BCIwiki (bciwiki.com) — A study published in Brain: A Journal of Neurology on October 23, 2026, reports a prospective single-centre cohort investigation by researchers affiliated with Columbia University Medical Center and NewYork-Presbyterian Hospital. It examined whether clinically unresponsive patients could control a brain-computer interface (BCI) early after acute brain injury. Participants included 41 clinically unresponsive, critically ill patients with acute brain injury, alongside convenience samples of 8 healthy volunteers and 22 conscious critically ill patients.
The self-paced, one-way BCI analyzed electroencephalography (EEG) in real time to modulate a tone's frequency and pitch. Patients alternately received instructions to attempt repeated opening and closing of their right hand to make the beeps faster, and to stop to make them slower. BCI control was achieved by 13 of 41 clinically unresponsive patients (32%), all 8 healthy volunteers, and 9 of 22 conscious critically ill patients (41%). Separately, machine learning analysis of EEG recorded during motor commands identified cognitive motor dissociation (CMD) in 8 of 41 patients (20%): preserved consciousness without the ability to communicate through motor responses.
Median signal delay for BCI control was 49.8 seconds in clinically unresponsive patients, with an interquartile range of 39.9-75.6 seconds. It was 17.1 seconds in conscious patients, with an interquartile range of 7.8-19.9 seconds, and 15.5 seconds in healthy volunteers, with an interquartile range of 12.9-21.2 seconds.
Among the 41 clinically unresponsive patients, 4/13 with BCI control (31%) recovered clinical consciousness by hospital discharge, compared with 1/28 without control (4%). The odds ratio (OR) was 21.2, with a 95% confidence interval of 2.3-930.7. Consciousness recovery at discharge was assessed using the Coma Recovery Scale-Revised. At 12 months, 6/13 patients with control (46%) recovered functional independence, compared with 2/28 without control (7%); the OR was 9.9, with a 95% confidence interval of 1.8-81.6. Functional independence was defined as a Glasgow Outcome Scale-Extended score of 4 or higher, denoting the ability to function independently for at least 8 hours.
The authors suggest that BCI control may help predict outcomes after acute brain injury and provide initial support for future bedside bidirectional communication. However, this study observed associations between one-way BCI control and recovery; it did not establish a rehabilitation effect or demonstrate bidirectional communication. All 37 respondents to a survey about potential uses and challenges strongly supported future BCI use. Those invited included conscious patients and family members of both conscious and clinically unresponsive patients.