Machine learning models identify multimodal measurements highly predictive of transdiagnostic symptom severity for mood, anhedonia, and anxiety

Background: Insights from neuroimaging-based biomarker research have not yet translated into clinical practice. This translational gap could be because of a focus of psychiatric biomarker research on diagnostic classification, rather than on prediction of transdiagnostic psychiatric symptom severity. Currently, no transdiagnostic, multimodal predictive models of symptom severity that include neurobiological characteristics have been described. Methods: We built predictive models of three common symptoms in psychiatric disorders (dysregulated mood, anhedonia, and anxiety) from the Consortium for Neuropsychiatric Phenomics dataset (n=272) which contains clinical scale assessments, resting-state functional-MRI (rs-fMRI) and structural-MRI (sMRI) imaging measures from patients with schizophrenia, bipolar disorder, attention deficit and hyperactivity disorder, and healthy control subjects. We used an efficient, data-driven feature selection approach to identify the most predictive features from these high-dimensional data. Results: This approach optimized modeling and explained 65-90% of variance across the three symptom domains, compared to 22% without using the feature selection approach. The top performing multimodal models retained a high level of interpretability which enabled several clinical and scientific insights. First, to our surprise, structural features did not substantially contribute to the predictive strength of these models. Second, the Temperament and Character Inventory scale emerged as a highly important predictor of symptom variation across diagnoses. Third, predictive rs-fMRI connectivity features were widely distributed across many intrinsic resting-state networks (RSN). Conclusions: Combining rs-fMRI with select questions from clinical scales enabled high levels of prediction of symptom severity across diagnostically distinct patient groups and revealed that connectivity measures beyond a few intrinsic RSNs may carry relevant information for symptom severity.