Three Years of Daily BCI Use Gives the Field Its First Real Endurance Data

10. Three Years of Daily BCI Use Gives the Field Its First Real Endurance Data

Casey Harrell, a man living with ALS, has used a surgically implanted brain-computer interface to produce speech for nearly three years, logging thousands of hours of active use since his first decoded sentences in 2023. The device, developed by researchers at UC San Francisco and UC Berkeley, reads motor cortex signals and translates intended speech into audible output. Harrell now uses the system independently at home, without a research team present, making him what the project team calls "the first power user" of a speech BCI in a clinical context.

That framing matters competitively. Neuralink's N1 implant and Synchron's Stentrode have both generated strong short-window demonstrations, but neither company has published sustained, years-long daily-use data from a single patient. Harrell's record is the closest thing the field has to an endurance benchmark. It shifts the question from "can a BCI decode speech" to "does a BCI hold up under real-world daily load." That is a harder question, and right now only one data point answers it. For Neuralink, which has implanted multiple patients but released limited longitudinal outcome data, this is a quiet pressure point. Synchron's endovascular approach avoids open-brain surgery but has not yet shown comparable communication throughput over comparable time.

The broader pattern is a field moving from proof-of-concept to durability trials, whether anyone has formally declared that transition or not. The next competitive signal to watch is whether Neuralink publishes multi-year outcome data from its 2024 cohort, and whether any team demonstrates signal stability across electrode sites over a similar three-year window. Harrell's case also raises the regulatory bar: if FDA reviewers now have a three-year daily-use reference point, shorter trial windows will face harder questions about what long-term safety and performance actually look like.

Source: MIT Technology Review