Heightened Depression and Executive Functioning Bidirectionally Impair One Another: Evidence from Within-Person Latent Change and Cross-Lagged Models

Abstract

The impaired disengagement hypothesis (Koster, De Lissnyder, Derakshan, & De Raedt, 2011) posits that executive dysfunction can predict future heightened depression and anxiety due to chronic poor attentional control and repetitive negative thinking across prolonged durations. Simultaneously, scar theories (Ottaviani et al., 2016) assert that increased psychopathology may forecast subsequent executive functioning (EF) deficits because of wear-and-tear of psychophysiological systems over protracted timescales. However, most work on EF-psychopathology relations have been cross-sectional, which precludes causal inferences. Thus, this study aimed to test the within-person relations between EF, depression and anxiety. Older adult participants (n = 856) averaged 81.59 years of age (SD = 7.10, range = 70–110, 58.53% females, 76.87% Whites). Assessments were conducted across four waves, each spaced approximately 2 years apart over a span of 8 years. EF was assessed using six behavioral measures (animal fluency, controlled oral word association test, backward counting, serial 7s subtraction, backward digit span, symbol digit modalities test). Depression and anxiety symptoms were measured with the Neuropsychiatric Inventory. Bivariate dual latent change score (BCS) and random-intercept cross-lagged panel models (RI-CLPM) were conducted to minimize measurement error as well as adjust for autoregressive effects, regression to the mean, and between-person variances. Within persons, RI-CLPMs revealed that prior greater depression symptoms forecasted lower subsequent EF, but not vice versa (d = -0.29 vs. -0.03). BCS models showed that within-person rise in depression symptoms at a time-lag predicted EF decrements at the next time-lag, but not the opposite (d = -0.20 vs. 0.14). Further, significant, small-to-moderate, negative between-person relations between EF and depression or anxiety severity were observed (d = -0.42 to -0.26). Neither within-person cross-lagged relations (d = -0.09–-0.06) nor change-to-future change associations (d = -0.04–0.04) were found between anxiety and EF. Prospective, within-person findings offer some evidence for developmental scar theories as opposed to vulnerability models. We discussed theoretical and clinical implications, such as potentially reducing the risk of future executive dysfunction-related neuropsychiatric disorders by targeting depression symptoms.