Season 3 Episode 10: "Can We Predict Psychiatric Diagnoses? Advances in Computational Psychiatry"- Natalia Mota MD PhD

In this episode of Making Therapy Better, Dr. Bruce Wampold speaks with psychiatrist and researcher Natalia Mota about a rapidly emerging field with major clinical implications: computational psychiatry.

Bruce opens by revisiting the long-standing tension between clinical judgment and statistical prediction, tracing it back to Paul Meehl’s landmark work and Daniel Kahneman’s insights about confidence versus accuracy. From there, the conversation moves into Natalia’s core contribution: using natural language processing and graph theory to quantify patterns in speech that clinicians often sense intuitively—especially in cases involving formal thought disorder.

Natalia explains how she represents language as speech graphs (words as nodes, sequences as directed edges) to measure connectedness, fragmentation, repetition cycles, and complexity. This approach enabled her team to distinguish—using early narratives like dream reports and emotionally cued storytelling—between clinical trajectories that later converged on schizophrenia versus bipolar disorder, with high sensitivity and specificity. Importantly, the discussion highlights why this work differs from “black box” AI: interpretability matters, and predictions must be anchored in psychopathology—not confounded social variables.

The episode also explores a crucial caution: language is shaped by education, literacy, development, culture, and socioeconomic context. Natalia shares findings showing that narrative complexity correlates strongly with formal education in typical populations, and discusses how severe symptoms can override expected developmental gains. The conversation expands beyond psychosis into related applications, including research on aging and Alzheimer’s disease, where different memory systems appear to drive changes in narrative connectedness.

The episode concludes with a broader reflection on today’s youth mental health landscape: rising rates of psychological distress, the role of social context and technology, and why we should think like epidemiologists and public health professionals—protecting the school and social environment, keeping the human in the loop, and avoiding premature pathologization.