The "real" AI: Sima Mofakham, Chuck Mikell's AI that sees signs of consciousness (tbrnewsmedia.com)

Stony Brook neurosurgeons Sima Mofakham and Chuck Mikell announced SeeMe, an AI that detects subtle, purposeful facial movements indicative of covert consciousness in patients with acute brain injury. In a 37‑patient study published in Nature Communications Medicine, SeeMe—using a simple camera and open‑source software—quantified microscopic changes in facial expression before and after verbal cues, compared each patient against their own baseline over days, and flagged trajectories of improvement up to four days earlier than bedside clinicians. The tool is noninvasive, inexpensive and built to be scalable for resource‑limited settings. SeeMe’s significance is twofold: it provides an objective, continuous behavioral signal that can reveal missed consciousness (echoing recent work that 15–25% of apparently unresponsive ICU patients have higher brain function) and it could guide earlier, personalized treatment and family counseling. The team emphasizes SeeMe as a supplemental “magnifying glass” alongside EEG and imaging rather than a lone decision maker. They’ve built in safeguards (temporal/spatial localization and reverse‑engineering of cues) and designed the model to be sensitive to weak signals; planned NIH‑funded trials aim to refine sensitivity/specificity, compare detections with standardized exams and neurophysiology, develop a deployable app, and explore pairing SeeMe with vagus nerve stimulation to amplify therapeutic responses.