Natural Language Processing and Conversational Technologies for Clinical Practice
Theme overview
Natural language processing (NLP) can be used in mental health to analyse language patterns in written or spoken communication to identify symptoms of mental health disorders such as depression, anxiety, and bipolar disorder. NLP can also be used to create personalized therapy and support for clients by analysing their communication and tailoring interventions to their specific needs. Chatbots can be used in mental health to provide accessible and low-cost support to individuals with mental health concerns. They can also monitor mental health symptoms and provide referrals to mental health professionals if necessary.
Whilst much of the work regarding NLP, chatbots and mental health focuses on analysing an individual’s language for insights into their mental health or using chatbots as virtual therapy agents, we are conducting work on using NLP and chatbots to inform and aid the mental health clinician.
Firstly, by analysing therapy transcripts we can generate insights into the language and dialogue characteristics that are associated with better psychotherapy outcomes. This in turn could lead to tools that guide clinicians in their psychotherapy practice, including systems that provide real-time feedback during the course of a therapy session.
Secondly, apart from using chatbots as virtual therapists, the capacity for modern conversational technology to generate extended dialogue opens up the possibility of creating chatbots that can simulate mental health clients. We are exploring this idea and our Client101 project is researching and developing a conversational platform that can be used by mental health professionals, particularly trainee clinicians, to engage in simulated therapy interactions and practice psychotherapy techniques.
Projects/funding
- Using conversational agents and natural language processing to train and evaluate psychologists
- Client101: Embedding a psychotherapy training chatbot platform into University of Melbourne psychology curricula
Publications
- Daniel Cabrera Lozoya, Jiahe Liu, Simon D’Alfonso, and Mike Conway. 2024. Optimizing Multimodal Large Language Models for Detection of Alcohol Advertisements via Adaptive Prompting. In Proceedings of the 23rd Workshop on Biomedical Natural Language Processing, pages 514–525, Bangkok, Thailand. Association for Computational Linguistics.
- Daniel Lozoya, Alejandro Berazaluce, Juan Perches, Eloy Lúa, Mike Conway, and Simon D’Alfonso. 2024. Generating Mental Health Transcripts with SAPE (Spanish Adaptive Prompt Engineering). In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 5096–5113, Mexico City, Mexico. Association for Computational Linguistics.
- D. C. Lozoya, S. D’Alfonso and M. Conway, "Identifying Gender Bias in Generative Models for Mental Health Synthetic Data," 2023 IEEE 11th International Conference on Healthcare Informatics (ICHI), Houston, TX, USA, 2023, pp. 619-626, doi: 10.1109/ICHI57859.2023.00109. keywords: {Analytical models;Systematics;Medical treatment;Mental health;Feature extraction;Transformers;Data models;Generative Models;Natural Language Processing;Mental Health;Bias;Fairness in AI},
Theme members
- Daniel Cabrera Lozoya, PhD Candidate
Kejian Cui, PhD Candidate