Martijn van Teffelen

Imagery-enhanced cognitive restructuring: Efficacy 89 4 Second, we analyzed the fixed effects of intervention on hostile beliefs, aggressive tendencies, and state anger after the emotional stressor at one-week follow-up (T2b). When M4/M5 (pre- emotional stressor) was selected as reference category, results demonstrated that hostile belief ratings ( b = 0.45, B = 6.35, SE = 1.56, t = 4.08, p < .001), aggressive tendencies ( b = 0.25, B = 2.22, SE = .96, t = 2.31, p = .024) and anger VAS ( b = 1.42, B = 41.39, SE = 3.20, t = 12.93, p <.001) significantly increased from pre- (M4/M5) to post-emotional stressor (M6). To test if the emotional stressor fully counteracted the obtained effects of I-CR and CR on hostile beliefs and aggressive tendencies in session 1, we ran the same model with baseline values of hostile beliefs and aggressive tendencies (M1) as reference category (see Table 3). Hostile belief ( b = -0.34, t = -2.45, p = .017) and aggressive tendencies (b = -0.28, t = -2.03 p = .046) remained significantly reduced compared to AC at one-week follow-up (M5) for participants in the I-CR condition. Also, compared to baseline (M1), hostile belief scores in the CR condition were significantly lower than scores in the AC condition ( b = -.32, t = -2.31, p = .024), but not in terms of aggressive tendency scores. In sum, observations on the sustainability of the intervention efficacy revealed that the efficacy of I-CR on hostile belief and aggressive tendencies was sustained. Moreover, the efficacy of CR was sustained in reducing hostile beliefs, but not in reducing aggressive tendencies. DISCUSSION The present research investigated if the efficacy of CR for transdiagnostic hostility would enhance through adding mental imagery. We expected that I-CR would be more sustainably efficacious in reducing the believability of hostile beliefs (H1), aggressive tendencies (H2), state anger (H3) and trait-hostility (H4) compared to traditional CR and an AC condition. The findings revealed that both I-CR and CR were superior to the AC group in changing hostile beliefs, aggressive tendencies, and state anger. We found that one session of I-CR holds the potential of more efficaciously changing hostile beliefs compared to traditional CR. Although actual hostility traits were not reduced more strongly in I-CR and CR than in our control condition, we observed sustained reductions in hostile beliefs and aggressive tendencies. Our findings provide convergent data with previous work that evidenced the potential of mental imagery to increase treatment efficacy for psychopathologies (Arntz & Weertman, 1999; Beck, 1985; Edwards, 1990; Ehlers et al., 2005; McEvoy et al., 2015; Smucker et al., 1995). One explanation for our finding that I-CR was superior to CR in changing hostile beliefs is that I-CR results in more elaborate processing and changing of meaning of participants` ideographic events. This is in line with the ideas of Beck (1985) on the superiority of changing “hot” (i.e., affectively valenced) cognitions when compared to “cold” (i.e., affectively