Examining Trajectories of Momentary Affect Surrounding NSSI and Alcohol Use: A Network Analysis and Bayesian Approach

This study aimed to identify specific affective antecedents of non-suicidal self-injury (NSSI) and alcohol use, and to examine affective trajectories before and after these behaviors.

Fifty-six undergraduates (71% female, 25% non-White, 20.2 ± 1.3 years old) meeting criteria for probable alcohol or substance use disorder participated in 28-days of ecological momentary assessment, completing four daily surveys assessing affect, self-punishment cognitions, and engagement in NSSI and alcohol use. Contemporaneous and time-lagged graphical vector autoregressive (GVAR) models were used to examine affect dynamics surrounding NSSI and alcohol use. Zero-inflated Bayesian binomial regression was used to investigate the relationship between momentary affect and likelihood of engaging in either NSSI or alcohol at the next prompt. Multilevel modeling was used to examine trajectories of affective states pre- and post-engagement in NSSI and alcohol use.

GVAR models revealed significant relationships between jitteriness and being afraid with next timepoint NSSI, and excitement with next timepoint alcohol use. Bayesian regression with horseshoe priors identified self-punishment and loneliness as predictors of NSSI. Moreover, Bayesian regression revealed that excitement predicted alcohol use among those without alcohol use disorder (AUD), while belongingness predicted alcohol use among those with AUD. Multilevel models revealed self-punishment and loneliness followed a negative quadratic trajectory around NSSI events. Excitement and belongingness did not relate to slopes of time.

The findings suggest NSSI is used to reduce negative emotional states, whereas alcohol consumption initially enhances positive emotional states in individuals without an AUD and may be used to increase belongingness in those with an AUD.