Abstract

This study was designed to investigate the characteristics of selective attentional bias in anxiety disorder patients. Specifically, we examined the content-specificity hypothesis and the preattentive bias hypothesis utilizing the computer-form modified Stroop task. Panic(n=20), socially anxious(n=20), and normal(n=20) subjects participated in the experiment and four types of words(panic-related, social anxiety-related, positive, and neutral words) were used as experimental stimuli. Half of the words were presented supraliminally and the other half were presented subliminally with backward masking procedure. The 'emotionality' of words was rated at pre- and post-experiment and the equality of emotionality level was confirmed. Results revealed that the selective attentional bias in anxious patients was confined to worry-congruent threatening words, while worry-incongruent threatening words and emotionally positive words did not elicit the attentional bias. These results support the content-specificity hypothesis of cognitive theory. The pattern of selective bias was observed in the subliminal condition as well as supraliminal condition. Although anxious patients were not aware of the meaning of the presented words, they showed selective attentional bias to worry-congruent threatening words. These results suggest that selective attentional bias also operates at the preattentive level. Finally, the results of this study were discussed in the light of clinical implications. Limitations of this study and suggestions for future studies were also discussed.