Same-object Benefit Or Cost? The Object-based Attention In The Attentional Blink

When two targets (T1 and T2) are inserted in a rapid serial visual presentation (RSVP) stream of distractors, T2's identification is impaired when it is presented within 200-500ms after T1, this transient deficit in performance is called the attentional blink (AB). When the target belongs to the attended object, it is usually identified more quickly and more accurately than when it belongs to the other objects. This is typically labeled as the same-object benefit effect.There had been found that the target identification appearing on the same-object was much accurate than presenting on the different-object when it was presented in the AB, indicating the AB same-object benefit. However, it is not the case in Conci et al (2009)'s study, in which same-object T2 identification in the AB was much more impaired, relative to the different-object T2 identification, revealing the AB same-object cost. It is counterintuitive, and contrary to previous findings of object-based attention using the cue-target paradigm, in which a same-object benefit was usually obtained within the SOA of 200-600 ms between cue and target.We assume that the object file is "locked" only when the period of T1 consolidation overlaps the period of establishing object representation (i.e. the bars or rectangles). The possible mechanism is that the representations of T1 and the rectangle are integrated to create a new object file. This process of creating new object-file leads to a refractory period of the whole object-file, because of the resources consuming process of T1 consolidation, and causes a same-object cost. In contrast, if the object representation has been stably established before the processing of T1, the operation of the object-file is only updating rather than creating during the T1 consolidation, and thus the object-file is not "locked". At this time, the same-object benefit would be observed as in the traditional cue-target paradigm.According to this assumption, whether the object representation is stably established in a short time is a critical factor determining whether a same-object cost or a same-object benefit is observed. To test this,5 experiments were conducted in the present study.Experiment 1 was to replicate the basic finding of Conci and Miiller, i.e. the same-object cost, in our laboratory setup. In the present experiment, we adopted a similar RSVP-variant of the spatial-cueing paradigm to expect a similar result as Conci and Miiller's.Experiment 2 was to examine whether a same-object benefit would be obtained at the long SOA between rectangles and T1 by previewing rectangles for 1 s (cf. Shomstein & Behrmann, 2008; Shomstein & Yantis,2004) before the onset of the RSVP stream. In Experiment 2, we presented the rectangles for 1 s to establish a stable object representation, and predicted that it could guide attentional selection with its overpoweringly beneficial influence in the relatively complex RSVP streams. Experiment 3 had two purposes. One of them was to test whether the crucial factor was the SOA between rectangles and T1 rather than that between rectangles and RSVP streams to influence on the object representation. The second one was to directly compare the effects of short and long SOA between rectangles and T1 within subject and within the same experimental session in the mixture manner. We predicted that if there is not ample time to establish an object representation (i.e., the SOA between rectangles and Tlwas short), then the same-object cost would be observed. If, however, sufficient time is to be provided to establish a robust object representation, then the same-object benefit would be observed.In Experiment 4, we outlined two bars in color (red or green) to further investigate the determinant factor of object representation on AB same-object effect. In Experiment 5, we decreased the difficulty in processing T1 to further examine whether the object representation was locked by T1 processing to lead to AB same-object cost.These results suggested that establishing object files which represents visual object information is a time-and resources-consuming process, and then dynamically affects the target processing in the AB task. If the object representation is not stably established, its object file is apt to be locked by the processing of T1 which is presented at a short interval, leading to a totally new object file. The process of creating new object files causes more impaired AB, demonstrating the AB same-object cost. Once the object representation is stably established, however, the subsequent T1 processing makes its object file updated, not discarded. Updating the pre-object file is a relatively less-consuming resources process, which enhances the processing of T2 within the same object, revealing the same-object benefit.