Science Watch

In the days and weeks after the Sept. 11 terrorist attacks on the United States, the world was continually confronted by the horrific images: a hijacked jet plowing into the World Trade Center's south tower minutes after another had crashed into the north tower; the burning towers collapsing to rubble; the enormous, blackened gash in the Pentagon; the slice of smoldering Pennsylvania earth where a fourth hijacked plane crashed after passengers apparently thwarted its killing mission.

Surely, these tragic sights are inscribed indelibly in the memories of millions of people. According to a recent study of anxiety and visual attention, for people who tend to be anxious even in the best of times, such fearsome images--as well as more mundanely threatening stimuli in the environment--may be especially inescapable, not only reflecting heightened anxiety but also, perhaps, perpetuating it. The new study, published in the December issue of the Journal of Experimental Psychology: General (JEP: General) (Vol. 130, No. 4), sheds new light on long-observed attention biases in anxiety, suggesting that attention is not only more quickly drawn to threatening stimuli for people who are anxious than for people who are not, but also that anxious people have trouble disengaging their attention from such stimuli.

The research was conducted with participants who were not experiencing clinical levels of anxiety. If it is replicated in clinical populations, suggests the study's lead author, Elaine Fox, PhD, of the University of Essex, it may help explain why people with anxiety disorders often have difficulty banishing worrisome thoughts from their minds.

The findings are "an important step forward theoretically, in understanding the relationship between cognition and emotion," comments clinical psychologist Richard J. McNally, PhD, who studies anxiety disorders and cognition at Harvard University. "Fox has isolated pretty convincingly where the problem with attention seems to lie."

Measuring disengagement

For the past 15 years, investigations into how threatening stimuli influence visual attention have relied largely on two methods for assessing attentional bias, known as the Stroop and dot-probe procedures. In both methods, threat-related or negative stimuli and neutral visual stimuli are placed in competition with each other, and participants' task is to attend to one and ignore the other.

Studies using these methods have consistently shown that threat-related stimuli interfere with people's processing of neutral stimuli--especially for people who are highly anxious. For example, investigations using the Stroop procedure have shown that anxious participants are slower to name the color in which threat-relevant words are printed than they are for neutral words, suggesting that the emotion-laden words' meaning soaks up attention resources that could otherwise be used for processing other information.

"These experiments have been interpreted as evidence that anxious people were shifting their attention toward the location of threat," observes Fox. But she and her colleagues suspected that an alternative--or at least an additional--mechanism might be at work. "It may not just be that anxious people are shifting their attention toward threat faster," she explains, "but that when they notice threat, they can't disengage from it as quickly."

To test that notion, the researchers adapted a procedure developed by University of Washington psychologist Michael I. Posner, PhD, and colleagues, known as "exogenous cueing." The procedure provides a test of visual attention that--unlike other measures--distinguishes between two components of attention: attentional shift and attentional disengagement.

In a series of experiments, each of which included several hundred trials, a small circle was briefly presented on either the left or right side of a computer screen. Participants' task was to quickly indicate, by pressing one of two keys, on which side of the screen the circle had appeared. In each trial, presentation of the circle was preceded by a "cue" stimulus, which sometimes appeared on the same side of the screen that the circle would appear and sometimes appeared on the opposite side. In some experiments, the cues were threatening, positive or neutral words, and in some experiments, the cues were angry, happy or neutral faces.

The researchers were interested in participants' response speeds on so-called invalid trials, in which the target circle appeared on the opposite side of the screen from the cue. This condition would require participants to withdraw their attention from the cue and transfer it to the opposite side of the screen.

The team reasoned that if anxiety is associated with an impaired ability to disengage from threatening stimuli, then anxious participants would be slower to respond on invalid trials when they had been cued with threatening stimuli--in some experiments, fear-provoking words, and in others, angry faces--than when they had been cued with neutral or positive stimuli. In contrast, there should be no such attentional bias for nonanxious participants.

The results of the experiments supported those hypotheses, leading Fox and her colleagues to conclude that anxiety--at least at subclinical levels--is associated with a decreased ability to disengage from threatening stimuli.

The new research "leads to a potential re-evaluation of the literature, which has been focused on the idea that anxious people are shifting toward threat," argues Fox.

"It's very clever work," comments Arne Öhman, PhD, a professor of psychology at the Karolinska Institute in Stockholm, Sweden, who also studies attention and emotion. "Fox and her colleagues present a compelling argument for the idea that an entirely different mechanism from what we have believed may be at work in anxiety. They provide a valuable contribution in pointing to these alternative explanations."

But, Öhman notes, the new findings don't rule out that threatening stimuli also capture attention more quickly than do neutral stimuli.

In fact, in a study published in the September issue of JEP: General (Vol. 130, No. 3), he and colleagues reported that participants were faster to detect pictures of fear-relevant targets--snakes and spiders--in a field of fear-irrelevant pictures than they were to detect fear-irrelevant targets. Further, they found, participants who reported being afraid of snakes or spiders were especially prone to this attentional bias. It will take further work, Öhman and Fox agree, to untangle the circumstances under which attention shifting and attentional disengagement biases occur.


The discovery that anxiety is linked to slower disengagement from threatening stimuli raises as many questions as answers, observe the study's authors. First, the finding poses an evolutionary puzzle: Why would such a mechanism be adaptive?

One possibility, Fox suggests, is that the disengagement delay in response to threatening stimuli allows animals--including humans--to conduct more detailed cognitive processing of potential threats in their environment. Alternatively, she and her colleagues speculate, it may be that the disengagement bias is a vestige of a broader "freezing" mechanism, which might have evolved to protect animals from attack.

Another important question, Fox says, is whether the disengagement bias that she and her colleagues observed in non-clinically anxious participants also occurs for people with more intense anxiety--and whether the millisecond-level attentional bias that they have observed is associated with the chronic worrying and rumination that people with anxiety disorders often experience.

"If this low-level mechanism is replicated at a higher level, then it could be a sustaining mechanism whereby people who are anxious tend to dwell on threat stimuli and remember them more," speculates co-author Riccardo Russo, PhD, also of the University of Essex. The team is now examining these questions in studies of people diagnosed with generalized anxiety disorder. Ultimately, Fox hopes, the research may lead to clues for curbing anxiety disorders.

"If we can actually train people to quickly shift their attention away from threatening stimuli, potentially we can reduce the attention bias and ultimately reduce anxiety," she says.

This article is part of the Monitor's "Science Watch" series, which reports news from APA's journals.