DECODING CHRONIC PAIN: SYMPTOM DIMENSIONS, PAIN MODULATION, AND PERIAQUEDUCTAL GRAY CONNECTIVITY

Abstract

Chronic pain is a complex condition that may not be adequately explained by diagnosis alone, motivating a symptom-based approach to phenotyping. Such symptom-defined phenotypes, informed by cognitive and neural mechanisms, may enable more precise stratification and improve treatment outcomes. This study examined 159 patients with fibromyalgia and chronic back pain, along with 72 healthy controls, and grouped patients into high- and low-severity phenotypes based on pain intensity, disability, and affective burden. Groups were then evaluated across behavioral, cognitive, and neural markers. In an expectation-induced pain modulation task, high-burden patients showed impaired modulation when positive expectations were violated, which predicted greater catastrophizing and hypervigilance. Resting-state fMRI demonstrated altered periaqueductal gray (PAG) connectivity, with high-burden patients exhibiting more negative dorsolateral/lateral PAG–dorsomedial prefrontal coupling. Stepwise machine learning classified phenotype membership with accuracy above chance, improved by integrating neuroimaging with behavioral features. These findings suggest a severe chronic pain phenotype with distinct behavioral and neural markers, supporting mechanism-based stratification.