision in the natural world is to monitor subjects as they perform tasks outside of the laboratory. To this end, we have developed a wearable system to record eye movements of subjects while allowing natural eye, head, and body movements. The experiments performed with the wearable eyetracker reveal that low-level metrics (e.g., fixation duration and saccade amplitude) vary by task, but there is much agreement with laboratory studies. A notable exception is the occurrence of very short fixations. Previous studies have reported fixations less than 150 msec only under special conditions (e.g. ‘express saccades’), but fixations < 150 msec were common in the real world.
The tasks used in traditional laboratory studies are measured in msec or seconds; real-world tasks can take minutes or hours. Extended tasks may benefit from the execution of high-level routines. Our studies show that high-level strategies emerge during complex, extended tasks. While most fixations are related to an immediate action, some are made on objects relevant for future actions. It is evident that these ‘look-ahead’ fixations represent a task-dependent strategy and not a general behavior elicited by the salience of objects in the environment as posited by bottom-up models of visual attention. Looking ahead to objects of future relevance may support the conscious percept of an environment seamless in time as well as in space. Identification of such high-order perceptual strategies demonstrates the value of studying visual perception during ongoing complex, natural tasks.
This PDF is available to Subscribers Only