BioLLMAgent: A Hybrid Framework with Enhanced Structural Interpretability for Simulating Human Decision-Making in Computational Psychiatry

2026-03-05

AI Summary

This study addresses the trade-off between interpretability and behavioral realism in computational psychiatry, where traditional reinforcement learning models offer strong cognitive interpretability but limited behavioral fidelity, while large language models (LLMs) generate human-like behaviors yet lack transparent cognitive structure. To bridge this gap, the authors propose BioLLMAgent, a novel hybrid framework that integrates an interpretable reinforcement learning engine—modeling core decision-making mechanisms—with an external LLM that generates naturalistic behavior, coordinated through a decision-fusion mechanism. The framework maintains high parameter identifiability while accurately reproducing human behavior across six clinical and healthy datasets, including the Iowa Gambling Task (parameter correlations > 0.67). It successfully simulates cognitive behavioral therapy and, through multi-agent simulations, reveals that group-based educational interventions may outperform individualized treatment strategies.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Computational psychiatry faces a fundamental trade-off: traditional reinforcement learning (RL) models offer interpretability but lack behavioral realism, while large language model (LLM) agents generate realistic behaviors but lack structural interpretability. We introduce BioLLMAgent, a novel hybrid framework that combines validated cognitive models with the generative capabilities of LLMs. The framework comprises three core components: (i) an Internal RL Engine for experience-driven value learning; (ii) an External LLM Shell for high-level cognitive strategies and therapeutic interventions; and (iii) a Decision Fusion Mechanism for integrating components via weighted utility. Comprehensive experiments on the Iowa Gambling Task (IGT) across six clinical and healthy datasets demonstrate that BioLLMAgent accurately reproduces human behavioral patterns while maintaining excellent parameter identifiability (correlations $>0.67$). Furthermore, the framework successfully simulates cognitive behavioral therapy (CBT) principles and reveals, through multi-agent dynamics, that community-wide educational interventions may outperform individual treatments. Validated across reward-punishment learning and temporal discounting tasks, BioLLMAgent provides a structurally interpretable"computational sandbox"for testing mechanistic hypotheses and intervention strategies in psychiatric research.

Problem

• computational psychiatry
• interpretability
• behavioral realism
• reinforcement learning
• large language models

Innovation

• hybrid framework
• structural interpretability
• computational psychiatry
• reinforcement learning
• large language models