Two Years of Telepathy

Neuralink is building brain-computer interfaces (BCI) that will return autonomy to people with unmet medical needs. Our first product, Telepathy, aims to enable people with paralysis to directly control computers, phones, and robotic limbs using their thoughts alone. Telepathy records neural activity directly from the brain regions responsible for the muscles of the hands and arms and translates those signals into digital commands, bypassing damaged neural pathways. 

Building such technology requires exploration—from understanding the information encoded in neural signals, to the intricate anatomy and physiology of the brain, to the user experience of telepathic control. Our work is driven by our partnership with the Neuralnauts: the early participants of clinical trials who step forward to pioneer BCI progress. 

With a growing crew of 21 Neuralnauts enrolled in trials worldwide, we are making exciting progress. This article highlights what we have learned from our trial participants, the milestones we have reached together, and how their participation is driving BCI technology forward.

Telepathy aims to enable people with paralysis to engage with the world as naturally as they would with their own hands. 

Initially, participants control a computer cursor by attempting to physically move their hand. Yet within minutes, they often forget about their hand, finding the cursor simply moves to their intended destination. Some participants, such as Noland, the first Neuralink recipient, have even reported the cursor arriving at the right place before they consciously realized where they wanted it to go.

Eventually, participants no longer attempt moving their hands nor their cursor – they simply move, as if the BCI was a natural extension of their body. This experience is most profound when participants go beyond cursor control on a screen and control a physical robotic arm. Nick, unable to move his limbs for four years, was able to control a robotic arm telepathically to complete basic tasks from feeding himself to even scratching an itch. What stuck with Nick, however, went beyond independence – it was recovering what he had thought was gone forever: the feeling of moving his arm. 

Beyond the subjective participant experience, we also quantify the speed and precision with which intent translates into action by measuring the information transfer rate. The more quickly and accurately a user selects a series of targets, the higher the information transfer rate and the better their control.

Using a computer mouse with their hands, able-bodied people transmit around 8-10 bits per second on average (calculate your own score here). Multiple Neuralink participants have reached and even sometimes exceeded this range. For instance, Nick achieved over 10 BPS within his first week of using the BCI. 

While performance metrics like Webgrid quantify restored ability, the ultimate test for regaining independence is found in how it changes people’s lives. In some cases, that test is literal. 

Ten years ago, Noland sustained a paralyzing spinal cord injury just before his final year of college. Forced to leave his degree unfinished and return home, Noland transitioned from an independent, ambitious college student to depending on others for everyday tasks, leaving him with a diminishing sense of purpose. 

With Telepathy, self-improvement became his new motivation. Being able to control a computer with full independence, Noland began waking up early, studying math, reading, and even learning new languages. Ten years after leaving college, Noland has returned to college to pursue a degree in neuroscience.

With Noland paving the way, more participants are now using Telepathy for education. Sebastian, one of our newest Neuralnauts, is a 23-year-old medical student who sustained a spinal cord injury two years ago during winter break. Before Neuralink, Sebastian relied on voice commands to operate his computer, making his return to school difficult. 

With Telepathy, Sebastian has drastically increased his productivity as he pursues his medical degree. From annotating research papers to completing interactive assignments to discreetly multitasking during lecture, Sebastian is using his Neuralink for up to 17 hours a day. 

Neuralnauts are also using their restored cursor control to create. Audrey, our first female Neuralnaut, sustained a spinal cord injury twenty years ago. For nearly two decades, Audrey has not directly controlled a computer, relying on her partner to complete everyday tasks. 

Despite her limited experience with computers, Audrey mastered Telepathy and found a new love for making art. Using Telepathy to make intricate pieces, she visually conveys her story through abstract art. After gaining recognition online, she aims to open a physical gallery to further showcase her work and inspire others.

What started as a creative outlet has transformed into an act of sovereignty. Through creating her own original art on her own time, Audrey has been able to create something that feels truly hers. 

From education to creative expression, Neuralink participants are discovering ways to restore experiences that seemed lost forever. But for some, it is returning something even more foundational: the ability to talk. 

Amyotrophic lateral sclerosis (ALS) is a devastating disease. People with ALS gradually lose control of nearly every muscle in their body, ultimately leading to paralysis of their entire body. Once they lose the ability to breathe and speak, up to 95% of people with ALS decline life-sustaining ventilation, in part due to their inability to talk to loved ones. 

To restore the ability for people with ALS to meaningfully interact with others, we are building faster communication systems by finding clever ways to translate neural data into text. After working with several of our Neuralnauts, such as Jake, we discovered that despite the Neuralink being implanted on just one side of the brain, we are still able to receive strong signals from both hands.

Building off this finding, we have been exploring ways to create a ten finger keyboard for the mind. By mapping their ten fingers to different letters analogous to a physical keyboard, participants have reached typing speeds of up to 40 words per minute. 

Jake, who is able to speak, has used this newfound typing ability to support his general computer use. Translating these findings, we aim to give nonverbal people with late-stage ALS the ability to converse faster and more effortlessly.

Looking forward, we aim to push communication to conversational speeds of 140 words per minute through our recently launched clinical trial called VOICE. Reading signals from brain regions involved with speech production, the goal of this study is to restore real-time speech for people with severe speech impairment caused by neurological conditions such as  ALS or stroke. 

Going beyond communication speed, our Neuralnauts have devised clever ways to be part of the room again. Brad, our first participant with ALS, noted that a major frustration of his paralysis was being unable to look around the room. When attending his son’s regional robotics competition, for instance, Brad was unable to see his son compete due to his inability to move his neck. Finding a lightweight camera that swivels 360 degrees, Brad attached it to his wheelchair and now moves it freely with his telepathic cursor, looking around at will.

With our Neuralnauts pushing the frontiers of BCI technology, the next step is to make those capabilities accessible to everyone that needs it. Every brain implant must adapt to individual neural architecture influenced by factors as subtle as fluctuations in brain motion, disease state, skull thickness, and blood vessels in the brain. 

A primary aim of our expanding clinical trials is to better understand these variations and improve both our hardware and the overall procedure for every participant. Each participant adds to our growing understanding, advancing the path for the next. 

For instance, when Noland became the first person to receive a Neuralink, we gained crucial insights into implant thread retraction, highlighting surgical and postoperative improvements needed to maintain consistent signal quality. With these learnings, we observe a general trend of signal improvement across participants. Notably, after implementing these mitigations, we have since seen higher signal quality across 18 of the subsequent 20 participants. Even with this improved signal quality, however, we still see variance in BCI performance, which appears to be correlated with both anatomical variability such as intracranial spacing, and condition, such as stage of their ALS. 

To make fluid, hand-like control accessible to everyone, regardless of anatomy or neurological condition, we must continue to improve our implant and surgical process. In the coming months, to improve the implant we are (1) increasing the volume of neural signal received by increasing from 1000 to 3000 electrodes and (2) preserving this signal over time by exploring mechanical features to improve thread retention. For the surgical process, we are investigating inserting implant threads directly through the dura mater surrounding the brain, reducing invasiveness of the procedure. 

Working closely with regulatory bodies and hospital sites, we aim to safely iterate on our hardware to deliver improved devices to participants and maintain our current record of zero serious device-related adverse events. This dedication to safety provides the foundation for the accelerating expansion of our Neuralnaut crew, from three participants in 2024 to multiple participants per month in 2025. 

Looking forward, the path is clear: to build a brain interface that restores independence at scale, we depend on the pioneers willing to lead. For that reason, we thank the Neuralnauts for the progress we made so far and for the progress still to come in 2026. 

If you or someone you know has quadriplegia due to a spinal cord injury, ALS, or brainstem stroke, consider joining the Neuralink Patient Registry. 

If you are interested in joining us in solving these problems, apply to open roles.