The HarmonyResearch Group

Hello!

I work at the intersection of psychology, psychiatry and computer science. My research asks what we can learn about health and illness from how people freely describe their experiences and symptoms in their own words – usually the first thing a clinician asks for. I combine psychological theory, psychometrics and artificial intelligence (AI), including natural language processing and machine learning, to develop and validate assessments of mental and physical health (e.g., depression, anxiety, PTSD, suicidality and well-being).

A central aim of my work is to develop language-based methods that are useful and trustworthy in clinical settings, held to the standards medicine applies to diagnostic tests. Together with clinical partners – including psychiatry researchers at the World Trade Center Health Program – I validate these assessments in clinical populations against expert best-estimate diagnoses, with models pre-registered before prospective testing (see the Sequential Evaluation with Model Pre-registration approach). I also work on transparent reporting (see the LEADING reporting guideline; www.leading-guideline.org) and on practical issues related to safety, bias, privacy and clinical accountability. I develop and maintain the R Language Analysis Suite, including text (www.r-text.org), topics (www.r-topics.org) and talk (www.r-talk.org), along with the Language-Based Assessment Model (L-BAM) Library, so that validated models are openly available for others to apply, replicate and audit.

I founded Minding Health (www.minding.health), where we work to translate these methods into clinical decision support for mental health assessment. I also collaborate with partners such as Mind (one of Sweden’s largest suicide prevention helplines) and SOS Alarm to explore how language-based methods can support safe, responsible, and evidence-informed practice in real-world services.

I am Co-Director of the Human Language Analysis Beings (HLAB) at Vanderbilt University’s College of Connected Computing (www.humanlanguage.org). HLAB develops human-centered NLP and machine-learning methods, with the goal of advancing AI language understanding and supporting more accurate, data-driven assessment and understanding of human and social psychological health.

In my non-academic time: I love spending quality time with my family and playing football.

Interests

  • Psychology
  • Mental Health
  • Artificial Intelligence
  • R

Publications

← All people