Selective processing of threat cues in subjects with panic disorder: The effect of awareness and specificity of threat cues

A Stroop color-naming task was used to deliver supraliminal and subliminal stimuli to normal subjects and subjects with panic disorder. In both conditions, subjects color-named panic threat stimuli and social phobia stimuli as well as neutral control words. They also performed the original Stroop task of color-naming color words in both conditions. In supraliminal trials, both groups showed greater Stroop interference to panic threat cues compared to neutral control words, although, unlike much other research in this area, subjects with panic disorder did not show a greater magnitude of Stroop interference to panic threat cues than did normal subjects. Neither group differentially responded to social phobia threat cues compared to their neutral control words. In subliminal trials, there was no evidence of the selective processing of threat cues. No differences were observed in either group for panic threat or social phobia threat cues compared to their neutral control words. A lexical decision task was used both as a manipulation check, and as a means to test whether the valence of a stimulus could be detected when delivered outside of awareness. Though the manipulation proved effective, the valence of cues was not detected in this procedure. The strength of this study lies in its many procedural checks and controls. Methodological improvements include manipulation checks, matching on variables which might affect outcome, checks on the appropriateness of the stimuli, and a carefully controlled stimulus presentation. Critical analysis of other research is outlined. Failure to replicate findings reported by other investigators is discussed in light of cognitive theories of psychopathology. These results, taken together with other findings, suggest that the attentional bias for processing threatening information assumed to be associated with panic disorder needs further exploration.