https://digimindlab.ai/research/Recent content in Research Interests on DigiMind LabHugoenhttps://digimindlab.ai/research/digital/Mon, 01 Jan 0001 00:00:00 +0000https://digimindlab.ai/research/digital/We study telemedicine and other digital delivery methods to provide mental health treatment at scale. Working with digital health stakeholders and real-world samples of more than 10,000 patients, we have found that digital delivery preserves the safety and efficacy of both psychotherapy and pharmacological interventions while expanding access. Because every interaction in these settings is captured through the digital surface, they also let us study therapeutic processes with NLP and language models.https://digimindlab.ai/research/computational/Mon, 01 Jan 0001 00:00:00 +0000https://digimindlab.ai/research/computational/We develop machine learning methods that move beyond diagnostic categories to characterize how symptoms are organized and how they change over time. This includes VISTA-SSM, our clustering method for noisy, irregular, and incomplete longitudinal data, and trajectory modeling that identifies distinct courses of adjustment to stress, loss, and treatment and the predictors that distinguish them. We also use network analysis to map relationships among symptoms and, most recently, to characterize the temporal dynamics of patient emotions turn by turn in psychotherapy samples of over thirty thousand patients.https://digimindlab.ai/research/human-ai/Mon, 01 Jan 0001 00:00:00 +0000https://digimindlab.ai/research/human-ai/As conversational AI enters mental health care, we examine how people interact with it. We test whether language model responses are perceived as empathic and supportive, analyze tens of thousands of real conversations between users and AI to characterize safety-relevant behavior, and study engagement and outcomes when generative AI is deployed for social and mental health support. We also build adversarial user simulations that expose how systems fail, so that these interactions can be evaluated and improved before and during deployment.https://digimindlab.ai/research/llms/Mon, 01 Jan 0001 00:00:00 +0000https://digimindlab.ai/research/llms/We study the empirical and regulatory foundations needed to deploy large language models (LLMs) safely in mental health care. We conducted among the first real-world evaluations of a conversational AI agent built for anxiety and depression, and we design simulated patients to stress-test conversational systems before they reach people. To guide development, we authored a research framework for LLM-based interventions and an implementation framework covering regulatory considerations, ethical safeguards, and deployment barriers specific to mental health.https://digimindlab.ai/research/nlp/Mon, 01 Jan 0001 00:00:00 +0000https://digimindlab.ai/research/nlp/We build natural language processing (NLP) tools for scalable, objective mental health monitoring, so that care no longer depends only on infrequent self-report. We proposed a research framework that maps NLP intervention targets, evidence gaps, and methodological standards, and we conducted the first evaluation of large language models’ ability to measure psychiatric functioning. We identify linguistic, acoustic, and emotional markers in communication between patients and clinicians that track symptoms and predict treatment outcomes, enabling passive monitoring at scale.