Biology subjects

Mellem, M. S.

Publications and source records attributed to Mellem, M. S..

2 recordsLinked to original sources

bioRxiv2018

Highly Predictive Transdiagnostic Features Shared across Schizophrenia, Bipolar Disorder, and ADHD Identified Using a Machine Learning Based Approach

The Diagnostic and Statistical Manual of Mental Disorders (DSM) is the standard for diagnosing psychiatric disorders in the United States. However, evidence has suggested that symptoms in psychiatric disorders are not restricted to the boundaries between DSM categories, implying an underlying latent transdiagnostic structure of psychopathology. Here, we applied an importance-guided machine learning technique for model selection to item-level data from self-reported instruments contained within the Consortium for Neuropsychiatric Phenomics dataset. From 578 questionnaire items, we identified a set of features which consisted of 85 items that were shared across diagnoses of schizophrenia (SCZ), bipolar disorder (BD), and attention deficit/hyperactivity disorder (ADHD). A classifier trained on the transdiagnostic features reliably distinguished the patient group as a whole from healthy controls (classification AUC = 0.95) and only 10 items were needed to attain the performance level of AUC being 0.90. A sum score created from the items produced high separability between patients and healthy controls (Cohens d = 2.85), and it outperformed predefined sum scores and sub-scores within the instruments (Cohens d ranging between 0.13 and 1.21). The transdiagnostic features comprised both symptom domains (e.g. dysregulated mood, attention deficit, and anhedonia) and personality traits (e.g. neuroticism, impulsivity, and extraversion). Moreover, by comparing the features that were common across the three patient groups with those that were most predictive of a single patient category, we can describe the unique features for each patient group superimposed on the transdiagnostic feature structure. Overall, our results reveal a latent transdiagnostic symptom/behavioral phenotypic structure shared across SCZ, BD, and ADHD and present a new perspective to understand insights offered by self-report psychiatric instruments.

pathology↗

bioRxiv2018

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.

neuroscience↗