Semantic Processing in Unresponsive Patients With Acute Brain Injury

Background and Objectives Early prognostication after acute brain injury (ABI) is challenging because behavioral responsiveness may not accurately reflect preserved cognitive function. Neural responses to semantic incongruity can identify language processing despite the absence of overt behavior. We investigated whether bedside EEG could detect semantic processing in unresponsive patients with ABI and whether its presence was associated with long-term outcomes. Methods In this prospective 2-center cohort study, unresponsive adults with ABI admitted to neurologic intensive care units between 2022 and 2025 underwent bedside EEG during a language paradigm comparing congruent and incongruent sentence endings (e.g., “my favorite color is green/soon”). Semantic processing was defined by differential event-related responses within centroparietal channels (300–500 milliseconds) using cluster-based permutation testing with false discovery rate correction. Outcomes were assessed for recovery of consciousness and functional recovery (Glasgow Outcome Scale–Extended score ≥4) at 12 months by blinded interviewers. Diagnostic performance was evaluated using sensitivity, specificity, predictive values, accuracy, and likelihood ratios. Results Among 105 patients (mean age 58 ± 19 years; 21% women; 53% Hispanic/Latino), the most common etiologies were intracerebral hemorrhage (45%), traumatic brain injury (43%), and subarachnoid hemorrhage (10%). Withdrawal of life-sustaining therapy occurred in 31%. Semantic processing was detected in 13 patients (12%). Recovery of consciousness occurred in 77% of patients with semantic processing compared with 49% without semantic processing (odds ratio [OR] 3.5; 95% CI 0.9–13.5, p = 0.08). Functional recovery occurred in 42% with semantic processing compared with 14% without semantic processing (OR 4.5; 95% CI 1.2–16.6, p = 0.03). Semantic processing was highly specific but insensitive for predicting both recovery of consciousness (specificity 94%, sensitivity 18%) and functional recovery (specificity 92%, sensitivity 29%). After excluding patients who underwent withdrawal of life-sustaining therapy or died of non-neurologic medical complications, specificity increased to 100% for recovery of consciousness and 95% for functional recovery. Discussion EEG evidence of semantic processing can be detected in a subset of unresponsive patients with ABI during the acute phase of injury and is associated with favorable long-term outcomes. Semantic processing was a specific but insensitive biomarker of recovery, suggesting potential utility as a rule-in marker of preserved cognitive function. These findings are hypothesis-generating and require validation in larger multicenter cohorts before clinical implementation.