A Computational Phenotype of Disrupted Moral Inference in Borderline Personality Disorder

Borderline personality disorder (BPD) is a serious mental disorder characterized by marked interpersonal disturbances, including difficulties trusting others and volatile impressions of others' moral character, often resulting in premature relationship termination. We tested a hypothesis that moral character inference is disrupted in BPD and sensitive to democratic therapeutic community (DTC) treatment.

Participants with BPD (n = 43; 20 untreated and 23 DTC-treated) and control participants without BPD (n = 106) completed a moral inference task where they predicted the decisions of 2 agents with distinct moral preferences: the "bad" agent was more willing than the "good" agent to harm others for money. Periodically, participants rated their subjective impressions of the agent's moral character and the certainty of those impressions. We fit a hierarchical Bayesian learning model to participants' trialwise predictions to describe how beliefs about the morality of the agents were updated by new information.

The computational mechanisms of moral inference differed for patients with untreated BPD relative to matched control participants and patients with DTC-treated BPD. In patients with BPD, beliefs about harmful agents were more certain and less amenable to updating relative to both control participants and participants who were treated with DTC.

The findings suggest that DTC may help the maintenance of social relationships in BPD by increasing patients' openness to learning about adverse interaction partners. The results provide mechanistic insights into social deficits in BPD and demonstrate the potential for combining objective behavioral paradigms with computational modeling as a tool for assessing BPD pathology and treatment outcomes.