Comparing scanpaths during scene encoding and recognition: A multi-dimensional approach

When viewing a stimulus too large to be apprehended in a single glance, observers make a series of fixations and saccades which occur in a particular sequence and which are associated with a location, duration and so on. In two-dimensional stimuli such as pictures and websites the placement and characteristics of these scanpaths show a large degree of variation both between and within individuals. A good deal of progress has been made in understanding this variation in looking patterns, both in terms of the “bottom-up” visual features which make certain locations more likely to be fixated, and in terms of the contribution of “top-down” processes such as the current task demands and the expertise of the viewer (Groner & Menz, 1985; Ballard & Hayhoe, 2009; Itti & Koch, 2001; Tatler, Hayhoe, Land, & Ballard, 2011; Underwood, Foulsham, & Humphrey, 2009). There has also been increasing interest in potential individual differences in eye movements, with findings indicating that each particular observer may move their eyes similarly in different tasks (Andrews & Coppola, 1999).