Two neurosurgeons stood over a patient who hadn’t spoken clearly in years. Their hands moved with practiced precision, placing the final bit of hardware—a silver coin-sized transceiver beneath her sternum—before sealing the incision. This wasn’t just another implant procedure.
It was the first time in human history that a fully wireless, fully implantable brain-computer interface (BCI) had been permanently placed to restore speech. No wires piercing the skull, no tether to a lab computer rig, no long commutes to a university facility just to type three words. The system was live.
The patient? A Michigan woman living with advanced amyotrophic lateral sclerosis (ALS), a motor neuron disease that was slowly robbing her of movement, then muscle control, and finally the simple, human joy of holding a conversation.
The device? Paradromics’ Connexus BCI—a 421-electrode cortical array nestled directly on the brain’s motor cortex, feeding neural data to that chest transceiver, which then broadcasts decoded thoughts as synthesized text and speech over Bluetooth.
Matthew Willsey, M.D., Ph.D., who led the surgery at Michigan Medicine, put it bluntly: “This has the potential to be a major step forward as we work toward our goal of helping treat people with paralysis who otherwise lack efficient and effective therapies for preserving communication.”
Here’s what actually happened in that OR, why this moment matters more than any lab demo ever could, and what six years of follow-up will reveal about real-world reliability, safety, and whether this device can truly restore natural communication.
Why the Wireless Architecture Changes Everything
Let’s rewind—just a bit. If you’ve read any BCI headlines over the past decade, you’re probably picturing someone with a big, wired electrode plug coming out of their head, tethered to a giant computer in a university lab. That’s mostly accurate: until now.
Early intracortical BCIs required a physical wire passing through a permanent skull opening, creating two relentless problems: infection risk and usage limitations. Every time the patient wanted to communicate, they had to travel to a facility, plug in, and hope the system didn’t misfire. That wasn’t just inconvenient—it was dignity-stripping.
The Connexus BCI flips that script. When the patient imagines speaking, her brain’s motor cortex still fires electrical signals along the neural pathways originally dedicated to moving her tongue, lips, and vocal cords—even if those muscles no longer function. The 421 microelectrodes on the cortical array capture these precise patterns of electrical firing.
Those signals travel wirelessly to a small, implantable transceiver in her chest, which then broadcasts the data to an external receiver (like a phone or tablet). Advanced machine-learning algorithms decode the neural rhythms in real time, mapping specific patterns to distinct words or characters.
Matt Angle, Ph.D., CEO and founder of Paradromics, captured the stakes: “Our goal is to restore natural communication for people who have lost the ability to speak and help them stay connected with their loved ones.”
That “connected” part is crucial. In ALS, communicative isolation often precedes depression and institutionalization. Restoring speech isn’t just about convenience; it’s about autonomy, and the ability to say no when needed, to share a quiet thought at bedtime, or simply tell someone you love them without typing it character by character on an eye-tracking system.
The Connect-One Study: Safety First, Speech Second
This procedure wasn’t part of a mass rollout. It marked the launch of the Connect-One Early Feasibility Study (EFS), a national clinical trial operating under an FDA Investigational Device Exemption (IDE) granted in November 2025.
Three sites are enrolling participants: University of Michigan, UC Davis (where David M. Brandman, M.D., Ph.D., leads the study as Lead PI), and a third yet-to-be-confirmed site. All focus on adults with profound communication deficits due to motor neuron diseases—primarily ALS and primary lateral sclerosis (PLS).
Why an Early Feasibility Study? Because the primary, absolute goal right now is proving long-term safety, bio-stability, and recording integrity of the hardware inside the human body. If the implant holds up over six years—the trial’s planned follow-up window—then efficacy studies can expand.
Dr. Willsey explains the immediate priority: “We are incredibly excited to investigate the potential of this wireless BCI to restore communication for people who have lost the ability to speak due to neurological disease or injury.”
That six-year window is telling. Previous BCIs failed because they degraded inside the brain after months or a year or two. This study will track not just whether patients can talk, but whether the electrodes stay sensitive enough to pick up clear signals year after year, and whether the chest transceiver continues to function without failure.
Behind the 421 Microelectrodes: Density and Decoding
The cortex doesn’t send clean, labeled packets of text when we think about speaking. It sends electrical storms—firing neurons, oscillating rhythms, and complex spatiotemporal patterns across hundreds of thousands of cells.
The Connexus array packs 421 microelectrodes onto a tiny, flexible cortical implant. That density matters. More electrodes mean more data points, better signal-to-noise ratios, and the ability to disambiguate subtle speech motor patterns that older arrays—often limited to 32, 64, or maybe 96 electrodes—simply couldn’t resolve.
Think of it like upgrading from a flip phone to a smartphone: you’re not just getting more pixels; you’re gaining the ability to record and replay full conversations, not just the occasional “yes” or “no.”
Once data flows from the array to the chest transceiver, machine-learning models take over. These aren’t rigid rule-based systems. They’re deep neural networks trained on hours of prior speech motor recordings—often gathered from epilepsy patients undergoing presurgical monitoring, whose brains are already exposed for electrode placement. In that June 2025 temporary implant, Willsey and Oren Sagher, M.D., Director of Functional Neurosurgery at U-M Health, confirmed the device could safely reside in the brain and reliably record signals.
Now, with permanent placement and continuous data collection, those models will be refined in real time. The patient’s own brain activity becomes the training ground for better decoding algorithms.
Dr. David Brandman from UC Davis puts it this way: “The Paradromics device has incorporated decades of learning from intracortical BCI research, and this study represents the next big step to investigate whether a fully implanted and wireless BCI can restore communication.”
A New Clinic, a New Mindset—And the ALS Care Ecosystem
Michigan Medicine didn’t start this effort overnight. Dr. Willsey launched the Brain-Computer Interface Clinic in 2025, a dedicated clinic for patients seeking access to emerging neuromodulatory therapies. It’s not just surgery; it’s longitudinal care, psychological support, and device training—all wrapped into one program.
That clinical infrastructure fed directly into the Connect-One trial. When your patient needs daily motor neuron disease management plus BCI evaluation, they can’t juggle appointments across three different institutions. Michigan Medicine keeps everything under one roof.
Dr. Stephen Goutman, M.D., M.S., Director of the Stanford Morris ALS Clinic and Associate Director of the Scott Pranger ALS Center, spoke to this integration: “It is critical to preserve communication for all those living with motor neuron disease… [it] preserves independence and quality of life.”
That independence piece can’t be overstated. Most assistive communication devices require caretaker setup, calibration, or frequent recalibration due to signal drift. A wireless BCI that runs at home—without wires, without daily cable management—changes the time burden from hours to minutes.
And let’s not forget: this isn’t just about ALS. Primary lateral sclerosis (PLS), progressive supranuclear palsy (PSP), spinal muscular atrophy (SMA), and even high-level spinal cord injuries all share one painful similarity—the loss of expressive speech. If Connexus succeeds in ALS, it could become the template for restoring voice in dozens of other conditions.
The Road Ahead: Six Years, One Patient at a Time
This first participant won’t be the last—but she will be the most closely watched. Every six months for the next half-decade, her implant’s signal quality, battery health (if applicable), and biocompatibility markers will be assessed. She’ll also undergo regular cognitive and communication assessments, quality-of-life measures, and device usability trials.
Even in the short term, success won’t look like flawless speech on Day 30. Expect slow, precise decoding—first characters, then short phrases, gradually accelerating as the algorithm adapts to her unique neural patterns. The real win is sustainability: if Day 12-month performance exceeds Day 6 months, we’re on the right track.
Looking out further, Paradromics hasn’t filed for an Investigational New Drug (IND) or premarket approval (PMA) yet. But with this early feasibility data, they’ll be preparing regulatory submissions within 2–3 years, paving the way for multi-center efficacy trials and eventual commercial availability.
For now, though, this is the moment that counts: a woman in Michigan sits up after surgery, looks at her tablet, thinks “Thank you,” and watches those words appear in clear, synthesized text on the screen. No wires. No help needed. Just her brain, a silver transceiver in her chest, and the quiet hum of progress.
The wireless future of neuroprosthetics didn’t arrive with a bang. It arrived in a hospital room, under quiet fluorescent lights, and it came to life—wirelessly, seamlessly, and forever changed the game for people living without a voice.