A new approach for identifying signs of hidden awareness in people who cannot speak or move after severe brain injury has been demonstrated by researchers at the University of Bath.
The system detects patterns of brain activity through a wearable headset using an advanced application of brain–computer interface (BCI) technology.
Across multiple experimental sessions, the researchers have uncovered signs of consciousness that were previously undetected in unresponsive patients.
This represents a potential advance in diagnostic methods and rehabilitation planning for patients. It also offers promising possibilities for future technologies that may help patients communicate without the use of voice or movement.
Published in Nature Communications Medicine, the study was carried out in patients with prolonged disorders of consciousness (PDoC) and locked-in syndrome (LIS) – conditions in which awareness may be preserved but cannot be outwardly expressed.
By recording brainwaves in these patients, this BCI technology detects when a person imagines a hand or arm movement, even when no physical movement is possible. Detection accuracy can improve if neurofeedback is provided over multiple sessions.
Structured approach
The study introduced a structured multi-session approach to assessing awareness in patients, combining:
Repeated training, where participants were taught to intentionally change their brain signals over time.
Study participants were asked to imagine actions such as lifting a weight with their left hand or lifting both feet. This led to distinct patterns of brain activity that could be detected, even in the absence of any physical movement, and translated into meaningful signals.
Real-time feedback, where participants received immediate sound-based feedback, confirming to them that the system had detected the correct pattern of imagined movement.
This real-time feedback has the potential to help participants refine their mental strategies across sessions, with their brain responses becoming clearer and more consistent over time – in a manner similar to how an uninjured person learns a physical or cognitive skill. Importantly, the researchers found that repeated, structured evaluations across multiple sessions was found to improve the detection of awareness in, and engagement from, unresponsive patients.
Staged questioning, where participants progressed to an exploratory yes–no question phase (for instance, they were asked to imagine one movement to indicate ‘yes’ and a different movement to indicate ‘no’).
These questions were structured to explore different types of mental capacity and awareness.
Study outcomes
The protocol was tested in 42 participants aged between 17 and 73 years, recruited across multiple NHS and Irish clinical sites.
Key findings included:
31 of 42 participants (73.8%) showed reliable intentional modulation of brain activity – i.e. consistent patterns or rhythms in the signals – when asked to imagine specific movements.
Approximately 90% of those participants progressed to the phase of the study designed to elicit yes-no responses.
Brain responses often became more consistent across sessions.
When used alongside standard behavioural tests, the multi session BCI approach improved detection of minimal conscious state from 39% to 69%, helping identify awareness that might otherwise go unnoticed.
Why behavioural tests alone are not enough
Standard bedside assessments rely heavily on observable movement such as eye movements, reflexes and simple command following. However, when injuries are severe enough to prevent physical movement, these assessments can substantially underestimate awareness.
Previous research suggests that up to 40% of patients in a minimally conscious state may be misdiagnosed as ‘awake but unaware’ because signs of cognitive activity are missed.
Brain-based assessments offer a means to detect purposeful responses even when a person cannot express themselves behaviourally. While earlier studies have demonstrated that a single assessment could reveal covert awareness, a single session represents only a minimal assessment window.
The researchers of the new study anticipated that such signals could be trained over time through repeated assessment where feedback was also provided, potentially leading to stronger evidence that consistently points to the presence of awareness and consciousness.
Lead author Dr Naomi du Bois, a researcher at the Institute for the Augmented Human (IAH) at the University of Bath, said: “This work shows how brain based response to structured questions could complement bedside assessment and help clinicians detect hidden awareness earlier.”
Senior author Professor Damien CoyleInstitute for the Augmented Human director of the IAH and a researcher in the Department of Computer Science at Bath, said: “The novelty here is the move beyond single-session assessment. We’ve shown that a structured multi-session BCI framework, with training, feedback and staged questioning, can operate in real clinical environments or in the home or care home, and strengthen the reliability of detecting signs of awareness.
“This creates a pathway toward improved diagnosis and may ultimately support patients to interact and communicate basic responses in some cases.”
The study was sponsored by Ulster University and funded by the UK Engineering and Physical Sciences Research Council (EPSRC) through a UK Research and Innovation (UKRI) Turing AI Fellowship (2021–2025) awarded to Professor Damien Coyle (Grant EP/V025724/1).