关键词: IOR, 三维空间, 虚拟现实技术, 视觉搜索

Abstract: Inhibition of return (IOR), typically explored in cueing paradigms, is a performance cost associated with previously attended locations. It has been suggested as a crucial attentional mechanism that biases orientation towards novelty. Our living environment turns out to be a complex three-dimensional space which entails ecological functioning of visual search. In Wang et al.'s (2015) three-dimensional experiment, participants' attention was (re)oriented in a straight line along the depth plane to induce location-based IOR. However, their operation could have an inconsistent depth distance, contaminating the IOR effect. Our study thus aimed to explore the range of IOR spreading in the depth of three-dimensional space.
Two depth distances were manipulated in the study: short distance for Experiment 1 and short distance for Experiment 2. The types of cue-target were divided into Valid (the target appeared in both the same depth plane and hemispace of the vision as the cue), X-Invalid (the target appeared in the same depth plane, but different hemispace of the vision as the cue), Z-Invalid (the target appeared in the same hemispace of the vision, but in different depth plane as the cue), and X-Z-Invalid (the target appeared in the different depth plane and hemispace of the vision). The cue-target in depth of target (closer depth plane vs. farther depth plane) was crossed with the cue-target types, forming a 2 × 4 factorial design in the two experiments. At the beginning of each trial, either of the four locations in the near or far space was cued for 300ms, and was followed by an interval of 200ms inter-stimulus. Then the spatial location of the central fixation cross was cued for 200ms to divert attention away from the previously by cued location. After another period of 150 or 250ms, the target was presented for 250ms at one of four locations, either cued or un-cued, with equal probability for four cue types. The validity of clues was 25% in the two experiments.
Our results revealed a main effect of cue types in both the short, F (3, 28) = 13.914, p < .001, and long-distance conditions, F (2, 60) = 7.691, p = .001. The reaction time of Z-invalid cues was significantly longer than that of X-invalid cues in short distance, p < .001, while in the long-distance condition there was no significant difference. The amount of IOR, which was obtained by subtracting the mean response time of invalid cues from that of valid ones, was used as the indicator of the IOR effect. The amount of IOR in Z-invalid condition in short distance was significantly smaller than that in long distance condition. Regardless in the near (p = .017) or far (p = .011) space conditions, the amount of IOR of Z-invalid cues in short distance was significantly smaller than that in long distance. The main findings were replicated in the Experiment 3, in which the objects at the close plane were subtending the same visual angle as those at the far plane.
In conclusion, the effect of IOR has a spread range in depth in three-dimensional space, such that IOR spreads from the cued position as depth distance extends.

Key words: inhibition of return (IOR), three-dimensional space, virtual reality technology, visual search