Explainable Depression Detection in Clinical Interviews with Personalized Retrieval-Augmented Generation

2025-03-03

AI Summary

Depression automatic detection faces three key challenges: untrustworthy black-box models, post-hoc explanations prone to hallucination, and non-personalized retrieval strategies. To address these, we propose a clinically grounded, interpretable detection framework tailored for psychiatric interviews. First, we design a personalized query generation module that dynamically integrates user-specific background with clinical diagnostic criteria. Second, we introduce an event-centric social-intelligence knowledge retrieval mechanism to enhance LLMs’ understanding of psychosocial context and mental health–relevant events. Third, we employ retrieval-augmented generation (RAG) to produce fine-grained, clinically verifiable explanations. Evaluated on a real-world benchmark, our method significantly outperforms state-of-the-art neural networks and LLM-based baselines, achieving a +4.2% F1 improvement while ensuring traceable, evidence-based diagnostic reasoning. This work delivers the first end-to-end AI solution for mental health assessment that supports both personalized inference and clinical consensus validation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Depression is a widespread mental health disorder, and clinical interviews are the gold standard for assessment. However, their reliance on scarce professionals highlights the need for automated detection. Current systems mainly employ black-box neural networks, which lack interpretability, which is crucial in mental health contexts. Some attempts to improve interpretability use post-hoc LLM generation but suffer from hallucination. To address these limitations, we propose RED, a Retrieval-augmented generation framework for Explainable depression Detection. RED retrieves evidence from clinical interview transcripts, providing explanations for predictions. Traditional query-based retrieval systems use a one-size-fits-all approach, which may not be optimal for depression detection, as user backgrounds and situations vary. We introduce a personalized query generation module that combines standard queries with user-specific background inferred by LLMs, tailoring retrieval to individual contexts. Additionally, to enhance LLM performance in social intelligence, we augment LLMs by retrieving relevant knowledge from a social intelligence datastore using an event-centric retriever. Experimental results on the real-world benchmark demonstrate RED's effectiveness compared to neural networks and LLM-based baselines.

Problem

• Automated depression detection lacks interpretability in clinical interviews.
• Post-hoc LLM generation for explainability suffers from hallucination issues.
• Traditional retrieval systems fail to personalize for diverse user backgrounds.

Innovation

• Retrieval-augmented generation for explainable depression detection
• Personalized query generation using user-specific background
• Event-centric retriever for social intelligence enhancement