Grant: CIHR Fellowship - PA: PPHE Research - Priority Areas

Engagement between the patient and the therapist is a key treatment factor for individuals with depression and has been known to improve therapeutic results. After COVID-19, online therapy has become a well-received alternative to delivering care. However, it remains unclear whether patients are as engaging online as in person. This project aims to develop a software system that automatically measures the patient-therapist engagement level in an online therapy session. The therapists can use this objective measure to gauge the effectiveness of sessions. As men, women and 2S/LGBTQI+ have different mannerisms and social norms, tailored subsystems will be created for different genders. As individual patients may have specific comfort zones and preconceptions about therapists, the system can provide insights on whether the patient-therapist pairs are good fits from interactive parameters extracted from the sessions. The system will use multiple data sources, including text, voice, facial expressions, and clinical information. This project leverages video recordings from a previous in-person study to build an initial model. The model will then be refined and tested using recordings of Zoom therapy sessions. Advanced computer science techniques, including voice and speech analysis, video data processing, and natural language processing (NLP), will be used to extract meaningful parameters related to engagement. Speech analysis will detect emotional tone, energy, and voice patterns, while video data will track gaze direction, facial expressions, and body language such as head movements and gestures. NLP will assess conversation dynamics, including emotional tone, relevance, and depth of thought. Once developed, this software system will inform therapists of patient engagement in real-time during online therapy, helping to ensure that virtual care remains as effective as in-person sessions.