Our lab leverages technological breakthroughs in artificial intelligence (AI) and natural language processing (NLP), particularly the emergence of large language models (LLM), alongside machine learning (ML) and signal processing, which now enable automated and continuous encoding of nonverbal and verbal behavior from video, audio, text, and physiological signals (Schwartz et al., 2023). By focusing on granular in-session elements using automated methods, we are able to scale larger samples and yield more precise insights into the intrapersonal and interpersonal dynamics that facilitate (or block) a positive outcome (Imel et al., 2025).  

For example, utilizing a novel, question-specific multimodal AI framework, which dynamically fuses text, audio, and video data to analyze distinct symptoms, can revolutionize automated depression assessments (Mandal et al., 2025). In another study, using social media posts, state-of-the-art LLMs have outperformed anonymous online peer responses (Liu et al., 2025). Also, using transformer-based emotion recognition models to analyze 139,061 patient utterances, we found that coherence in negative emotions significantly predicted patients’ improved functioning (Atzil-Slonim et al., 2024).

References:

Imel, Z. E., Creed, T., Kious, B., Althoff, T., Atzil-Slonim, D., & Srikumar, V. (2025). A Framework for Automation in Psychotherapy. Current Directions in Psychological Science, 0(0).

Schwartz, B., Uhl, J., & Atzil-Slonim, D. (2023). Assessments and measures in psychotherapy research: going beyond self-report data. Frontiers in Psychiatry, 14, 1276222.

Mandal, A., Atzil-Slonim, D., Solorio, T., & Gurevych, I. (2025, May). Enhancing Depression Detection via Question-wise Modality Fusion. In Proceedings of the 10th Workshop on Computational Linguistics and Clinical Psychology (CLPsych 2025) (pp. 44-61).

Liu, C. C., Arnaout, H., Kovačić, N., Atzil-Slonim, D., & Gurevych, I. (2025). Tailored Emotional LLM-Supporter: Enhancing Cultural Sensitivity. 

Atzil-Slonim, D., Eliassaf, A., Warikoo, N., Paz, A., Haimovitz, S., Mayer, T., & Gurevych, I. (2024). Leveraging natural language processing to study emotional coherence in psychotherapy. Psychotherapy, 61(1), 82.