Measuring Mental Health Variables in Computational Research: Toward Validated, Dimensional, and Transdiagnostic Approaches

TL;DR

Computational mental health research often relies on invalid or inappropriate measures of psychopathology, compromising model validity. The paper argues for validated, dimensional, and transdiagnostic measurement approaches and provides practical guidance to adopt them. It discusses the limitations of self-report, the superiority of clinician-administered or well-validated instruments, and the dimensional and transdiagnostic frameworks (HiTOP, RDoC) to better capture psychopathology and comorbidity. It also outlines risks of poor measurement, such as reduced power and biased conclusions, and proposes actionable steps and data-sharing strategies, including an n+1 measurement approach and large clinical NLP datasets. The work aims to improve the validity and clinical utility of computational models in mental health.

Abstract

Computational mental health research develops models to predict and understand psychological phenomena, but often relies on inappropriate measures of psychopathology constructs, undermining validity. We identify three key issues: (1) reliance on unvalidated measures (e.g., self-declared diagnosis) over validated ones (e.g., diagnosis by clinician); (2) treating mental health constructs as categorical rather than dimensional; and (3) focusing on disorder-specific constructs instead of transdiagnostic ones. We outline the benefits of using validated, dimensional, and transdiagnostic measures and offer practical recommendations for practitioners. Using valid measures that reflect the nature and structure of psychopathology is essential for computational mental health research.