Predicting long-term prognosis in comatose patients through brain network analysis under name-evoked stimulation

Accurate prognosis assessment of comatose patients remains a significant challenge in neurocritical care. Growing evidence indicates that brain connectivity is integral to the maintenance of consciousness and may be linked to its recovery. In this study, we recorded bedside electroencephalography (EEG) from comatose patients during an auditory oddball name-calling task to investigate task-related dynamic causal modeling (DCM) connectivity and to examine whether connectivity strengths correlated with patients' functional recovery. Our findings reveal that a bidirectional model, incorporating reciprocal connectivity among the superior frontal gyri, superior parietal lobules, and primary auditory cortices, was significantly associated with the neural processing of name-calling stimuli in comatose patients. Furthermore, the strength of these DCM connections demonstrated a capacity to predict long-term prognostic outcomes, as evaluated via the Glasgow Outcome Scale-Extended scale. Together, these results provide evidence supporting the potential of DCM-derived biomarkers in evaluating functional prognosis in comatose patients. (ChiCTR2000033586).