Current Research and Scholarly Interests
Precision Mental Health: From Brain Circuits to Better Treatment Decisions
Our research advances precision mental health by moving beyond diagnosis and trial-and-error care toward biologically informed, personalized treatment decisions. Our work connects brain circuits, behavior, symptoms and treatment response, from standardized functional imaging probes to circuit biotypes, treatment prediction and clinical decision support.
A central foundation is Precision Circuit Mapping (PCM), our standardized framework for translating individual differences in brain circuit function into clinically actionable measures for treatment selection.
1. Precision Circuit Mapping
PCM combines resting-state imaging with brief emotion and cognitive task probes to measure circuits across contexts. Task-based measures can reveal treatment-relevant differences in circuit engagement that are not captured by resting-state connectivity alone and have been important in distinguishing response across treatments.
The framework is designed for multisite and clinical workflows. A standardized imaging protocol can be completed in under 30 minutes, with end-to-end automated quality control, processing and norm-referenced patient-level circuit measures generated in under 15 minutes.
2. Brain Circuit Biotypes and Individual Differences
Mental health diagnoses encompass substantial biological heterogeneity. We identify reproducible patterns of dysfunction across circuits supporting cognitive control, attention, salience, positive and negative affect, and self-referential processing. Our brain circuit “biotypes” connect these differences with symptoms, behavior and clinical outcomes.
3. Precision Therapeutics and Treatment Selection
We test whether circuit measures can distinguish likely response and nonresponse across treatments. Our studies span antidepressant medications, behavioral interventions and neuromodulation, mechanistically selective repurposed treatments, and emerging therapeutics including ketamine, MDMA-assisted therapy and psilocybin.
Across different treatment mechanisms, we ask: Which treatment is most likely to work for which patient, and what biological information is needed to make that decision?
4. Explainable, Human-Centered AI for Clinical Decision Support
We develop explainable, human-centered AI to translate patient-level circuit information into transparent clinical decision support for treatment selection. Rather than relying on black-box prediction, our approach is designed to show clinicians the circuit measures contributing to a patient’s profile, the treatment-relevant evidence associated with those measures, and how that information can inform a clinical decision.
Our goal is circuit-guided treatment selection that augments rather than replaces clinical judgment, preserving interpretability, traceability to evidence and clinician oversight. The emphasis is on transparency, usability and clinical utility, not prediction alone.
5. From Clinical Decision Support to Real-World Use
Translation has been a goal of our research from the outset. We develop and test the clinical, technical, regulatory and implementation pathways required to move validated precision measures and decision support into real-world mental health care. Our work spans multisite implementation, standardized acquisition and clinical workflows, prospective clinical studies, regulatory translation and evaluation of clinical utility.
The central question is not simply whether a biomarker is statistically significant, but whether it can be measured reliably, delivered efficiently, interpreted transparently and used to improve a real clinical decision.
Ultimately, our goal is to replace prolonged trial-and-error treatment with objective information that helps people receive an effective treatment sooner.
Learn more about our research:
Stanford Center for Precision Mental Health: http://med.stanford.edu/pmhw
