Leading the way to the target: eye movements in visual search

Abstract

Searching for objects is part of our daily life. Generally, this cannot be accomplished without eye movements. These eye movements form a path towards the target. In this thesis we focus on several factors that shape this path. Eye movements in search often aim for objects resembling the target. Nevertheless, observers also frequently select conspicuous objects that do not resemble the target, particularly following short saccade latencies. In Chapter 2 we focus on the selection of conspicuous objects as a function of the saccade latency. We find that following short saccade latencies only conspicuous objects that can be processed quickly are selected. While following longer latencies less conspicuous objects requiring longer processing time are also selected. This suggests that the temporal aspects of encoding visual information are a determining factor in saccadic selection. When a target is surrounded by similar objects this can deteriorate the quality of peripheral information, a phenomenon known as crowding. When a target is inconspicuous due to crowding the selection of conspicuous objects will not result in an efficient strategy. However, little research has been devoted to the role of crowding in search. In Chapter 3 we show that when a target can be distinguished peripherally, eye movements are indeed biased towards locations where there is little crowding and the target is more conspicuous. Interestingly, when the target cannot be discriminated peripherally, this bias reverses to areas with strong crowding, where the target is least conspicuous. The surrounding objects are only part of what determines the perception of the target. The background of the target is also an important factor. In Chapter 4 we investigate the role of the background in search. Manipulating the orientation or spatial frequency of the background (while keeping the properties of target and distractor constant) results in search times that are longest for a small target-background contrast. When varying the background luminance, however, we find that the longest search times occur when the background is darker than both target and distractors, regardless of the target-background contrast. Thus, background plays an important role in search. This role depends on the individual contrast of both target and distractors with the background and the type of contrast (luminance, orientation or spatial frequency). Finally, it is tempting to assume that each eye movement is the result of analysis during the directly preceding fixation. However, it is also possible that a single decision is made for several consecutive eye movements. In Chapter 5 we demonstrate that the latter is in fact the case. We show that up to three eye movements are planned ahead in visual search. Furthermore, the ability to deviate from the planned path depends on whether new or not visual information can be processed sufficiently fast. In all, we have presented studies focusing on fundamental aspects of selection by eye movements that have not received much attention in previous studies using artificial displays.