Our brains constantly receive vast amounts of information from our surroundings. Attention allows us to selectively process this information, enabling us to ignore any irrelevant information, so we can interpret and attend to more important information.
However, attention has temporal limits, this is demonstrated by a phenomenon known as the attentional blink (AB). AB can be investigated using a rapid serial visual presentation (RSVP) methodology. Stimuli such as letters, digits, words or pictures are briefly presented in the same location, at a rate of approximately 10 stimuli per second. Participants are asked to identify one or more of these stimuli, termed targets. Stimuli between the targets are masks or distractors. Single targets, even though presented for only 100ms can be reported accurately. However, research has shown that when searching for two masked targets (T1 and T2), there is a severe impairment for detection of the second target, when the two targets are presented within less than 500ms. This finding is known as the AB. Participants however, can accurately report T2 when told to ignore T1, or the targets are separated by more than 500ms. (Shapiro Arnell & Raymond, 1997a; Raymond, Shapiro & Arnell,1992)
There are a number of cognitive explanations for the AB, Raymond et al. (1992) put forward the idea that inability to detect T2 is down to T1 using up attention resources meaning there are insufficient resources left to process T2. But following their further 1997 research they refined this idea suggesting, while T2 is not processed to a level where it can be reported, they were incorrect to suggest that no aspects of T2 are processed. There a number of studies which support this claim. Luck, Vogel & Shapiro (1996) found that if T2 was a word presented during the time period where the AB occurs it produced a negative event-related potential known as the N400, which is a marker that the brain is processing meaning. In another study Shapiro, Ward & Sorensen (1997b) used a three target AB paradigm where T2 was presented during the AB period. Results highlighted that although participants were unable to report T2, it did act as a prime for T3 and ability to detect T3 was better when T2 was semantically related. Both these studies lead to the conclusion that T2 undergoes some degree of higher-level processing, including meaning but does not enter our consciousness.
Based on the idea of T2 undergoing limited processing a number of new theories to explain AB have emerged. Shapiro et al. (1997a) suggested an interference theory, whereby T1, T2 and the masks which appear simultaneously after them are all encoded in a short-term storage buffer where they will interfere and compete with each other. Differential weighting of T1 over T2 due to T1 being the first task means there is accurate report of T1 at the expense of T2. The interference theory predicts the number of items competing in the buffer will affect the magnitude of the AB. Evidence from Issak, Shapiro & Martin (1999) supports this and found AB increases with number of competitors which share conceptual features.
An alternative theory is proposed by Chun & Potter (1995) who suggest a 2-stage model. In stage one stimuli are processed to a preliminary stage where features and meaning are registered but it cannot be reported. In stage 2 the stimuli are consolidated so they can be reported. They explain AB due to the fact that T2 cannot enter stage 2 as it is occupied by T1.
The aim of this study was to try and replicate the findings found in the previous study on AB conducted by Raymond et al. (1992). It was hypothesised that detection of T2 will be less accurate in dual-detection task compared to a single-detection task. As well as that T2 detection will be more accurate when there is a longer delay between T1 and T2 compared a short delay.
Method
Participants:
An opportunistic sample of 283 healthy undergraduate students at the University of Birmingham.
Design:
There were 2 independent variables (IV) in this experiment. The first was the type of task; this was either a single-detection or dual-detection task. The dual-detection condition required subjects to look out for red letters (T1) and report their identity at the end of the stream as well as to look for whether the letter X (T2) appeared in the stream. In the single-detection task subjects had to ignore red letters and report if the letter X (T2) was shown in the stream. The second IV was the position of T2 in the stream. T2 either occurred 700ms after T1, this was the long delay condition (Figure 1) or 300ms after T1, the short delay condition. (Figure 2)
The dependent variable (DV) for this experiment was T2 detection performance, and the percentage of participants who correctly detected the letter X in each of the different conditions was recorded.
A within-subject design was used for the study and all participants completed trials in each condition; dual-detection task with short delay, dual-detection task with long delay, single-detection task with short delay and single-detection task with long delay.
Figure 1. Long delay condition where T2 is presented 700ms after T1
Figure 2. Short delay condition where T2 is presented 300ms after T1
Materials:
A Microsoft PowerPoint slideshow was used to show a sequence of 12 random letters. The letters were presented at the rate of 7Hz. In the stream one letter was red, this was the first target (T1). The letter X followed T1 with a probability of 50%, this was target 2 (T2).
Procedure: This experiment was comprised of 8 trials. In each trial the instructions for the task were displayed on a computer screen, participants then clicked the mouse to begin the study. An RSVP methodology was used in each trial and instructions were followed by the stream of 12 letters. The first 4 trials required participants to complete the dual-detection task, in the remaining trials participants completed the single-detection task. In half the trials T2 occurred 700ms after T1, and in the other half T2 occurred 300ms after.
Participants results were collected and collated at the end of the study in order to calculate the percentage of correct T2 detection.
Results
Our results demonstrated that T2 detection was higher when participants completed the single-detection task compared to the dual-detection task. T2 detection performance was also better in the long delay condition, when T2 was presented 700ms after T1, in both the single-detection and dual-detection task. (Figure 3)
Figure 3 also shows the greatest percentage of participants, 76.8%, correctly detected T2 in the condition comprised of the single-detection task and long delay. While, the lowest percentage of participants correctly detecting T2, 57.3%, occurred during the dual-detection task with short delay.
Figure 3. Percentage of participants who correctly detected letter X (T2) in the different conditions.
Discussion
Our results support the hypothesis proposed and as predicted T2 detection was less accurate in the dual-detection task and when there was a shorter delay of 300ms between T1 and T2. This could suggest that AB is affected by task difficulty, and only occurs in a specific period of time.
Our findings are also in line with existing literature. For instance, Raymond et al. (1992) also reported that participants had no difficulty only detecting one letter, the T2, and in the single-detection task 85% of participants were able to correctly identify T2 in They also found that only 60% of participants could detect T2 in a dual-detection task between 180-450ms after T1 was presented. This is similar to our finding that T2 detection was worse when there was a dual-detection task and T2 was presented 300ms after T1. This consistency with results from previous studies indicates a high level of external reliability and supports the work by Raymond et al. (1992).
The findings in this study however, may be limited as they cannot offer conclusive evidence to support some of the existing theories explaining AB. For example, each trial in this study used RSVP with a stream of 12 letters. But to prove the interference theory proposed by Shaprio et al. (1997a) trials with a varying number of distractors may be needed to demonstrate that AB in due to competition between the targets and masks in the buffer. Similarly, to support the 2-stage model by Chun & Potter (1995) a modification to the study may be needed so it is possible to determine how much processing T2 undergoes. This could involve the use of EEG or priming techniques as seen in previous studies (Luck et al. 1996 ; Shapiro et al., 1997b) or alternatively the use of other neuroscience techniques such as fMRI to see if brain areas linked to processing meaning are active during the period when T2 is presented.
Furthermore, there are a number of questions left unanswered by both the present and existing studies on AB. For example, most existing theories suggest that the masks or distractors play a crucial role in the AB, but their exact role is not specified. A recent study by Nieuwenstein, Potter & Theeuwes (2010) suggests that the AB can occur without distractors. They conducted a RSVP experiment in which T1 and T2 were separated by a blank interval and the AB was still found to occur. These results could be interpreted to suggest that it our inability to rapidly reengage attention, shortly after disengaging attention following identification of T1 which leads to the AB, rather than the effect of distractors or depletion of processing resources. This has implications for existing theories and thus a new model for AB may be needed.
Another question which needs answering is why processing of T2 is relatively unaffected when it is presented immediately after T1. The work by Nieuwenstein et al. (2010) could also offer an explanation for this. Their work suggests target obdurate stimuli will elicit the deployment of attentional processing resources. Therefore, if T2 immediately follows T1 attentional engagement will be sustained, allowing T2 to be identified. If a distractor, or blank interval was to follow T1 then there will be a discontinuation of the input which kept attention engaged, and so attention will disengage prior to the presentation of T2 meaning it cannot be reported.
Overall, it is clear that while results found in this study are in line with others there is a lack of certainty in what causes these results, demonstrated by the number of contradicting and conflicting theories presented to in order explain AB. In the future further research into the AB could have a focus on, the role of distractors and timing of the AB in order to create a single unified model.
References
Chun, M. M., & Potter, M. C. (1995) A two-stage model for multiple target detection in rapid serial visual presentation. Journal of Experimental Psychology: Human Perception and Performance, 21, 109-127.
Isaak, M. I., Shapiro, K. L., & Martin, J. (1999) The attentional blink reflects retrieval competition among multiple rapid serial visual presentation items: Tests of an interference model. Journal of Experimental Psychology: Human Perception and Performance, 25, 1774-1792.
Luck, S. J., Vogel, E. K. and Shapiro, K. L. (1996) Word meanings can be accessed but not reported during the attentional blink. Nature, 382, 616-618.
Nieuwenstein, M. R., Potter, M. C., & Theeuwes, J. (2010). Unmasking the attentional blink. Journal of Experimental Psychology: Human Perception and Performance, 35, 159- 169.
Raymond, J. E., Shapiro, K. L., & Arnell, K. M. (1992) Temporary suppression of visual processing in an RSVP task: An attentional blink. Journal of Experimental Psychology: Human Perception and Performance, 18, 849-860.
Shapiro, K. L., Arnell, K. M., Raymond, J. E. (1997) The attentional blink. Trends in Cognitive Sciences, 1, 291-296.
Shapiro, K. L., Driver, J., Ward, R., & Sorensen, R. E. (1997) Priming from the attentional blink: A failure to extract visual tokens but not visual types. Psychological Science, 8, 95-100.