How can artificial intelligence and large language models help more people get science-backed treatments for mental disorders? How do psychological processes like depression and anxiety show up in human language? What is the underlying structure of psychopathology? I use computational linguistics and machine learning to address these questions.
I am a postdoctoral researcher at the Stanford Institute for Human-Centered AI working in the Computational Psychology and Well-Being and Fidelity, Adaptation, Sustainability, and Training labs. I did my PhD in Psychology at the University of Pennsylvania in the Boundaries of Anxiety and Depression lab. I’m part of the World Well-Being Project. My research has been supported by the National Science Foundation.
I use evidence-based therapies, including CBT for depression and insomnia, exposure therapy for the anxiety disorders, and cognitive processing therapy and prolonged exposure for PTSD.
I write about the future of AI and mental heath, including how large language models can improve research and treatment, what needs to happen to do this effectively, and how we will know when these applications are ready to deploy.
I also develop products that use large language models to improve assessment and treatment of psychopathology. Right now I’m building an AI-based CBT homework helper that aims to make therapy homework better for patients, and an AI-based platform to help therapists learn gold-standard treatments.
I have also used natural language processing to characterize and measure (and disentangle!) depression and anxiety and repetitive negative thinking. Some of my current work investigates what drives the relationship between depression and first-person singular pronoun use (e.g., I, me, mine).
Depression and anxiety are not categorical (they are dimensional – severity matters) and cut across traditional diagnostic classes (they are transdiagnostic). The prevailing classification system, DSM-5, is neither transdiagnostic nor dimensional. This is a problem! My work has developed and tested a transdiagnostic, dimensional anxiety classification system as a possible alternative.
Beyond these research streams, I’m interested in developing behavioral measures of psychological constructs, like using a joystick to measure characteristics of thought. I also use empirical methods to disentangle or understand the overlap between closely related psychological constructs, like worry and rumination.
Stade, E. C., Eichstaedt, J. E., Kaysen, D. Salecha, A., Greenberger, A., Singhbi, S., & Wiltsey Stirman, S., (preprint). TherapyTrainer: Using AI to train therapists in written exposure.
Stade, E. C., Wiltsey Stirman, S., Held, P., Schwartz, H. A., & Eichstaedt, J. E., (preprint). Designing clinical psychological AI that reduces suffering: Challenges and technical considerations.
Stade, E. C., Eichstaedt, J. E., Kim, J. P., & Wiltsey Stirman, S., (preprint). Readiness Evaluation for AI Deployment and Implementation for Mental Health: A review and framework.
Ganesan, A. V., Varadarajan, V., Lal, Y. K., Eijsbroek, V. C., Kjell, K., Kjell, O. N. E., Dhanasekaran, T., Stade, E. C., Eichstaedt, J. C., Boyd, R. L., Schwartz, H. A., & Flek, L. (preprint). Explaining GPT-4’s schema of depression using machine behavior analysis.
Stade, E. C., Wiltsey Stirman, S., Ungar, L., Boland, C. L., Schwartz, H. A., Yaden, D. B., Sedoc, J., DeRubeis, R. J., Willer, R., Eichstaedt, J. E. (2024). Large language models could change the future of behavioral healthcare: a proposal for responsible development and evaluation. npj Mental Health Research.
Rai, S., Stade, E. C., Giorgi, S., Francisco, A., Ungar, L., Curtis, B., Guntuku, S. C. (2024). Key language markers of depression depend on race. PNAS.
Stade, E. C., Ungar, L., Eichstaedt, J. C., Sherman, G., Ruscio, A. M. (2023). Depression and anxiety have distinct and overlapping language patterns: Results from a clinical interview. Journal of Psychopathology and Clinical Science.
Stade, E. C., DeRubeis, R. J., Ungar, L., & Ruscio, A. M. (2023). A transdiagnostic, dimensional classification of anxiety shows improved parsimony and predictive noninferiority to DSM. Journal of Psychopathology and Clinical Science.
Stade, E. C., Ungar, L., Havaldar, S., Ruscio, A. M. (2023). Perseverative thinking is associated with features of spoken language. Behaviour Research and Therapy.
Stade, E. C., & Ruscio, A. M. (2022). A meta-analysis of the relationship between worry and rumination. Clinical Psychological Science.
Stade, E. C., Cohen, R., Loftus, P., & Ruscio, A. M. (2022). A novel measure of real-time perseverative thought. Clinical Psychological Science. 10(3), 534–552.
Costi, S., Soleimani, L., Glasgow, A., Brallier, J., Spivack, J., Schwartz, J., Levitch C. F., Richards, S., Hoch, M., Stade, E. C., Welch, A., Collins, K. A., Feder, A., Iosifescu, D. V., Charney, D. S., & Murrough, J. W. (2019). Lithium continuation therapy following ketamine in patients with treatment resistant unipolar depression: a randomized controlled trial. Neuropsychopharmacology, 44, 1812–1819.
Murrough, J. W., Stade, E. C., Sayed, S., Ahle, G., Kiraly, D. D., Welch, A., Collins, K. A., Soleimani, L., Iosifescu, D. V., & Charney, D. S. (2017). Dextromethorphan/quinidine pharmacotherapy in patients with treatment resistant depression: A proof of concept open label clinical trial. Journal of Affective Disorders, 218, 277-283.
Stade, E. C., & Iosifescu, D.V. (2016). Using EEG for treatment guidance in major depressive disorder. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 1(5), 411-422.
OSF | ORCID | Google Scholar