(PS2-C51) Examining the Roles of Reward Responsiveness and Rumination in the Prediction of Non-suicidal Self-injury During Late Adolescence

Background: Non-suicidal self-injury (NSSI), or the deliberate harm to one’s bodily tissue without suicidal intent, is conceptualized as a highly prevalent, emotion regulation strategy among adolescents (Nock, 2010). Although past research indicates that individuals with elevated levels of reward responsiveness (RR) are prone to the reinforcing properties of NSSI (Burke et al., 2015), contrary evidence suggests that higher RR may predict fewer NSSI acts (Jenkins et al., 2013). These contradictory findings may be due to limitations in subjectively measuring RR; thus, investigations using objective measures are needed. A separate literature suggests that rumination exacerbates NSSI frequency (Selby et al., 2008). That is, rumination on negative affect and dampening of positive affect may amplify one’s negative mood state, strengthening the desire to engage in NSSI. Despite evidence that both RR and rumination contribute to NSSI occurrence during adolescence, limited work has explored the interplay between these factors simultaneously. We sought to address these gaps by examining the interaction of RR and rumination in the prediction of NSSI frequency.

Method: Adolescents and young adults (n = 219; age 14-21) with moderate to high levels of reward sensitivity completed measures of RR, rumination, and NSSI frequency. At baseline, participants completed multimodal measures of RR: the Behavioral Activation Scale (BAS; Carver & White, 1994) and the Card Arranging Reward Responsivity Objective Test (CARROT; Al-Adawi et al., 1998). Additionally, participants completed the Ruminative Responses Scale – Brooding subscale (RRS-B; Treynor et al., 2003) and the Responses to Positive Affect Scale – Dampening subscale (RPAS-D; Feldman et al., 2008). At a follow-up visit (M = 9.7 months), participants completed the Form and Function Self-Injury Scale (Jenkins & Schmitz, 2012); a self-report measure of lifetime and past year NSSI frequency. To reduce frequency count variability, NSSI frequency was collapsed into five categories (0, 1, 2-5, 6-20, and 20+; Whitlock et al., 2013).

Results: Aside from RRS-B and RPAS-D, which shared small correlations with NSSI frequency (r = .13-.16, all p < .049), no measures of RR (all r < .03) were reliably associated with NSSI frequency over the past year, all p > .687. When entered into the same model, RRS-B (ß = .18, p = .010), but not BAS (ß = -.01, p = .909), significantly predicted past year NSSI frequency at follow-up and drove the model’s significance, F(3, 210) = 2.681, p = .048. However, their interaction did not predict NSSI frequency over the past year (R² change = .01, p = .315). A separate set of analyses found that RPAS-D better accounted for the variance in past year NSSI frequency (ß = .14, p = .040) than CARROT scores (ß = .02, p = .752).



Conclusion: Ruminative brooding and dampening, rather than RR or their interaction, prospectively predicted NSSI frequency among late adolescents. Self-report and objective measures of RR did not better explain variations in past year NSSI occurrence. Our results suggest that even among individuals with elevated levels of RR, rumination may serve as a better predictor of NSSI frequency six months to a year later. Implications for theory, research, and practice will be discussed.