We report 2 experiments investigating hypervigilance to pain signals in healthy individuals. In order to allow a detailed analysis, we decomposed attention in 3 different processes: 1) initial shifting, 2) engagement, and 3) disengagement. We used a spatial cueing paradigm in which the location of targets is validly or invalidly predicted by spatial cues. in the first experiment, cues were differentially conditioned to create pain signals, allowing the investigation of engagement and disengagement of attention. Because this procedure does not allow an adequate isolation of the shift component, we introduced a new adaptation of the spatial cueing paradigm in the second experiment, in which targets instead of cues were differentially conditioned. We replicated previous findings, showing enhanced engagement to and retarded disengagement from pain signals compared to control signals. In addition, we demonstrated that participants were still hypervigilant to pain signals after extinction. We were not able to demonstrate speeded shifting to pain signals. instead we found a generally faster detection of pain signals compared to control signals. We conclude that hypervilance to pain signals does not emerge as rapid initial shifting to the pain signal but rather as enhanced processing once it is detected. Theoretical and clinical implications of these findings are discussed. Perspective: We investigated the fundamental processes associated with hypervigilance to pain. Our findings suggest that treatment approaches using attention techniques should not focus on preventing patients to shift their attention to the pain but rather on diminishing the threat value and learning to disengage from their pain.