Abstract
Performing a specific task in a real-world environment necessarily means selectively sampling and using only a small subset of the information available in the scene. An interesting question is how the selectivity of that sampling is affected by familiarity, both with the environment and with the task itself.
To address this question we designed an experiment in which participants performed one of two complex everyday tasks (using either a microwave or a coffee machine), in a real-world environment (a mock-kitchen) populated with a combination of task-relevant and task-irrelevant (distractor) objects (see Supplementary Figures). Participants were either familiar with both the environment and the correct equipment for their task (the familiar condition), familiar with the environment and the equipment for the other task (the switched condition), or familiar with neither (the control condition). While participants performed these tasks, we tracked both their gaze patterns and their positions
in the environment.
Analyses show that, as expected, familiar participants performed the tasks significantly faster than participants in the switched condition. Rather surprisingly, however, participants in the switched condition were also slower than those in the control condition, suggesting the possibility that familiarising to the requirements of another task interfered with their actual task. Familiar participants spent less time fixating both relevant and irrelevant objects compared to the control condition, whilst switched participants spent less time fixating irrelevant objects than control participants, but more time fixating relevant objects. This difference is mostly due to a small number of task-critical sub-object regions.
In sum, we show selectivity of information use for behavior (in terms of gaze patterns) following familiarisation in a real-world task, which will enable transfer to Virtual Reality of real-time
availability of task-relevant information and control of interfering task-irrelevant information.