We are constantly bombarded with a lot more information than we can process at once. We use our attentional abilities to filter out unwanted information and to focus on those things that appear particularly relevant—for example, right now I am trying to ignore all the buzz around me in a Star Alliance lounge at Heathrow while focusing on the screen in front of me.
Sometimes the filtering process is imperfect and we cannot help but be drawn to stimuli that we would rather ignore. For example, in an earlier post we noted how difficult it can be to name a picture when it is adorned with an interfering word, say “bird” written on a rabbit.
The temptation to say “bird” is difficult to resist, even though we know that our task is to name the picture instead. Clearly, we cannot ignore printed information, whether we like it or not, as we have noted here previously on several occasions.
But this is not the only way in which our attention can fail: sometimes attention fails not because unwanted stimuli intrude, but because the brain requires some time to process stimuli to a durable state before allowing subsequent stimuli in. One of the most striking demonstrations of this type of failure is the so-called attentional blink (AB) paradigm, in which participants must detect two stimuli in a stream of rapidly presented non-relevant stimuli.
The figure below illustrates the basic procedure. Participants are instructed to detect the only white letter in the sequence (“V” below, which we generically call T1 from here on), and then whether another specific letter (“X” below, generically called T2) appears in the stream of stimuli following T1. (Alternatively, in some experiments both targets might be of a distinct color and participants report all targets in that color.) At the end of the sequence, participants must report both T1 and T2.
The pervasive finding in this paradigm is that given correct detection and report of T1, people are far less likely to report T2 accurately when it is close to T1 but not immediately adjacent to it—that is, when some 200-600 milliseconds elapse between T1 and T2. This temporal “blink” must reflect an attentional phenomenon because it only occurs if people must report T1; detection of T2 is not impaired by T1 when only T2 needs to be reported. The next figure shows canonical data:
The blue line corresponds to the condition where both targets must be reported, whereas the red line is for trials on which only T2 must be reported. The effect is very powerful and you can observe it yourself by participating in the following demonstration online:
Although the AB phenomenon has been known for more than 25 years, it is still not clear what the underlying brain mechanisms are. A recent article in the Psychonomic Bulletin & Review explored some of those possible underlying mechanisms. Researchers Kimron Shapiro, Simon Hanslmayr, James Enns, and Alejandro Lleras reasoned that the T2 deficit in the attentional-blink paradigm might be linked to the brain oscillations that are generated by the temporal presentation rate of the stream in which the two targets are embedded.
Most previous study on the AB have presented the visual stream at a rate of 10 stimuli per second; that is, at a frequency of 10 Hz. It turns out that in the human brain, 10 Hz corresponds to the frequency of alpha oscillations. Alpha oscillations were one of the earliest synchronous patterns to be detected in the brain using electroencephalography (EEG). By the time we reach age 3, the alpha oscillation at 10 Hz is typically firmly established.
Numerous studies have linked alpha oscillations to perceptual performance, with the general finding that high alpha activity predicts a low level of perceptual performance. Shapiro and colleagues therefore hypothesized that alpha oscillations might similarly be linked to the magnitude of the attentional blink.
To test this possible link, Shapiro and colleagues kept the temporal lag between T1 and T2 constant (and equal to the lag that has produced the AB in countless previous studies). Instead, the researchers varied the frequency of the visual stream in which the targets were embedded. Four different frequencies of presentation were used (Theta: 6.26 Hz, Alpha: 10.3Hz, Beta: 16.0 Hz, and Gamma: 36.0 Hz). The figure below shows the various experimental sequences:
As shown by the vertical red lines, the lag between T1 and T2 was the same irrespective of stream frequency, and the lag between T2 and the next distractor stimulus was kept constant to equalize potential perceptual interference effects.
The results from one of the experiments reported by Shapiro and colleagues are shown below (they conducted two studies, both of which yielded nearly identical results):
The figure shows the magnitude of the AB, computed as the difference in accuracy between reporting T1 and T2 at a lag of 300 ms (the lag that most reliably yields an AB). It is clear from the figure that the AB is greatest when the stimuli are presented in the alpha and beta range of frequencies (10 Hz and 16 Hz, respectively).
The results indicate that the attentional blink is not just a function of the temporal lag between T1 and T2, as most existing theories have postulated or taken for granted. Instead, the frequency of surrounding events also determines the magnitude of the AB.
Shapiro and colleagues provide a provocative interpretation of those data: the researchers suggest that the “… effects point to underlying neural circuitry that opens and closes access to conscious perception, potentially both endogenously and exogenously.”
This conclusion is supported by previous work which has linked alpha (and beta) oscillations with inhibition of neural assemblies, and the fact that visual stimuli are more often missed when alpha (and beta) activity is particularly high.
These results point to the possibility that brain oscillations play an important role in cognitive information processing.
Article featured in this blog post:
Shapiro, K. L., Hanslmayr, S., Enns, J. T., & Lleras, A. (2017). Alpha, beta: The rhythm of the attentional blink. Psychonomic Bulletin & Review. DOI: 10.3758/s13423-017-1257-0.
