Current understanding of eye guidance

Visual sampling of the scene is at a premium in human vision. High acuity vision is restricted to the small foveal region at the centre of vision and the ‘‘saccade and fixate’’ strategy means that this window of high-resolution vision is only directed at an average of three to four locations in each second. It is therefore not surprising that understanding the factors that underlie decisions about where to allocate this valuable resource of foveal vision have been at the heart of eye movement research since the earliest objective recordings (Wade & Tatler, 2005). The earliest work on viewing complex scenes found that there were certain locations in scenes that were consistently looked at by most observers (Buswell, 1935). Buswell called these locations ‘‘centers of interest’’ and asked what it was about these locations that made them ‘‘interesting’’. He considered that there were two possibilities: That it was something external in the stimulus that attracted the viewers’ eyes; or that it was something more internal to the viewer that reflected higher level cognitive ‘‘interest’’. Buswell favoured the latter explanation, and he famously showed that presenting the same image but with different instructions fundamentally changed the places that a viewer fixated; an observation that was confirmed by Yarbus (1967). Given this demonstration and early favouring of top-down control of where we place our eyes in complex scenes, it is perhaps surprising that much of the subsequent consideration of how we view complex scenes has focused on the role of low-level factors in guiding attention. Why might this be? A wealth of literature on visual search demonstrated that basic visual features