Development and application of a diagnostic algorithm for posttraumatic stress disorder

Abstract

Intact cognitive functions rely on synchronous neural activity; conversely, alterations in synchrony are thought to underlie psychopathology. We recently demonstrated that anomalies in synchronous neural interactions (SNI) determined by magnetoencephalography represent a putative PTSD biomarker. Here we develop and apply a regression-based diagnostic algorithm to further validate SNI as a PTSD biomarker in 432 veterans (235 controls; 138 pure PTSD; 59 PTSD plus comorbid disorders). Correlation coefficients served as proximities in multidimensional scaling (MDS) to obtain a two-dimensional representation of the data. In addition, least absolute shrinkage and selection operator (LASSO) regression was used to derive a diagnostic algorithm for PTSD. Performance of this algorithm was assessed by the area under the receiver operating characteristic (ROC) curves, sensitivity, and specificity in 1000 randomly divided testing and validation datasets and in independent samples. MDS revealed that individuals with PTSD, regardless of comorbid psychiatric conditions, are highly distinct from controls. Similarly, application of the LASSO regression-derived prediction model demonstrated remarkable classification accuracy (AUCs≥0.93 for men, AUC=0.82 for women). Neural functioning in individuals with PTSD, regardless of comorbid psychiatric diagnoses, can be used as a diagnostic test to determine patient disease status, further validating SNI as a PTSD biomarker.