CFiCS: Graph-Based Classification of Common Factors and Microcounseling Skills

2025-03-28

AI Summary

This study addresses the fine-grained identification of generic therapeutic factors and micro-level counseling skills in psychotherapy dialogue texts—a challenging task due to context sensitivity and hierarchical nesting. To tackle this, we propose the first inductive graph neural network framework integrating a therapy-concept heterogeneous graph with ClinicalBERT. Specifically, we construct an explicit heterogeneous graph encoding hierarchical concept relationships and jointly leverage ClinicalBERT’s semantic representations and GNNs for inductive node embedding—enabling generalization to unseen, zero-connection samples. Additionally, we design a hierarchical multi-task classification architecture to jointly optimize macro-level factor and micro-level skill recognition. Experiments demonstrate that our method significantly outperforms baselines—including Random Forest, BERT-based multi-task models, and standalone graph models—on both macro- and micro-level F1 scores, particularly improving fine-grained skill classification accuracy. The approach provides an interpretable and generalizable computational foundation for modeling psychotherapeutic processes.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Common factors and microcounseling skills are critical to the effectiveness of psychotherapy. Understanding and measuring these elements provides valuable insights into therapeutic processes and outcomes. However, automatic identification of these change principles from textual data remains challenging due to the nuanced and context-dependent nature of therapeutic dialogue. This paper introduces CFiCS, a hierarchical classification framework integrating graph machine learning with pretrained contextual embeddings. We represent common factors, intervention concepts, and microcounseling skills as a heterogeneous graph, where textual information from ClinicalBERT enriches each node. This structure captures both the hierarchical relationships (e.g., skill-level nodes linking to broad factors) and the semantic properties of therapeutic concepts. By leveraging graph neural networks, CFiCS learns inductive node embeddings that generalize to unseen text samples lacking explicit connections. Our results demonstrate that integrating ClinicalBERT node features and graph structure significantly improves classification performance, especially in fine-grained skill prediction. CFiCS achieves substantial gains in both micro and macro F1 scores across all tasks compared to baselines, including random forests, BERT-based multi-task models, and graph-based methods.

Problem

• Automatic identification of therapeutic change principles from text.
• Challenges due to nuanced, context-dependent therapy dialogue.
• Hierarchical classification of common factors and counseling skills.

Innovation

• Graph machine learning with ClinicalBERT embeddings
• Heterogeneous graph for therapeutic concepts
• Graph neural networks for inductive learning