Increased attention towards threatening stimuli in both the external and internal environments is thought to be a factor in the causation and maintenance of pathological anxiety. Attentional biases for threatening information have been demonstrated in anxiety disorders, however the cortical mechanisms involved remain unclear. In this investigation, an Emotional Stroop task consisting of neutral, positive, depression-related and anxietyrelated words, was used to investigate attentional biases in 14 Panic Disorder patients and 32 psychiatrically healthy controls. The standard colour-word Stroop was also performed to determine whether any general cognitive deficits exist in Panic Disorder. Steady-state probe topography (SSPT), a brain electrical activity imaging methodology, was used to investigate participants' brain activity during performance of the tasks. It was hypothesised that Panic Disorder is associated with specific biases for disorder-specific information and thus patients would exhibit increased interference for anxiety-related words only, compared to neutral words. Mean reaction times for the Standard Stroop was similar for the two groups. For the Emotional Stroop task, neither group showed an interference effect for any emotional category. However, Panic Disorder patients performed the Emotional Stroop significantly more slowly than the Controls. The SSPT data suggest that the Standard and Emotional Stroop tasks are associated with different patterns of brain activity in the Control and Panic Disorder groups despite the similarities in the reaction time data. Specifically, the Standard Stroop was marked by strong temporo-parietal excitation in the Panic Disorder group only. In addition, anterior SSVEP patterns further differentiated between the Control and Panic Disorder groups. The most striking finding for the Emotional Stroop was strong sustained bilateral temporo-parieto-occipital excitation in the Panic Disorder group. In addition, a subgroup of the Controls exhibited increased interference for anxiety-related words and therefore the brain activity for this group and the remainder of Controls who did not show interference was analysed separately. It was found that the presence of interference for anxiety-related words was associated with right prefrontal inhibition prior to response. Other time-varying changes in the SSVEP further distinguished between the subgroup of Controls who showed an interference effect and those who did not.