Illusory conjunctions reflect the time course of the attentional blink

Abstract

Illusory conjunctions in the time domain are binding errors for features from stimuli presented sequentially but in the same spatial position. A similar experimental paradigm is employed for the attentional blink (AB), an impairment of performance for the second of two targets when it is presented 200-500 msec after the first target. The analysis of errors along the time course of the AB allows the testing of models of illusory conjunctions. In an experiment, observers identified one (control condition) or two (experimental condition) letters in a specified color, so that illusory conjunctions in each response could be linked to specific positions in the series. Two items in the target colors (red and white, embedded in distractors of different colors) were employed in four conditions defined according to whether both targets were in the same or different colors. Besides the U-shaped function for hits, the errors were analyzed by calculating several response parameters reflecting characteristics such as the average position of the responses or the attentional suppression during the blink. The several error parameters cluster in two time courses, as would be expected from prevailing models of the AB. Furthermore, the results match the predictions from Botella, Barriopedro, and Suero's (Journal of Experimental Psychology: Human Perception and Performance, 27, 1452-1467, 2001) model for illusory conjunctions.

Individual Differences in the Attentional Blink: The Temporal Profile of Blinkers and Non-Blinkers

“…Although the finding of suppression as a function of lag corresponds with findings from previous studies [43], [44], [50], all of these papers reported substantially more suppression.…”

“…Whereas both our study and that of Popple and Levi [49] employed integral dimensions of the stimuli as the relevant features (colour and shape), Vul et al [43] as well as Chun [44] used composed targets (a letter surrounded by an annulus or coloured frame). Although the study by Botella and colleagues [50] did use colour as an integrated target feature, they introduced a task-switch by varying the colour of the two targets, and possibly reduced the effectiveness of colour as a target-specific feature by also varying the colour of each distractor in the stream. It is thus not inconceivable that the latter studies introduced additional factors into the AB task that further complicated the binding and subsequent selection of targets.…”

“…Errors in the temporal selection process have been studied before [44], [49], [50], but individual differences were not considered and differences in the total number of errors were not controlled for. Given that intrusions of items following T2 are inherently related to the total number of errors made, we studied the relative number of intrusions, allowing comparisons between blinkers and non-blinkers in the type of intrusions irrespective of the total rate of response errors.…”

“…In EEG studies, suppression is reflected in the P3 component that is absent or strongly attenuated during the AB [53], and also the n2pc (associated with the allocation of attention) is known to be affected [55]–[57]. Similarly there is quite some evidence supporting the idea that attentional selection is delayed during the AB, provided by behavioural studies [27], [43], [44], [50], [58]–[60] and EEG studies [3], [53], [61], where the latter have revealed that when the second target was reported correctly at short time intervals, the P3 component was delayed in comparison to longer intervals.…”

Attention blinks for selection, not perception or memory: Reading sentences and reporting targets.

“…Several studies of the attentional blink (e.g., Chun, 1997) have noted that such misbindings are increased under conditions that produce an attentional blink, so that the failure to report T2 is sometimes because of the report of T2 + 1 or T2 − 1 instead. Such misbindings do not account for the whole of the attentional blink effect in those studies, however (e.g., Botella, Privado, Gil-Gomez de Liano, & Suero, 2011; Vul, Nieuwenstein, & Kanwisher, 2008). Nonetheless, we carried out analyses in which report of the T2 + 1 or T2 − 1 word was counted as correct, along with report of the correct word.…”