Ascertaining Diagnosis Classification With Elicited Speech

Major psychiatric disorders often occur together in the same patient and it can be difficult to distinguish between disorders with overlapping symptoms. The majority of patients with mental illness receive treatment in settings where it is not feasible to conduct detailed diagnostic interviews or neuropsychiatric testing. A cost-effective and efficient tool is needed for accurate diagnosis. Looking at language is an efficient way of making sense of the brain because it is easily observed and reflects brain circuitry. Automated natural language processing (NLP) can help us do this objectively, efficiently, and with high sensitivity. The investigators aim to use linguistic features extracted using natural language processing to aid in diagnostic classification and in predicting dimensional symptoms. This will be done by obtaining clinical diagnoses, language samples, and self-report scales from 604 participants with a variety of psychiatric symptoms. Adolescents will also be included and the investigators will try to predict diagnosis in adolescents using models that were built on adult data.