New AI Tool Detects Consciousness in Brain Injury Patients Earlier Than Traditional Methods
A new artificial intelligence (AI) tool, called SeeMe, developed by researchers at Stony Brook University, is demonstrating the ability to detect signs of consciousness in patients with acute brain injury and coma days before clinicians can observe physical responses. The findings, published in Nature Communications Medicine, offer a potential breakthrough in diagnosing and predicting recovery outcomes for individuals with severe brain trauma.
SeeMe bridges the gap between a patient’s internal awareness and observable physical signs. According to Dr. Romina Mofakham, senior author of the study and associate professor of neurosurgery and electrical and computer engineering at Stony Brook, “Just because someone can’t move their limbs or speak doesn’t mean thay aren’t conscious. Our tool uncovers those hidden physical efforts by patients to show they are conscious.”
The system utilizes high-resolution video and computer vision to analyze involuntary facial reactions to simple verbal commands, such as “open your eyes” or “show me a smile.” These subtle responses, frequently enough imperceptible to the human eye, are then processed using machine learning algorithms.
A clinical study involving 37 patients with acute brain injury revealed that SeeMe detected purposeful movement an average of four days before the clinical care team recognized physical movements. Furthermore,patients who showed early responses detected by SeeMe were significantly more likely to regain consciousness and experience improved functional outcomes upon hospital discharge.
“This is not just a new diagnostic tool,it’s a potential prognostic marker,” explains Dr. David Mikell, neurosurgeon and co-lead investigator. He emphasizes the value of data-driven insights for families seeking information about their loved ones’ recovery prospects, stating, “Families often ask us how long it will take for a loved one to wake up, or if they ever will. This study helps us answer those questions with more confidence, grounded in data, not just experience or instinct.”
The researchers highlight the ethical implications of accurately identifying consciousness in TBI patients, noting that misdiagnosis can lead to inappropriate medical decisions, including the withdrawal of care or limited access to crucial neurorehabilitation.
SeeMe offers a practical solution to the need for objective assessment tools. It is noninvasive, inexpensive, and scalable, requiring only a standard camera and open-source software – making it accessible even in resource-limited healthcare settings.
Looking ahead, the team plans to conduct larger clinical trials and pursue regulatory approval, with the goal of integrating SeeMe into standard ICU practice alongside existing monitoring methods like EEG. They envision a multi-modal platform that leverages AI to provide a more comprehensive understanding of a patient’s neurological state,ultimately allowing for personalized care and optimized rehabilitation. Mofakham concludes that SeeMe exemplifies how AI can empower patients, enabling them to “speak without words” and regain independence.
The research was funded by institutional seed grants supporting collaboration between the neurosurgery and electrical and computer engineering departments at Stony Brook University.
