Flipbook

CHAPTER 8 202 because this could prevent relapse. For example, a patient might practice going to the supermarket with and without taking his anxiolytic medication, bringing his mobile phone or partner, and quickly sitting down when he feels dizzy. When these safety behaviors become part of the extinction context, their availability after therapy may not present a context shift after treatment and trigger a return of fear. Treanor and Barry (2017) proposed to incorporate safety behaviors in the later stages of exposure, because the discrepancy between the expected and actual outcome of exposure is expected to be smaller in later stages than in the early stages of exposure therapy. Therefore, the negative effects of a conditional inhibitor on inhibitory learning may be reduced. Future research that investigates the effect of incorporating safety behavior in the final stages of exposure therapy on the long-term effects on treatment outcomes is needed. FUTURE RESEARCH The studies presented in the current dissertation include healthy and subclinical samples, and are performed in a laboratory setting. As is more often the case with experimental psychopathology research, this may raise questions about whether the findings generalize to real-world settings (van den Hout, Engelhard, & McNally, 2017). We chose this approach, however, because we aimed to investigate the learning principles of safety behavior in the persistence of irrational fears. The fear conditioning model has proven to be a robust and valid preclinical laboratory model for the exposure and the return of fear (Milad & Quirk, 2010; Vervliet et al., 2013). Additionally, results obtained in subclinical samples are often informative for clinical populations (van den Hout et al., 2017). For example, studies with analogue samples are useful for understanding obsessive-compulsive (OC) related phenomena, because OC symptoms are prevalent in nonclinical populations, and the maintenance factors are similar to those in clinical populations (Abramowitz et al., 2014). Nevertheless,

RkJQdWJsaXNoZXIy MTk4NDMw