Testing associations between assessments of cognitive flexibility and eating disorder symptoms in adolescent bulimia nervosa

Objective: Cognitive rigidity, or difficulty adapting to changing demands, is commonly observed in anorexia nervosa. Less is known, however, about cognitive flexibility (CF) in bulimia nervosa (BN) and, particularly, adolescence. Clarifying this relation and best assessment practices may guide informed clinical decision-making. The current study compared how two measures of CF (i.e., Wisconsin Card Sort Task [WCST] and Trail Making Task [TMT]) relate to BN symptoms among adolescents. Methods: Data from a subsample (n = 78) of adolescents with BN were analyzed. Linear and hurdle regressions were used to compare the effects of WCST perseverative errors and TMT performance on Eating Disorder Examination Global Scores, objective binge episodes, and self-induced vomiting episodes (SVEs) at baseline and end-of-treatment (EOT). Results: Neither CF measure associated with baseline BN symptoms. TMT performance positively associated with the likelihood of engaging in SVEs at EOT (beta = 0.47, p = 0.01, 95% confidence interval [CI] = [0.11-0.84]) and, among adolescents who endorsed >= 1 SVE at EOT, WCST perseverative errors (beta = 0.05, p = 0.005, 95% CI = [0.01-0.08]) positively associated with SVE frequency at EOT. Discussion: The overall lack of associations between CF and outcomes suggests that cognitive rigidity may not be as relevant to the clinical profile of adolescent BN as for anorexia nervosa. In the few significant associations that emerged, the WCST and TMT uniquely predicted the severity of vomiting at EOT in this sample. Given the lack of CF deficits, future work should aim to test the role of other executive functions (e.g., impulsivity), in addition to CF, to determine which deficits are present in adolescent BN and may predict outcomes. Public Significance: Patients with eating disorders often have difficulties thinking flexibly, which may interfere with their recovery. We tested two ways of measuring flexible thinking in adolescents with BN. Overall, flexible thinking was not associated with symptom-level outcomes. However, less flexible thinking at the start of treatment predicted self-induced vomiting at EOT. If findings are replicable, then assessing and addressing flexible thinking could improve outcomes for adolescents with BN.