Cognitive Bias Modification: An Intervention Approach Worth Attending To

People with emotional disorders display biased patterns of cognition, operating to favor the processing of emotionally negative information (1). A particularly robust finding is that anxiety disorders are characterized by an attentional bias toward threatening information (2). Cognitive accounts have implicated this attentional bias in the genesis and maintenance of anxiety pathology (3). However, it is only recently that clinical researchers have developed training procedures capable of directly modifying cognitive biases and have sought to evaluate the capacity of these procedures to therapeutically ameliorate emotional dysfunction. Encouraging early findings have led to rapid expansion of this cognitive bias modification literature across the past 3 years (4).

Much of this work has made use of a computerized attentional training task, designed to reduce selective attention to threatening stimuli by requiring participants to process visual probes consistently presented in screen locations distal to threatening words or images. Initial proof-of-concept studies confirmed that single-session delivery of this computer training task successfully modifies selective attentional response to threat and also alters the intensity of anxiety reactions to laboratory stressors, both in healthy adults (5) and in children (6). Several randomized controlled trials employing placebo control conditions since then have shown that more extended exposure to this attentional bias modification procedure, configured to reduce attention to threat, can alleviate clinical symptoms in adults diagnosed with generalized anxiety disorder (7) or social anxiety disorder (8). Recently, a small case series (9) showed that clinically anxious children given extended exposure to this attentional training procedure evidenced a reduction in anxiety symptoms. However, until now, no randomized controlled trial has investigated whether the attention bias modification approach can yield therapeutic benefits for children with clinical anxiety.

In this issue of the Journal, Eldar et al. (10) report the first randomized controlled trial to evaluate whether computerized attention bias modification can ameliorate the symptoms of pediatric clinical anxiety. Forty children with a current anxiety disorder (separation anxiety disorder, generalized anxiety disorder, specific phobia, or social phobia) were randomly assigned to one of three conditions. In the attention bias modification condition, children completed a version of the computerized probe task designed to train attentional bias away from threatening stimuli. On each trial, a fixation cross was followed by two photographs (an angry face and a neutral face) exposed for 500 msec. Immediately the photographs disappeared; a small probe was presented in the screen location where one of the two images had just been shown. This probe was a pair of dots aligned either horizontally or vertically, and the child's job was simply to press a button to indicate which type of probe appeared. In the condition designed to induce attentional bias away from the more threatening stimulus, probes always appeared distal to the location where the angry face had just been shown and proximal to the location where the neutral face had been shown. Numerous such trials (1,920 in total) were completed over four weekly sessions.

This attention bias modification condition was compared with two well-matched control conditions. In one control condition, children were exposed to the same angry-neutral face pairs and performed the same probe discrimination task, but the probes appeared with equal frequency in the location where either face had just been shown. In the other control condition, only neutral-neutral face pairs were shown in the probe task. Thus, neither control condition contained a training contingency to modify attentional response to threat, although they differed with respect to whether or not they exposed children to angry faces.

Eldar et al. first examined whether the attention bias modification condition successfully induced attentional avoidance of the threatening image. Inclusion criteria stipulated that all children should display a preexisting attentional bias toward the angry faces, and this attentional bias was found to be equivalent across groups at baseline. Importantly, attentional bias to the angry faces declined significantly across the 4-week period among participants in the attention bias modification condition but not among those in either control condition. The researchers then examined anxiety symptoms. Critically, there was a significant reduction in child and parent reports of anxiety symptoms as well as in clinician ratings of anxiety severity for participants who completed the attention bias modification condition but not for those in either control condition. At the end of the trial, 33% of the attention bias modification group no longer met diagnostic criteria for any anxiety disorder, compared with 13% and 0% for the two comparison groups. Hence, the study confirms that attentional bias to threat can be reduced in clinically anxious children through computerized training and that such bias modification serves to reduce anxiety symptoms and their severity.

This is the first randomized controlled trial to evaluate whether attention bias modification can alleviate clinical anxiety in children, and the study has some notable strengths. The comparison groups were well chosen, permitting the conclusion that it is change in attentional bias—and not merely performing a probe discrimination task or the exposure to angry faces—that results in the observed symptom improvement. These positive findings are particularly important given that current treatments for pediatric anxiety are insufficient and in view of the fact that child anxiety is a precursor to later psychiatric problems. Attention bias modification adds a new clinical weapon to our treatment armory, and its application in childhood may help protect against the development of subsequent psychopathology, thereby acting as a “cognitive vaccine” (11). The compatibility of attention bias modification with computer delivery could make it readily accessible to young people outside the clinic setting, perhaps using mobile technology. Furthermore, the study's demonstration of significant improvement in pediatric anxiety symptoms after only four weekly sessions indicates efficacy at low intensity.

Of course, the study has some limitations that future research could usefully address. The sample size is small, and replication with larger samples would provide reassurance that the observed effects are reliable. No follow-up data are reported, leaving uncertain the duration of the observed therapeutic benefits of attention bias modification in anxious children. The study does not identify the neurocognitive mechanisms governing attention bias modification-induced attentional change. A fuller understanding of these mechanisms could contribute to the development of more effective bias modification procedures. Perhaps the most significant limitation of the study is that it does not contrast the therapeutic effect of attention bias modification with that of more established intervention approaches, such as cognitive-behavioral therapy or pharmacotherapy. Such comparisons will be necessary in order to determine the true value of this intervention in the clinical setting. Future research should also evaluate the potential benefits of delivering attention bias modification in conjunction with conventional interventions, given the evidence that multimodal approaches are more effective than monotherapies in the treatment of pediatric anxiety (12).

Nevertheless, Eldar et al. have made a timely, distinctive, and significant contribution to the burgeoning body of evidence that attention bias modification can contribute to the alleviation of clinical anxiety. They have convincingly demonstrated that attentional bias to threat can be attenuated in clinically anxious children using a computerized training approach and have shown that this leads to a restriction in the breadth of anxiety symptoms experienced, a decline in their severity, and reduced rates of clinical diagnosis. Hence, this study gives good grounds for confidence that attention bias modification is likely to prove to be of therapeutic value in the treatment of pediatric anxiety.