My ‘brain-computer’ | ForkLog
Brain–computer interfaces (BCIs) are riding an investment wave, selling a story that once belonged to sci‑fi. The pitch: connect neurons to the network to restore lost function—and, eventually, supercharge healthy minds. After reviewing research outputs, company filings and clinical milestones, here’s a clear look at where the tech stands and who is actually shaping the market.
The field at a glance
BCIs split across three paths: invasive (electrodes in brain tissue), minimally invasive (under skull or inside blood vessels without penetrating cortex), and non‑invasive (EEG and similar). Despite the Neuralink-dominated headlines, progress across all three tracks is driven by diverse, often independent teams.
Who’s really building it
Neuralink’s most direct rival is Synchron. Its Stentrode is threaded through the jugular and seated in a blood vessel near the motor cortex—avoiding open-skull surgery and moving fast through regulators. Patients in the US and Australia already use it to control Apple devices by thought, though the vessel wall reduces signal bandwidth.
Blackrock Neurotech is the veteran backbone. Its wired Utah Array has powered leading academic breakthroughs since 2004 and remains the gold standard for high-fidelity recordings. Dozens of people have lived with these implants for years, and a next‑gen commercial platform is in the works.
Precision Neuroscience—founded by former Neuralink staff—offers an ultrathin electrode sheet placed on the brain’s surface via a tiny incision, avoiding penetration of neural tissue while promising higher channel counts than typical ECoG grids.
In China, Tsinghua University’s minimally invasive NEO implant sits beneath the skull and has already enabled a fully paralysed person to control a mechanical exoskeleton.
Foundational algorithms come from the long‑running BrainGate consortium; its founder, John Donoghue, earned the Queen Elizabeth Prize for Engineering for pioneering work on thought‑driven device control.
Competition is accelerating. Science Corporation raised $230 million to develop PRIMA, a retinal implant for macular degeneration that restores visual function by stimulating surviving retinal neurons.
Engineering advances are global. Russia’s Motorica builds advanced prosthetic limbs with neural control, while researchers at HSE University’s Centre for Bioelectric Interfaces push non‑invasive decoding of motor intent. And in November 2025, Phantom Neuro launched a registry to connect upper‑limb amputees with neurointerface trials—signalling broader access for the public.
Physics, biology—and the walls they build
The brain resists foreign hardware. Glial scarring around microelectrodes degrades signal quality over time. Neuralink’s first human case underscored the risk: some implanted threads failed within weeks. Even when stable, today’s systems move only hundreds of bits per second—enough for a cursor, text entry, or simple games. The sci‑fi dream of “telepathy” would require orders of magnitude more channels.
Power and heat compound the problem. Boosting compute to raise bandwidth strains batteries and risks warming surrounding tissue. A sustained increase of even about one degree Celsius near an implant can cause irreversible damage—an engineering dead end the field hasn’t escaped.
Non‑invasive realities
EEG and related tools face skull‑induced signal distortion and noise. At a specialist meeting at HSE University, experts flagged hurdles with modern “dry” electrodes: their sensitivity varies with humidity, skull shape and how well the cap is fitted. Traditional gel electrodes provide reliable contact and accuracy—but wearing them daily is impractical. For consumer‑grade BCIs, signal fidelity remains the chief bottleneck.
From clinics to consumer ambitions
Developers frame BCIs primarily as medical devices—to restore mobility, speech and sight. That framing helps with regulators, including the FDA. Yet as the tech matures, medicine becomes a springboard: companies increasingly eye cognitive augmentation, productivity tools and new interfaces for healthy users.
BCIs also create a new, intimate layer of data. If phones map digital habits, neural interfaces could, in principle, capture unconscious affective responses. Big Tech sees dual value: a faster control surface than any keyboard—and a feedback channel to measure marketing impact neuron‑by‑neuron. Governments and defense agencies have long bankrolled neurotech: the US DARPA funds research into nonverbal battlefield communication, low‑latency drone control, and algorithms to blunt stress and pain.
Minds in the loop
BCIs aren’t simple readouts; they’re closed‑loop systems where human and machine co‑adapt. Machine‑learning decoders continuously tune to the user, while the nervous system relearns to generate signals that are easier for software to classify—a phenomenon highlighted in recent Nature Machine Intelligence work.
That coupling blurs agency. Because a third‑party decoder mediates intent, users can quickly lose track of where their will ends and the algorithm’s “fill‑in” begins. A deep sense of embodiment emerges—akin to the rubber‑hand illusion but more fundamental. Thought‑driven cursor control can be as fatiguing as a physical workout.
Meanwhile, Stanford researchers have reliably mapped cortical signals to phonemes and text—decoding elements of “inner speech” and hinting at communication without moving lips or fingers.
Privacy, policy, culture
Direct brain–machine links raise unprecedented privacy risks. Early studies already decode fragments of inner dialogue and imagery with mixed success, suggesting a future where even mental states could leak. Brainwaves act like fingerprints for emotion and mental‑health traits.
Policymakers are reacting. In the US, the proposed MIND Act aims to safeguard “digital thought,” while the Federal Trade Commission is crafting rules to block unauthorised monetisation of neural data. Cultural anxiety is surfacing too: new theatre, like The Moon is Always Full, asks whether clinical BCI trials are safe—and what happens if digitised traces of minds escape into AI training sets.
What to watch next
BCIs have left sci‑fi and entered the grind of engineering. For now, they’re complex prototypes whose near‑term mandate is unequivocal medical benefit. But the corporate race is bigger than wheelchairs: multinationals are building the next interface layer, where phones and keyboards recede. The defining question of that era isn’t just who builds the hardware—it’s who owns, protects and profits from the signals produced directly by your neurons.
