Real‑time Brain‑Controlled Hearing Aids

Researchers demonstrated a system that reads brain activity (via implanted electrodes in consenting epilepsy patients) to identify which overlapping conversation a listener is attending to and then amplifies that voice in real time using machine‑learning algorithms. In lab tests the system improved intelligibility and reduced listening effort, and the work appears in Nature Neuroscience led by Nima Mesgarani at Columbia. — If this approach is portable and non‑invasive enough, it could transform assistive hearing technology while raising privacy, consent, safety, and regulatory questions about neurodata‑driven devices.