Summary: Brain responses to spoken language can be detected with EEG in unresponsive traumatic brain injury patients. The strength of those responses helps predict recovery.
Source: University of Birmingham
New research suggests that measuring how unresponsive patients with severe traumatic brain injury (TBI) process spoken language can offer valuable prognostic information for clinicians.
A research team at the University of Birmingham demonstrated that electroencephalography (EEG), a non-invasive method for recording brain electrical activity, can detect cortical responses to speech in patients who show no outward signs of awareness. By quantifying how strongly a patient’s brain synchronises to the rhythm of phrases and sentences, clinicians can improve predictions about recovery and make better-informed treatment decisions.
Importantly, this technique is entirely passive: it can be applied to patients in intensive care who are not able to follow commands or provide any voluntary response. The assessment does not require the patient to perform tasks or to show behavioural reactions, which makes it suitable for the critical early days after severe head injury.
In the study published in Annals of Neurology, the investigators recruited 28 patients with acute TBI who were sedation-free and not obeying commands within a few days of injury. While recording EEG, the team played streams of isochronous monosyllabic words that formed meaningful phrases and sentences. In healthy listeners, EEG activity aligns—or phase-locks—with the rhythmic structure of phrases and sentences only when speech is consciously understood. The researchers used this neural phase synchrony as an index of covert speech comprehension in the unresponsive patients.
The research team were able to obtain functional outcome data on 17 patients at three months and 16 patients at six months after injury using the Glasgow Outcome Scale Extended. They found a strong association between the acute strength of cortical tracking of phrases and sentences and later recovery: greater neural synchrony to speech predicted better outcomes. Statistical analysis showed a correlation above 0.6 (p < .007), and the speech-comprehension signal significantly improved prognostic accuracy beyond routine clinical measures such as the Glasgow Coma Scale and CT grading.
Routine assessments for severe TBI typically rely on behavioural observations and imaging. However, patients who remain unresponsive after sedation present a major challenge for prognosis and care planning. Other EEG-based methods have tracked imagined movements or command-following, but those approaches require an active, fairly complex mental response from the patient and can miss patients with limited cognitive resources. In contrast, passive measures of speech processing lower the bar for detecting residual cognition and extend the number of patients who might benefit from evaluation.
Lead author Dr Damian Cruse, from the University of Birmingham’s School of Psychology and Centre for Human Brain Health, states: “Our passive EEG measure provides a graded index of comprehension without requiring any behavioural response. That can reduce uncertainty during a critical period and support decisions about continuing life-sustaining treatment and directing rehabilitation resources to those most likely to benefit.”
About this TBI research news
Source: University of Birmingham
Contact: Beck Lockwood – University of Birmingham
Image: The image is in the public domain
Original Research: Closed access.
“Covert speech comprehension predicts recovery from acute unresponsive states” by Damian Cruse et al., Annals of Neurology. DOI: 10.1002/ana.25995
Abstract
Covert speech comprehension predicts recovery from acute unresponsive states
Objective
Patients with traumatic brain injury who remain unable to follow commands after sedation washout represent a major prognostic challenge. Reducing uncertainty at this stage can lead to better care decisions and more effective allocation of limited rehabilitation resources. Detecting covert cognition at the bedside—particularly residual speech comprehension—may help predict which patients are most likely to recover.
Methods
The study enrolled 28 acute TBI patients who were 2–7 days sedation-free and not obeying commands. While EEG was recorded, patients listened to streams of equally timed monosyllabic words that formed meaningful phrases and sentences. In healthy listeners, EEG synchronisation with phrase and sentence rhythms occurs only when speech is consciously comprehended. Thus, the presence and strength of this cortical tracking in unresponsive patients provide a passive measure of residual speech comprehension.
Results
Outcome data were collected for 17 patients at three months and 16 patients at six months using the Glasgow Outcome Scale Extended. Acute cortical tracking of phrases and sentences correlated strongly with later outcome (r > .6, p < .007), measured via inter-trial phase coherence. Linear regression showed that the strength of the comprehension response (beta = .603, p = .006) significantly improved prognostic accuracy beyond standard clinical variables such as the Glasgow Coma Scale and CT grade.
Interpretation
A simple, passive auditory EEG protocol substantially improves prognostic accuracy during a critical decision-making period. Because it does not require active participation, this approach is less vulnerable to cognitive limitations and can be applied to more patients, offering a practical tool to identify covert comprehension and guide early clinical choices in severe TBI care.