Several experiments have shown that we can reduce the frequency of analogue flashbacks with competing tasks presented during a trauma film (i.e. peri-traumatically). A "distraction" hypothesis suggests that any competing task may reduce flashbacks due to distraction and/or a load on executive control. Alternatively, a "modality" hypothesis based on clinical models of PTSD suggests that certain tasks will not protect against intrusions (Experiment 1) and could actually increase them (Experiment 2). Experiment 1 contrasted two concurrent tasks, Verbal Interference (counting backwards in threes) and Visuospatial tapping, against a no-task Control condition during trauma film viewing. The Visuospatial group had significantly fewer intrusions of the film over 1-week than the Control group. Contrary to a distraction account, the Verbal Interference group did not show this effect. Using a larger sample, Experiment 2 showed that the Verbal Interference group (counting backwards in sevens) had more intrusions (and inferior voluntary memory) than no-task Controls. We propose that this is in line with a modality hypothesis concerning trauma flashbacks. Disrupting verbal/conceptual processing during trauma could be harmful for later flashbacks.