Explaining recovery from coma with multimodal neuroimaging

The aim of this prospective, observational cohort study was to investigate and assess diverse neuroimaging biomarkers to predict patients' neurological recovery after coma. 32 patients (18-76 years, M = 44.8, SD = 17.7) with disorders of consciousness participated in the study. Multimodal neuroimaging data acquired during the patient's hospitalization were used to derive cortical glucose metabolism (¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography), and structural (diffusion-weighted imaging) and functional connectivity (resting-state functional MRI) indices. The recovery outcome was defined as a continuous composite score constructed from a multivariate neurobehavioral recovery assessment administered upon the discharge from the hospital. Fractional anisotropy-based white matter integrity in the anterior forebrain mesocircuit (r = 0.72, p < .001, 95% CI: 0.87, 0.45), and the functional connectivity between the antagonistic default mode and dorsal attention resting-state networks (r = - 0.74, p < 0.001, 95% CI: - 0.46, - 0.88) strongly correlated with the recovery outcome. The association between the posterior glucose metabolism and the recovery outcome was moderate (r = 0.38, p = 0.040, 95% CI: 0.66, 0.02). Structural (adjusted R² = 0.84, p = 0.003) or functional connectivity biomarker (adjusted R² = 0.85, p = 0.001), but not their combination, significantly improved the model fit to predict the recovery compared solely to bedside neurobehavioral evaluation (adjusted R² = 0.75). The present study elucidates an important role of specific MRI-derived structural and functional connectivity biomarkers in diagnosis and prognosis of recovery after coma and has implications for clinical care of patients with severe brain injury.