PhD Scientific Days 2025

Mental Health Sciences III.

Investigation of Emotional Instability in Depression and Borderline Personality Disorder Using a Computational Model

Name of the presenter

Unoka Sámuel

Institute/workplace of the presenter

Semmelweis University

Authors

Unoka Sámuel¹

1: Semmelweis University

Text of the abstract

Introduction:
Symptoms of depression are present in several psychiatric disorders such as Major Depressive Disorder (MDD), Bipolar Disorder (BD), and Borderline Personality Disorder (BPD), albeit with distinct pathomechanisms. Despite this, patients often achieve similar scores on standard depression questionnaires. We hypothesized that during risky decision-making, emotional reactivity is differently influenced by expected value (EV), certain reward (CR), outcome, and reward prediction error (RPE) across depression subtypes and healthy controls (HC).
Aims:
Our aim was to differentiate BPD, depression without personality disorder (DEP), and HC groups based on momentary mood fluctuations during risky decision-making, using computational modeling.
Methods:
A total of 33 BPD, 17 DEP, and 51 HC participants were included. The effect of risky choices on mood was assessed using an established paradigm, while influencing factors were analyzed via mathematical models. Hierarchical Bayesian analysis was applied to model the effects of EV, CR and RPE on mood. Parameters were compared based on their High Density Intervals (HDI).
Results:
RPE had a significantly stronger influence on mood in the BPD group compared to both DEP and HC groups (the 95% HDI of the difference did not include zero). In the DEP group, this weight was slightly lower than in the HC group. In contrast, the weight of CR was significantly higher in the DEP group than in both the BPD and HC groups. BPD participants also showed higher CR sensitivity than HC. Baseline mood was lowest in the DEP group, followed by BPD, with HC showing the highest mood levels (none of the 95% HDIs included zero).
Conclusion:
The three groups could be distinguished based on their model parameters, suggesting distinct emotional processing mechanisms across depression subtypes. These findings may support the development of novel diagnostic tools and targeted treatments, as well as contribute to understanding the biological underpinnings of depression.
Funding:
This research was supported by the Egyetemi Kutatói Ösztöndíj Program (EKÖP).

e-mail address: unokasamuel2003@gmail.com
the name of the University: Semmelweis University
the name of the Supervisor: Dr. Csukly Gábor
I would like to present it as an oral presentation.