Aims Individuals with Generalised Anxiety Disorder have an attentional bias towards threatening information. It is not known whether this results from facilitated engagement (faster orientation) or delayed disengagement (shifting attention away) from threat. Recent research has developed a new methodology designed to modify attentional disengagement from threat. Using this paradigm, the present study assessed the causal role of attentional disengagement from threat and its impact on worry.

Methods Twenty-four university students scoring below 56 on the Penn-State-Worry-Questionnaire were randomly assigned to either threat disengagement training, or non-threat disengagement training. Training was assessed using threat and non-threat test-trials. All participants then completed a novel worry task, assessing tendency, ability and persistency of worry. The hypothesis was that training to disengage from threat rather than non-threat stimuli would affect tendency, ability or persistence of worry.

Results Accuracy and test-trial reaction-time data indicated disengagement training was successful; compared to the non-threat disengagement group, the threat disengagement group had faster reaction-times for non-threat valence test-trials, experienced marginally non-significantly more negative intrusions during active worry, and found it significantly more difficult to worry, when required to engage solely with worry without interruption in the worry task.

Conclusions It is possible to manipulate attentional bias to disengage from threat information, leading to fewer negative thought intrusions during active worry and increased difficulty in engaging solely with worry, thus suggesting that impaired disengagement has a causal role in the ability to worry.