A speech implant drops its four-hour rule: reading Connect-One's first months.

On October 26, Paradromics said the FDA had approved an amendment to its Connect-One protocol that removes one line from the eligibility criteria: prospective participants no longer have to live within a four-hour drive of a study site. It sounds administrative. For a trial recruiting people with severe paralysis, it decides who can be considered at all. Candidates still have to travel to one of three sites, at UC Davis Health, Mass General Brigham and University of Michigan Health, and every other criterion stands. The company says it is still working out how to support people who travel long distances for screening, visits and follow-up.

The study itself moved quickly. The FDA authorized it under an Investigational Device Exemption in November 2025, and the first long-term implantation of the fully implantable, wireless Connexus device took place at University of Michigan Health in June 2026. Paradromics says that first participant now uses the system to communicate through synthesized speech, send text messages and make phone calls. In August the company reported a further FDA letter widening the personal computing devices participants may use.

Those details matter because everyday use is the point of the device. A speech decoder that works in a lab session is a research result. One that sends a text message from home is closer to a product. But the evidence is thin in the plain statistical sense: the releases describe one participant, the claims come from the company, and we found no peer-reviewed results from the study so far.

The durability question has a similar shape. On October 8, Paradromics reported more than four years of stable recording in two sheep implanted with cortical modules representative of the Connexus electrodes, with stable spike signal-to-noise ratio and no observed array migration. A separate study of the full platform in six sheep has held stable for nearly two years. If that transfers to people, it addresses the field's oldest worry, since implantable BCIs have commonly been assumed to trade performance for long-term stability. It remains a preclinical result in a small number of animals, announced by the company that makes the device.

A quieter signal came on October 19, when Paradromics and Siemens Healthineers announced a collaboration on multimodal imaging for BCI care, covering candidate assessment, neural targeting, surgical planning, post-implant localization and outcomes. Nobody builds that workflow for a one-off experiment. It suggests the company is planning for the routine clinical pathway that a device would need if it ever reached the market.

That market does not exist yet. Connexus is limited by US law to investigational use, and Synchron's own site states that its BCI system is not approved for commercial use in any geography. Three things would change our read of Connect-One: a second and third participant with comparable results, peer-reviewed performance data rather than press releases, and human stability measured over years rather than months. The four-hour rule was one barrier to getting those answers, and it is now gone.