Neuralink Human Trials Timeline 2024–2026: Patients, Telepathy, Blindsight & Global Expansion

Ten years ago, Noland Arbaugh sustained a paralyzing spinal-cord injury just before his final year of college, leaving him unable to move his arms or legs. In January 2024, he became the first person implanted with Neuralink’s N1 device, in a study called PRIME (Precise Robotically Implanted Brain-Computer Interface). The initial goal was narrow and concrete: let Noland control a computer cursor and digital devices using decoded neural activity, without moving his hands.

What actually happens is not “reading his mind.” When Noland thinks about moving his hand, his motor cortex produces electrical activity. N1’s electrodes record that activity, software decodes the pattern associated with intended movement, and a computer converts it into cursor motion. Neuralink is decoding specific, trained neural patterns for a defined task — not reading arbitrary private thoughts.

After Noland’s surgery, some of the implant’s threads retracted from brain tissue — a physical, mechanical problem, not a software bug. Fewer effective threads meant fewer usable electrode signals, and his initial BCI performance declined. Neuralink did not hide this: the company disclosed it publicly and responded by modifying its decoding algorithms and how signals were interpreted. Performance subsequently recovered, and on some measures improved beyond earlier levels.

Software did not physically reinsert the threads. It extracted more useful information from the signal that remained. That distinction matters: the first Neuralink problem was physical; part of the solution was software, working around reduced hardware, not fixing it.

In July 2024, a second PRIME Study participant — referred to publicly as Alex, living with a spinal-cord injury — received his implant at the Barrow Neurological Institute, going home the following day. Neuralink published a progress update on 21 August 2024. By his second day of use, Alex was using Autodesk Fusion 360 CAD software to design a mount for his own Neuralink charger, which was 3D-printed and attached to his laptop stand. He also returned to playing Counter-Strike 2, a game he had previously played using a mouth-operated controller.

CAD use matters because it moves BCI demonstrations beyond simple cursor tests toward genuinely productive computer work. Neuralink reported that Alex had no thread-retraction issue at the time of its update — that is a snapshot from August 2024, not a claim of permanent, all-patient stability.

Brad Smith received his implant in November 2024, becoming Neuralink’s third participant and its first with amyotrophic lateral sclerosis (ALS) — and its first non-verbal recipient, relying on a ventilator and able to move only his eyes before the implant. Using Neuralink, Brad began communicating with an AI-generated voice and controlling a swiveling camera mounted to his wheelchair, letting him look around a room freely for the first time in years. By the end of 2024, Neuralink had implanted three participants — the baseline the company’s own 2026 retrospective measures growth against.

Starting in November 2024, Neuralink’s clinical research expanded beyond the United States. Health Canada approved CAN-PRIME on 20 November 2024, Neuralink’s first international study, based at Toronto Western Hospital (University Health Network), targeting people with limited or no hand use due to ALS or cervical spinal-cord injury.

Neuralink announced UAE-PRIME on 14 May 2025, its first Middle East clinical trial, at Cleveland Clinic Abu Dhabi in partnership with the Department of Health Abu Dhabi — evaluating the N1 implant and R1 robot in adults with severe motor and speech impairment from spinal-cord injury, ALS, stroke or other qualifying neurological conditions.

Neuralink launched GB-PRIME on 31 July 2025 across two sites, University College London Hospitals (UCLH) and Newcastle Hospitals, approved by the UK’s MHRA, HRA and Health and Care Research Wales. The study was designed to enroll up to seven participants; UCLH reported that all seven had received surgery between October and December 2025.

Neuralink’s own 28 January 2026 update published an implantation timeline covering every clinical-trial surgery across 2024 and 2025: one in the first half of 2024, two in the second half of 2024, four in the first half of 2025, and thirteen in the second half of 2025 — a cumulative 20 surgeries through the end of 2025. 🟢

On 28 January 2026, exactly two years after Noland’s implant, Neuralink published a retrospective titled “Two Years of Telepathy,” announcing a growing crew of 21 Neuralnauts enrolled in trials worldwide. 🟢 The company said it maintained its record of zero serious device-related adverse events at that point — a company-reported figure worth taking seriously, but one that a small, closely monitored early-feasibility cohort cannot extend into a claim of long-term, population-scale safety.

The update also reported: after implementing surgical and postoperative mitigations learned from Noland’s thread retraction, Neuralink observed higher signal quality in 18 of the subsequent 20 participants. 🟢 The company said it plans to increase electrode count from roughly 1,000 to 3,000 electrodes in future hardware, and is investigating inserting threads directly through the dura mater to reduce surgical invasiveness — both labelled as planned, not yet deployed. 🟠

Beyond the cursor: robotic arm, education, art, communication

Nick, unable to move his limbs for four years, used Neuralink to telepathically control an assistive robotic arm to feed himself and scratch an itch — a documented step beyond screen-based cursor control into physical-world assistance. Noland returned to college to study neuroscience. Sebastian, a 23-year-old medical student, uses his implant for up to 17 hours a day for coursework. Audrey, Neuralink’s first female participant, uses her implant to create original digital art after two decades without directly controlling a computer.

For communication, Neuralink reported that despite the implant being placed on only one side of the brain, it can still receive signals related to both hands from some participants, enabling a “ten-finger keyboard for the mind” — participant Jake, who has ALS, reached typing speeds of up to 40 words per minute this way. Building on that, Neuralink has launched a dedicated clinical trial called VOICE (ClinicalTrials.gov NCT07224256), aiming to decode speech-related neural activity directly, with a longer-term goal of conversational speeds around 140 words per minute for people with severe speech impairment from ALS or stroke. 🟠

By mid-2026, public reporting and participant tracking indicated Neuralink had implanted at least 26 participants — up from the 21 the company itself announced on 28 January 2026. 🟡 One widely circulated tracking account identifies patient 26 as Lee Marten, a Vancouver police sergeant implanted on 20 May 2026, reported to be the first participant to receive Neuralink’s newer through-the-dura surgical approach. This account has not been confirmed directly by a Neuralink press update as of this writing, so it is presented here as community/media tracking, not an official company total.

Do not treat “26” as a Neuralink-confirmed figure unless the company itself publishes a matching total. The most defensible framing: Neuralink’s own last confirmed count was 21 (28 January 2026); public tracking suggests continued enrollment brought the total to at least 26 by mid-2026; replace this figure the moment a newer official Neuralink total is published.

Neuralink’s N1 implant is fully implanted, wireless and rechargeable, originally specified with 1,024 electrodes distributed across 64 flexible threads — each thinner than a human hair. (As above, Neuralink said in January 2026 that it plans to increase this to 3,000 electrodes in future hardware; that upgrade had not shipped as of September 2026.)

Those threads are too fine and flexible to place manually with the precision required, so Neuralink built the R1 surgical robot specifically to insert them — avoiding visible blood vessels and targeting precise cortical locations repeatably. R1 does not perform the whole surgery. A neurosurgeon still carries out exposure, craniectomy, implant mounting and closure; R1’s role is the precise thread insertion step within that procedure. Surgery is not fully automated, and Neuralink has not demonstrated it is routinely performed as outpatient care.

On 17 September 2024, the FDA granted Blindsight Breakthrough Device Designation — not early 2026, and not FDA approval for commercial use. Breakthrough status can provide a more interactive, expedited FDA development and review pathway for a qualifying device. It does not mean the device is FDA-approved, does not mean it has been proven effective in humans, and does not mean anyone’s blindness has been treated or cured.

Blindsight’s concept: bypass damaged eyes or optic nerves entirely by implanting a microelectrode array into the visual cortex and electrically stimulating it to create perceived visual patterns — the opposite data direction from Telepathy. Telepathy mainly reads neural activity out of the brain; Blindsight would write electrical stimulation into it to create an artificial sensory perception. Decoding movement intentions and creating useful visual perception are different neuroscience problems — success with one does not automatically prove the other.

Elon Musk has said Blindsight could eventually restore vision even to people blind since birth, and that high-resolution or extended-spectrum (infrared/ultraviolet) vision might eventually be possible. These are Musk/company goals and predictions, not independently demonstrated outcomes. In early August 2026, Musk said Neuralink expected to begin its first human Blindsight implants within roughly 6 to 12 months — meaning Blindsight remained an upcoming, not active, human trial as of September 2026.

Neuralink’s implant surgery carries the same category of risks as any neurosurgical procedure: infection, bleeding, tissue response, device failure and signal degradation over time, alongside the thread-retraction pattern already seen in Noland’s case. Neuralink reported maintaining zero serious device-related adverse events as of its January 2026 update — a genuinely positive early signal, but one drawn from a small, closely monitored cohort. It does not establish population-scale, long-term safety, and should not be read as “Neuralink has proven its device is safe” in any broader sense.

The most important unanswered scientific question may be durability: can electrodes remain safe, stable and useful for years or decades? Human Neuralink implant history began in January 2024 — under two years old as of this update — so long-term durability is genuinely unknown, not merely under-discussed.

Who owns your brain data?

A connected implantable medical device raises real questions about neural-data privacy, cybersecurity, informed consent, long-term device support, what happens if the company changes hands or shuts down, software updates and data portability. These are legitimate open questions for BCI technology generally, not evidence of any specific breach or vulnerability at Neuralink. This article does not describe attack methods and makes no claim that anyone’s neural data has been compromised.

Not for general commercial sale or use. Neuralink’s PRIME study operates under FDA investigational human-study authorization; Blindsight holds Breakthrough Device Designation. Neither of those is the same as FDA premarket approval for routine clinical use. In plain terms: the FDA has allowed Neuralink to test its device in humans under controlled trial conditions — it has not approved Neuralink as a commercially available medical treatment, and there is no path today for a member of the public to simply buy one.

Neuralink did not invent brain-computer interfaces. Other research groups and companies — including Synchron, Precision Neuroscience and Paradromics — have implanted BCIs in humans and demonstrated cursor control, communication or movement-related capabilities using different hardware approaches. Neuralink’s distinguishing features are its high electrode density, its fully implanted wireless design, and its purpose-built surgical robot — not sole ownership of the underlying science.

For most people, moving a computer cursor is so ordinary it barely registers as an action. For someone who cannot reliably move their hands because of a spinal-cord injury or ALS, the same action can represent something much larger: independence. That is why Neuralink’s most consequential demonstrations are not necessarily its most futuristic ones. A participant playing a game is interesting. A participant studying for a medical degree, making art, or looking around a room on their own for the first time in years is arguably more consequential.

The first Neuralink implant also showed why early human trials matter in the first place. The device worked. Then some threads moved. Performance declined. Engineers changed the software. Performance recovered. That sequence is less spectacular than a science-fiction headline, but it may be more important — experimental medicine advances by discovering what works, what fails, and whether the failure can be understood and addressed safely.