🧠💸 Tracking the Money: Who’s Funding Subvocal and Neural Interface Tech (2005–Present)

Targeted Individuals (TIs) frequently ask, “Where’s this tech coming from, and who’s behind it?” Let’s follow the money trail clearly:

🇺🇸 U.S. Government & Military Agencies

• 🔬 DARPA:
  • 2009: Launched Project Silent Talk ($4M) to detect and interpret “pre-speech” brain signals.
  • 2017: Started Neural Engineering System Design (NESD) ($65M) to create implants capable of interfacing with up to a million neurons.
  • 2018: Initiated Next-Generation Nonsurgical Neurotechnology (N³), funding six teams (Battelle, Carnegie Mellon, Johns Hopkins, PARC, Rice University, Teledyne) to develop wearable neural interfaces for military applications.
• 🚀 NASA Ames (2004): Pioneered subvocal speech recognition using throat sensors, achieving ~92% accuracy in silent communication.
• 🏛️ NIH & BRAIN Initiative (Obama administration): Invested billions in neural interface research, supporting non-invasive BCI research (e.g., Synchron’s “Stentrode” trials).

🏫 Academic Institutions

• 🎓 University of California Irvine (2008): $4M from U.S. Army for “synthetic telepathy” research using EEG.
• 🎓 Stanford & Brown University (BrainGate):
  • 2004: Enabled first neural-controlled computer cursor for paralyzed patients.
  • 2021: Achieved a neural handwriting system decoding brain signals at 90 characters per minute.
• 🎓 UC San Francisco:
  • 2021: Created a speech neuroprosthetic translating brain signals into complete sentences.
  • Partnered with Facebook Reality Labs.
• 🎓 MIT (AlterEgo):
  • 2018: Developed non-invasive device recognizing internal speech via facial neuromuscular signals with 92% accuracy.

💼 Private Companies & Startups

• 🚩 Neuralink (Elon Musk, 2016): Raised over $300M for brain implants enabling telepathic communication. Received FDA clearance for human trials in 2023.
• 🚩 Synchron:
  • Developed “Stentrode” implant inserted through blood vessels; funded by DARPA, private investors like Bill Gates and Jeff Bezos.
  • Leads global clinical trials in practical BCI communication.
• 🚩 Paradromics:
  • Emerged from DARPA’s NESD ($18M funding), focuses on decoding speech from neural signals.
• 🚩 Meta (Facebook) Reality Labs:
  • Invested in UCSF research (2017–2021) on brain-to-text interfaces.
  • Acquired CTRL-Labs for $500M, focusing on peripheral neural interfaces for AR/VR.

🌍 International Efforts

• 🇪🇺 EU BrainCom Project (€8.35M, 2017–2022): Developing cortical implants to restore speech in paralysis patients.
• 🇨🇳 China Brain Project (launched 2021): Significant investments in neural interfaces and “brain-control” technologies for both civilian and military use.
• 🇨🇳 CIBR/NeuCyber: Developing the semi-invasive “Beinao No.1” BCI implant, aggressively scaling up human trials.

🤔 Why This Matters for TIs

Understanding who finances and develops these technologies clarifies how sophisticated neural interfaces and silent communication systems became a reality. These projects, initially military-driven, have expanded globally into medical and commercial sectors. While many programs have legitimate medical aims, their dual-use potential means they can be (and likely are) repurposed covertly, raising significant ethical concerns around surveillance, privacy, and cognitive liberty.

Stay informed and vigilant. The more we know, the better we can advocate for transparency, oversight, and the protection of our cognitive freedoms.

The Global Push for Subvocal and Neural Interfaces (2005–Present)

U.S. Government Initiatives and Funding

Public agencies in the United States have heavily invested in subvocal communication and brain-computer interface (BCI) research since the mid-2000s. DARPA (Defense Advanced Research Projects Agency) has led many groundbreaking programs. In 2008, the U.S. Army (via the Army Research Office) awarded a $4 million MURI grant to a University of California Irvine team to explore “synthetic telepathy,” aiming for nonverbal soldier-to-soldier communication​news.uci.edu​latimes.com. This effort, led by cognitive scientist Michael D’Zmura, investigated using noninvasive EEG to decode imagined speech for battlefield use and for patients unable to speak​news.uci.edu​latimes.com. Building on that, DARPA launched Project Silent Talk in 2009 with $4 million in funding, seeking to detect “pre-speech” neural signals and transmit them as communication​wired.com​wired.com. The goal was to map EEG patterns to words, test if such brain signals generalize across people, and build a prototype system for covert communication over short ranges​wired.com​wired.com. This vision – essentially “synthetic telepathy” – was at the cutting edge of DARPA’s 2010 budget and marked one of the earliest major subvocal BCI programs​wired.com. 

Throughout the 2010s, DARPA continued to spearhead high-risk neurotechnology projects. Under the Obama administration’s BRAIN Initiative (launched 2013), DARPA joined NIH, NSF, and IARPA in a coordinated funding effort to advance neurotechnologies. The BRAIN Initiative provided $3.5 billion for neuroscience/BCI research from 2014–2024​gao.gov​gao.gov. DARPA’s contributions focused on high-bandwidth neural interfaces: for example, the Neural Engineering System Design (NESD) program (2017) awarded $65M across six teams to develop implantable interfaces that could interact with up to one million neurons​techcrunch.com​techcrunch.com. One team at Brown University worked on a “neurograin” wireless implant network to decode speech-related brain activity​techcrunch.com, while startup Paradromics (the sole small company funded) received about $18M to develop a prosthetic device for decoding and interpreting speech from brain signals​techcrunch.com. DARPA also emphasized two-way communication (writing back to the brain), pushing toward implants no larger than “two nickels stacked” that could both record and stimulate neurons​techcrunch.com. 

Meanwhile, DARPA recognized the need for non-surgical neural interfaces for healthy users (e.g. soldiers). In 2018 it announced the Next-Generation Nonsurgical Neurotechnology (N³) program. By 2019, N³ had funded six teams – at Battelle, Carnegie Mellon, Johns Hopkins APL, PARC (Xerox), Rice University, and Teledyne – to create high-resolution wearable brain interfaces for defense applications​darpa.mil​darpa.mil. These groups explored methods like novel optics, ultrasound, and electromagnetic nanoparticles to noninvasively read and write brain signals​darpa.mil. The envisioned outcomes included silent brain control of cyber systems or swarms of drones, enabling warfighters to “remain meaningfully involved” in split-second operations​darpa.mil​darpa.mil. N³ reflects the Department of Defense’s broader interest in hands-free command interfaces – essentially letting operators “think” to communicate with machines or each other. The U.S. Air Force, for instance, has tested BCI technology for pilots controlling drone swarms, and the Navy explored EEG-based threat detection (“mind-reading binoculars”) in parallel with DARPA’s efforts​pbs.org​wired.com. 

Other federal agencies have bolstered BCI and silent communication R&D. NIH has funded academic work on BCIs for patients, especially through the BRAIN Initiative. In 2021, NIH’s BRAIN program awarded $9.3M to a consortium led by Carnegie Mellon University to develop a minimally invasive BCI for communication in paralysis​mirm-pitt.net​mirm-pitt.net. This project, involving collaboration with Mount Sinai Hospital, University of Pittsburgh, and startup Synchron, is implanting Synchron’s Stentrode device in patients and aims to enable emailing or texting by thought alone​mirm-pitt.net​mirm-pitt.net. The Stentrode BCI (an endovascular electrode mesh) can be inserted via blood vessels without open brain surgery​mirm-pitt.net. Such NIH-funded trials are significant milestones toward a practical “brain-to-text” communication prosthesis. IARPA (Intelligence Advanced Research Projects Activity) has also contributed, primarily as a funding partner in the BRAIN Initiative and through programs focusing on neuroscience for intelligence applications​gao.gov​gao.gov. While IARPA hasn’t publicized a dedicated subvocal BCI program, it has run challenges on decoding neural signals (e.g. INSTINCT) and invested in related neurotech (such as the MICrONS project to map brain circuits), complementing DARPA’s more hardware-driven agenda. 

Finally, it’s worth noting NASA played a pioneering role in the mid-2000s. In 2004, researchers at NASA Ames developed a “subvocal speech” recognition system using neck-mounted sensors. They demonstrated that inaudible, internally spoken words produce detectable nerve signals in the throat, which their system could translate into commands with ~92% accuracy​spacenews.com​spacenews.com. In a NASA test, users silently mouthed words like “go”, “left”, “right” and even digits, and the system correctly recognized them to perform web searches​spacenews.com. This early achievement by Chuck Jorgensen’s team proved the concept of silent voice command interfaces​spacenews.com​spacenews.com. Such technology was envisioned for astronaut suits or noisy control towers where spoken communication is impractical​spacenews.com​spacenews.com. NASA’s work was one of the first public demonstrations that “internal speech” signals could be harnessed – a concept later echoed in DARPA’s programs and many academic projects.

Academic Research Labs and Universities

Academic institutions have been at the core of advancing BCIs for communication, often with government or philanthropic funding. A key academic consortium is the BrainGate team (Brown University, Stanford, Case Western, and others), which has worked on implanted BCIs for the severely paralyzed since the early 2000s. In 2004, BrainGate (supported by DARPA and Cyberkinetics Inc.) made history by enabling a locked-in patient to control a computer cursor by neural signals – effectively allowing rudimentary communication like typing out messages. Throughout the 2010s, BrainGate researchers continued to improve typing speeds via brain signals. By 2017, Stanford’s branch of BrainGate had a participant with paralysis using an implant to type 8 words per minute by point-and-click control of a keyboard​med.stanford.edu​med.stanford.edu. In 2021 they achieved a breakthrough: a BCI that decodes attempted handwriting motions from the motor cortex. In this study, a participant simply imagined writing letters by hand; machine learning decoded the neural patterns in real time, yielding text at 90 characters per minute (≈18 words per minute) – comparable to smartphone typing speed​med.stanford.edu​med.stanford.edu. This “mental handwriting” system, reported by Dr. Jaimie Henderson and Krishna Shenoy at Stanford, set a new speed record for BCI communication, demonstrating that tapping into the brain’s fine motor plans (for writing) can far outpace slower letter-by-letter selection methods​med.stanford.edu​med.stanford.edu. It was funded by sources including NIH, the HHMI, and DARPA, exemplifying how academic teams leverage multi-source funding for high-impact results. 

Another major academic player is UC San Francisco, where neurosurgeon Dr. Edward Chang’s lab focuses on decoding speech from cortical activity. In July 2021, UCSF (with support from Facebook Reality Labs and NIH) announced a notable milestone: a “speech neuroprosthesis” that enabled a man with severe paralysis to produce complete sentences by thought​ucsf.edu​ucsf.edu. In this trial, electrodes on the speech motor cortex captured the brain signals of attempted speech, and AI algorithms translated them into text displayed on a screen. It was the first time full words (as opposed to spelling letter-by-letter) were directly decoded from a human brain in real time​ucsf.edu​ucsf.edu. The system initially handled a limited vocabulary (around 50 words), allowing the participant to communicate at ~15 words per minute, but it demonstrated the feasibility of brain-to-text communication using natural speech brain patterns​ucsf.edu​ucsf.edu. This project was a partnership between academia and industry: Facebook’s Reality Labs provided funding and machine-learning expertise under a sponsored research agreement, while UCSF conducted the clinical research independently​ucsf.edu​ucsf.edu. The success illustrated how academic labs are translating years of neuroscience (Chang’s team had studied speech neurons for a decade) into practical communication aids. In 2022, the UCSF group went even further, unveiling a bilingual BCI that could decode sentences in two languages from brain signals​neurosurgery.ucsf.edu – pointing toward versatile real-world “speech prosthetics” for patients who have lost the ability to speak. 

Academic researchers have also pursued noninvasive paths to decoding internal speech. At the University of Maryland, Prof. Shihab Shamma’s team is studying “imagined speech” using EEG. In 2020, Shamma received a DURIP equipment grant to set up an advanced EEG lab to record the brainwaves of subjects thinking words without speaking​today.umd.edu​today.umd.edu. His team’s goal is an “EEG speech recognition system” – essentially like a Siri for your thoughts​today.umd.edu​today.umd.edu. They are collecting data of people listening to and imagining speech, then applying AI models to match brainwave patterns to words​today.umd.edu​today.umd.edu. This builds on Shamma’s decades of work in auditory neuroscience and follows earlier research (including by his former student Nima Mesgarani at Columbia) showing that internal or imagined speech produces recognizable neural signatures​today.umd.edu. Columbia University’s Mesgarani, in fact, demonstrated in 2019 that listening to speech and thinking of the same speech produce similar activity in the auditory cortex; by training deep learning models on neural recordings, his team reconstructed intelligible speech from brain signals​today.umd.edu. These academic efforts suggest a future where a noninvasive cap could pick up your internal monologue and translate it to text – a true “mind-reading” interface for communication. While invasive methods (implants) still have higher fidelity, the fact that universities are decoding words from scalp EEG and even fMRI has huge implications for broader, non-clinical use cases down the road. 

Academic innovation in “silent speech” interfaces isn’t limited to brain signals. Researchers have also tapped muscle signals and other physiological cues of internal speech. A notable example is the MIT Media Lab’s “AlterEgo” project, a student-led research effort from 2018. An MIT researcher demonstrates the AlterEgo wearable silent speech interface, which captures internal vocalization via neuromuscular signals on the face​newatlas.com​newatlas.com. The AlterEgo device is a lightweight mandibular headset that detects subtle electrical signals in the jaw and facial muscles when a user silently vocalizes words (moving the tongue and subvocal muscles but without sound)​newatlas.com. Arnav Kapur and Pattie Maes’ team at MIT proved that these signals can be translated by an AI into actual words. In trials, users wearing the device could perform tasks like arithmetic or chess by silently speaking commands; the system achieved about 92% transcription accuracy on a 20-word vocabulary after brief training​media.mit.edu​media.mit.edu. Essentially, AlterEgo reads the electrical impulses of internal speech (the “inner voice”) through surface electrodes, enabling a concealed, voice-free conversation with a computer​media.mit.edu​news.mit.edu. This academic prototype garnered wide attention as it allows discreet communication (the user hears replies via bone-conduction headphone) without any audible speech – ideal for assistive tech or environments where silence is golden​news.mit.edu​newatlas.com. Similar work has been pursued by other labs: for instance, early in the 2010s, researchers at Carnegie Mellon and Karlsruhe Institute of Technology used electromyography (EMG) sensors on the face and neck to recognize silently mouthed words, coining the term “silent speech interface.” This line of research – often funded by NSF grants or international agencies – runs parallel to brain-signal BCIs, and in some cases complements them (as in Facebook Reality Labs’ exploration of both brain and peripheral neural input for AR interfaces). 

In summary, universities and research institutes (UCSF, Stanford, Brown, MIT, Columbia, UMD, CMU, and more) have tackled the problem of neural or silent communication from multiple angles: invasive implants in speech motor areas, noninvasive brain decoding with EEG/fMRI, and peripheral nerve/muscle signal decoding. Many of these projects were enabled by federal grants (DARPA, NIH, NSF, Army) and often involved cross-disciplinary partnerships – neuroscientists, engineers, linguists, and computer scientists working together. We also see academia partnering with industry (UCSF + Facebook; CMU + Synchron; UW Madison using a Microsoft AI grant to tweet via BCI in 2009, etc.) to drive prototypes to demonstration. Academic labs have produced much of the core science and first proofs-of-concept that private companies are now racing to commercialize.

Private Sector Companies and Startups

In the last decade, private companies and startups have taken center stage in neural interface development, fueled by both venture capital and often by government contracts. One of the most prominent is Neuralink, co-founded by Elon Musk in 2016 with the vision of ultra-high-bandwidth brain implants. Neuralink’s R&D, funded privately (over $300M raised), has focused on a surgical BCI chip with thousands of electrodes for eventually enabling “telepathic” communication and machine control. By 2020–2021, Neuralink demonstrated its implant in pigs and monkeys – most famously, a monkey that mentally played Pong and later “typed” on a screen via brain signals. In a 2021 demo, Neuralink showed a monkey moving a cursor to spell out phrases on a virtual keyboard just by thinking of moving its hand. This suggested that a human with the same implant could potentially text or interface with a computer without using their body. In 2023, Neuralink received FDA clearance for its first human trials, and as of early 2025 it reportedly has a few human patients with its implant (though still behind some competitors)​reuters.com. Neuralink’s long-term goal explicitly includes direct brain-to-brain communication – essentially making “conceptual telepathy” possible – although initial applications target assistive communication for paralyzed patients. The company’s development has been accelerated by custom chip design, robotic surgeons, and a team of neuroscientists, and it has benefitted indirectly from the prior decade of DARPA/NIH implant research (some Neuralink team members came from academic BCI labs). While Neuralink hasn’t taken direct government funding publicly, it exists in an ecosystem catalyzed by those public investments (e.g. they leverage the open research on Utah arrays, brain decoding algorithms, etc.). 

Several startups focusing on BCIs for communication have roots in government-funded research or have secured such funding themselves. Synchron, founded in 2016 and originating from research at University of Melbourne, developed the Stentrode endovascular BCI. Synchron pursued a less invasive approach by deploying electrodes via blood vessels to the motor cortex. Early development was backed by Australian grants and DARPA’s Reliable Neural Interfaces program. By 2021–2022, Synchron gained significant private investment (including from tech billionaires Bill Gates and Jeff Bezos) and achieved human trials in Australia and the U.S., ahead of most rivals​reuters.com​reuters.com. As of 2025, Synchron’s device has been implanted in about 10 patients who have been able to control computers cursors or text interfaces by thought, making it the current leader in clinical BCI trials globally​reuters.com. Synchron’s endgame is to provide a commercial assistive device that lets people with paralysis regain communication – for instance, one patient with ALS used the stentrode to send text messages and emails hands-free. The company’s partnership with the NIH-funded consortium at CMU (mentioned earlier) underscores how private firms collaborate with academia on these goals​mirm-pitt.net​mirm-pitt.net. 

Another startup, Paradromics, based in Texas, spun out of DARPA’s NESD program. Paradromics has been working on high-channel-count implants (akin to a next-gen Utah array) and explicitly targeting speech prosthesis for locked-in patients as a first application. After receiving ~$18M from DARPA NESD in 2017​techcrunch.com, Paradromics raised additional venture funding ($20M seed in 2019​paradromics.com, $33M Series A in 2021) and developed a platform they call the “Direct Data Interface.” They achieved a lab prototype that can record from tens of thousands of neurons. The CEO has stated the company’s initial product aims to decode attempted speech in patients who cannot speak, effectively giving them a voice. Thus, Paradromics is a prime example of a private venture carrying forward a DARPA-funded concept (multi-neuron speech decoding) into a potential commercial clinical device. 

In the Big Tech arena, Meta (Facebook) Reality Labs (FRL) made headlines with its foray into brain interfaces. In 2017, Facebook announced an ambitious project to create a brain-to-text typing interface (targeting 100 words per minute) using noninvasive tech. FRL invested in academic partnerships, notably funding UCSF’s speech BCI research (2017–2021) which led to the successful demonstration of a speech prosthesis mentioned above​ucsf.edu. Although Facebook ultimately decided in 2021 to halt its direct brain-typing project for AR glasses (finding that invasive methods were needed to reach their speed goals)​theverge.com, the research they sponsored significantly advanced the field​ucsf.edu​ucsf.edu. Facebook pivoted to a more peripheral neural interface – in 2019 it acquired CTRL-Labs, a startup building EMG-based wristbands that read neural signals in arm muscles. CTRL-Labs’ technology, now part of Meta’s Reality Labs, enables users to control devices by intending to move their hand/fingers (useful for AR/VR control). While not “subvocal speech,” it represents a nonverbal neural interface for communication with computers. This acquisition (reportedly $500M) signaled the private sector’s belief that neural interfaces are key to the future of human-computer interaction (for AR, wearable computing, etc.). Other tech companies have also dabbled: for instance, Microsoft funded research on using EEG for typing, and Alphabet (Google) has supported neural speech decoding research via Google Brain and AI groups (e.g. helping analyze neural data with machine learning, though they have not launched a hardware BCI). 

A thriving niche of private companies builds noninvasive EEG-based communication tech. OpenBCI is a notable startup providing open-source EEG headsets and software to researchers and hobbyists in over 60 countries​openbci.com. Founded around 2014 via crowdfunding, OpenBCI’s affordable brainwave hardware has been used in countless BCI studies (including some on P300 spellers and neurofeedback communication aids). In 2020, OpenBCI partnered with VR companies (like Valve and Varjo) on “Project Galea,” integrating EEG, EMG, EOG and other biometrics into high-end VR headsets​roadtovr.com. This allows developers to experiment with brain signals in controlling or communicating with virtual environments – a step toward mainstream brain-machine interfaces in consumer tech. Although OpenBCI’s devices are lower resolution than medical BCIs, their widespread adoption and continued product development (backed by both enthusiasts and SBIR grants) show the private sector’s push to democratize brain interfaces. Similarly, companies like Emotiv (launched 2009) and InteraXon (Muse) have sold thousands of EEG headsets, initially targeting gaming or meditation, but also enabling basic brain-typing and communication apps for disabled users. These firms often collaborate with researchers; for example, in 2009 a University of Wisconsin student used an off-the-shelf EEG headset to post a Twitter message by brain, an early demo of EEG-to-text communication cited by media​pbs.org. 

Also worth mention are prosthetics and neurotech companies that support BCI communication research. Blackrock Neurotech (formerly Blackrock Microsystems) supplies the Utah electrode arrays and neurosignal amplifiers used by many academic groups (including BrainGate). They recently announced plans for a commercial BCI implant for assistive communication, leveraging decades of their tech in research. Medtronic and Abbott (neurostimulation giants) have shown interest in brain-computer applications beyond therapy, though mostly in pilot stages. New startups emerge regularly: e.g. MindMaze (Switzerland, BCI for stroke rehab), NextMind (France, visual EEG-based clicker, acquired by Snap Inc. in 2022), and Kernel (founded by Bryan Johnson, which pivoted to noninvasive EEG/functionalFNIRS helmets for cognitive monitoring). While not all are specifically about “subvocal” communication, they contribute pieces to the puzzle (better sensors, algorithms, funding impetus) that advance the overall domain of neural interfaces. 

Crucially, the private sector’s role is not just technology development but also financing and scaling. Venture capital funding into BCI startups surpassed $300M annually by the early 2020s​from-the-interface.com. This influx, outpacing many government research budgets, has accelerated progress and created a landscape where prototypes move rapidly towards productization. Startups often maintain ties with academia – for example, professors as scientific advisors or university labs as beta testers – and sometimes continue to receive government contracts even as private entities. DARPA and NIH grants have seeded several companies (e.g., Neuralink hired engineers from DARPA’s SUmmIT program; Synchron’s early DARPA grant; Paradromics’ NESD award). The result is a vibrant public-private ecosystem driving toward practical “thought-to-text” and “silent speech” solutions.

International and Global Contributions

Subvocal and neural interface research is a global endeavor. Outside the U.S., government agencies, academic consortia, and companies across Europe and Asia have mounted major programs. 

Europe (EU) – The European Union has funded large-scale projects on brain-computer interfaces, often under its Framework/Horizon programs. One flagship effort is the BrainCom project (2017–2022), a €8.35 million consortium of 10 partners across 6 countries​cogneuro.bio.lmu.de​cogneuro.bio.lmu.de. BrainCom (a FET Proactive project) is developing next-generation cortical implants to restore speech for people with paralysis or aphasia​cogneuro.bio.lmu.de​cogneuro.bio.lmu.de. It focuses on novel flexible electronics (like graphene-based ECoG grids) that can cover large speech areas of cortex with high resolution​cogneuro.bio.lmu.de​cogneuro.bio.lmu.de. By capturing distributed neural activity, they aim to decode complex language signals and eventually drive a speech prosthesis (e.g. a device that would “speak” the words a person intends to say)​cogneuro.bio.lmu.de​cogneuro.bio.lmu.de. This project unites European experts in neuroscience, engineering, and ethics, reflecting a strong interdisciplinary and international approach. The EU’s emphasis on collaborative research is also seen in initiatives like BNCI Horizon 2020, which coordinated BCI research roadmaps and user-needs assessments Europe-wide around 2013–2020, and the more general Human Brain Project (which, among many aims, developed brain simulating and neural data tools that can aid BCI decoding algorithms). European universities (TU Graz, EPFL, University of Tübingen, etc.) have long been leaders in EEG-based spellers and communication BCIs for assistive technology, often with support from national science foundations and EU grants. For instance, a team in Germany achieved a remarkable result in 2022: using a brain implant and neurofeedback, they enabled a completely locked-in patient with ALS to spell out messages at a few characters per minute by thinking – essentially establishing communication where there was none. This was not a single agency’s effort but built on decades of European medical BCI research aimed at restoring communication to those with “locked-in” syndrome. 

China – China has ramped up investment in brain-computer interfaces as part of a broader push in neuroscience and AI. The China Brain Project, launched in 2021 after years of planning, includes brain-machine interface technology as one pillar (alongside basic neuroscience and “brain-inspired” AI)​mgv.pku.edu.cn. Significant government funding has been allocated to BCI R&D in military, medical, and commercial sectors. By the mid-2020s, Chinese groups achieved notable milestones. In 2020, researchers at Tianjin demonstrated a brain-controlled car and a brain-typing system at a tech expo, drawing public attention. More substantially, in 2021 Chinese scientists (at Nankai University) reported a BCI that helped a paralyzed man type using brain signals via a headcap EEG with an input method optimized for Chinese characters. 

In 2022, reports emerged that the Chinese military was exploring “brain control” weaponry and communication, though details are scarce. What has become clear is China’s intent to catch up and even leapfrog in BCI technology. By 2025, a collaboration between the Chinese Institute for Brain Research, Beijing (CIBR) and a company NeuCyber Neurotech announced successful implantation of a new semi-invasive wireless brain chip in multiple human patients​reuters.com​reuters.com. Their chip, dubbed “Beinao No.1”, is placed on the surface of the brain via a small burr hole (less invasive than fully penetrating implants) and can transmit brain signals wirelessly. In March 2025, they revealed they had implanted this device in three patients and planned to reach 13 patients by year’s end, aiming to overtake Western trials in scale​reuters.com​reuters.com. If approved for a larger clinical trial of 50 patients in 2026, as they project, Beinao No.1 could become the world’s most widely tested implanted BCI​reuters.com​reuters.com. Chinese officials explicitly compare their progress to U.S. companies – noting, for example, that Synchron had 10 patients and Neuralink 3 at that time​reuters.com – underscoring a sense of competition. Additionally, Chinese researchers have touted developing the first two-way adaptive BCI with faster training (an achievement claimed in state media in 2021), and there are multiple startups, like NeuroXess in Shanghai, working on implantable electrode technology. NeuroXess plans a flexible shaft electrode for insertion into the motor cortex and has shown a monkey controlling a computer via a wireless implant (paralleling Neuralink’s demos)​tomshardware.com. The Chinese government has also issued ethical guidelines for BCI (in 2021) to encourage innovation while managing risks​pmc.ncbi.nlm.nih.gov, indicating that they foresee real deployment of these technologies. Overall, China’s BCI initiatives – backed by ministries, the PLA, and provincial innovation programs – encompass everything from mind-controlled drones to medical communication devices, in what could be described as a race to achieve practical “synthetic telepathy” and brain-control applications for both civilian and military use. 

Other Countries – Worldwide, many countries have niche contributions. Japan has explored BCIs for robotics and gaming; as early as 2008, Hitachi demonstrated an EEG “mind-control” toy train, and Honda showcased a Brain-Driven Robot arm. Japan’s government (MIC) funded a “BMI Project” (Brain Machine Interface) in late 2000s focusing on neuroprosthetics and communication aids for the disabled. Canada is home to InteraXon (Muse headband) and strong neuroscience programs (Toronto, Western), with some researchers decoding internal speech via near-infrared signals. Australia, aside from collaborating on Synchron, also hosted early BCI speller trials and an Army-funded project on “ultrasonic mind control” for pilots. Russia has had intermittent interest: Russian researchers in the 2010s claimed experiments on EEG-based telepathy and the Kremlin included neurotechnologies in its strategic tech roadmap, though concrete outputs are less visible. In the Middle East, projects like Qatar’s QRI have looked at BCI communication for disabilities, and Israel’s tech sector has small startups working on AI-based speech prosthetics. 

International collaborations are common. The EU, U.S., and Japan have jointly organized the International BCI Competition series to benchmark algorithms for P300 spellers and motor imagery communication, spurring global progress. Conferences like IEEE EMBC and the BCI Meeting foster cross-pollination of ideas globally. The result is that advances in one country (e.g. a German team’s signal processing algorithm, or a Taiwanese team’s dry EEG electrodes) quickly propagate and are adopted by others. As of 2025, the community is increasingly sharing data (e.g. through initiatives like the NeurotechX and OpenBCI communities, and open datasets from BCI challenges), which helps all players improve decoding of complex signals like imagined speech.

Milestones, Partnerships, and the Road Ahead

Over two decades, the pursuit of “silent” or direct neural communication has progressed from speculative research to concrete prototypes and early human trials. Below is a brief timeline of notable milestones and the key organizations behind them:

• Mid-2000s: NASA’s Ames Research Center demonstrates a 6-word subvocal speech recognizer using throat sensors (2004)​spacenews.com​spacenews.com. Cyberkinetics and Brown University conduct the first human trial of an implanted BCI (2004), letting a paralyzed man open emails via thought. DARPA’s early Brain-Machine Interface program (2002–2006) funds monkey neural control experiments and spawns the BrainGate trials. These set the stage for treating neural signals as a communication channel.
• Late 2000s: The U.S. Army/DARPA invest in synthetic telepathy – UC Irvine and partners begin studying EEG-based speech imagery (2008)​news.uci.edu. In parallel, emerging academic work on EMG-based silent speech (e.g. Carnegie Mellon) and noninvasive BCIs for spelling (Utrecht’s P300 speller, 2006) gain traction. In 2009, a University of Wisconsin student uses an EEG cap to post to Twitter by thought alone, a media-friendly demo of brain-to-internet communication​pbs.org.
• Early 2010s: Academic labs improve BCI spellers (Faster P300 interfaces, EEG dry electrodes from European projects). IARPA and DoD begin funding cognitive neuroscience (projects like ADAMS and ICArUS) that indirectly push BCI tech (e.g. better signal processing). In 2013, the U.S. BRAIN Initiative launches – by 2014 funding flows to dozens of BCI-related projects (e.g. decoding speech, high-density electrodes). Facebook, Inc. creates Building 8 (later FRL) and hires DARPA alumni to explore BCIs around 2016–2017, signaling Big Tech interest.
• Late 2010s: A wave of startups enters: Neuralink (2016), Kernel (2016), Synchron (2016) form, many drawing talent from academia. DARPA’s NESD program (2017) injects funds into Paradromics and academic teams at Brown, UCSF, UCLA, etc., accelerating high-bandwidth implant research​techcrunch.com​techcrunch.com. In 2018, MIT’s AlterEgo silent speech device is unveiled, demonstrating practical internal speech recognition without implants​media.mit.edu​media.mit.edu. That same year, DARPA announces N³ (2018) and by 2019 awards six teams for noninvasive BCIs​darpa.mil. On the international front, the EU’s BrainCom (2017) project kicks off focusing on speech prosthesis​cogneuro.bio.lmu.de​cogneuro.bio.lmu.de, and China lays groundwork for its Brain Project (announced in 2016, formalized by 2021). By 2019, Facebook’s partnership with UCSF yields initial results – decoding a small set of spoken phrases from brain activity in epilepsy patients​ucsf.edu. Also in 2019, Columbia University’s Nima Mesgarani uses AI to reconstruct heard speech from brain signals, hinting imagined speech could be next.
• Early 2020s: Despite a pandemic, progress accelerates. In 2020, a complete locked-in ALS patient in Germany communicates via an implanted BCI for the first time (letter-by-letter selection). Facebook Reality Labs refocuses on wrist EMG after concluding in 2021 that optical BCIs won’t hit their wpm targets​theverge.com – an example of a private pivot. 2021 stands out for milestones: Stanford’s high-speed “mindwriting” at 90 chars/min (Nature, May 2021)​med.stanford.edu; UCSF (Chang lab) publishes the first full-word neuroprosthetic speech in NEJM (July 2021)​ucsf.edu; Synchron completes human trials in Australia and gets FDA Breakthrough Device status in the U.S. Meanwhile, Neuralink shows its monkey typing and files for human trial approval. The NIH BRAIN Initiative funds the CMU-Mount Sinai-Synchron trial (late 2021) to test the Stentrode in 6 patients​mirm-pitt.net​mirm-pitt.net – a model academia-industry-clinical partnership.
• 2022–2023: The EU’s Horizon Europe program continues funding BCI consortia (e.g. Neurotwin and others for neurotech). Meta’s CTRL-Labs team debuts prototype AR wristbands. Synchron implants the first U.S. patient with a Stentrode (July 2022, at Mount Sinai in NYC) and raises $75M in venture funding. Neuralink finally secures FDA clearance in 2023 and begins recruiting for human trials, aiming to test its implant for assisting paralysis patients with communication. Chinese researchers at CIBR/NeuCyber publish results of their semi-invasive BCI in pigs and prepare for human tests, while another Chinese group (Harbin Institute) claims a noninvasive “mind-reading” headset that recognizes 100+ words with AI (though with lower accuracy, according to skeptics). By late 2023, Blackrock Neurotech in the U.S. announces it has kept implants in patients for 6+ years typing with BCIs, and plans a commercial communication BCI by 2025.
• 2024–2025: Competition and collaboration intensify. The GAO in the U.S. releases a technology assessment urging coordination on BCI development and noting that BCIs for communication in paralysis are nearing real-world deployment​gao.gov​gao.gov. New startups emerge targeting niches (speech decoding via ear EEG, throat sensor wearables for gaming chat, etc.). China (2025): CIBR/NeuCyber’s Beinao No.1 device is in 3 human patients and projected to rapidly expand trials​reuters.com​reuters.com, potentially giving China the largest cohort of BCI-communicating patients. This galvanizes international dialogues on standards and ethics. On the bright side, patients in multiple countries now use BCIs to text loved ones or operate speech synthesizers by thought – fulfilling the early promises of the field.

In conclusion, the development of subvocal and neural interface communication has been a multi-front effort: U.S. defense agencies drove early innovation and high-risk projects, health agencies and initiatives like BRAIN provided critical funding and clinical focus, academia delivered breakthroughs in decoding and prototypes, and private companies are translating these advances into scalable products. International contributions, especially from Europe’s collaborative projects and China’s state-driven programs, have added new impetus and approaches. Partnerships are ubiquitous – industry funds academia (Facebook/UCSF), governments fund startups (DARPA/Paradromics), universities spin-off companies (Brown → Cyberkinetics → Blackrock; Melbourne → Synchron), and consortium models blend all three. The “who did what” can be summarized as follows:

• Decoding silent/internal speech: NASA Ames (biomedical sensors) and DARPA (EEG telepathy) pioneered it; academic labs (UCI, Columbia, UMD) expanded it; startups like AlterEgo (MIT) built wearable solutions.
• High-resolution brain implants for communication: Developed via DARPA programs (BrainGate, NESD) by teams at Brown, Stanford, UCSF, etc.; now carried by companies (Neuralink, Paradromics, Blackrock) and clinical research (UCSF, Pittsburgh).
• Noninvasive BCIs and EEG-based communication: Advanced by European and U.S. labs (TU Graz, UMD, Wisconsin) with support from NSF/EU grants; commercialized in part by OpenBCI, Emotiv, and used in Meta’s AR plans.
• “Synthetic telepathy” for defense: Initially a DARPA/Army vision (Silent Talk); today partially realized in DARPA’s N³ teams (Battelle, CMU working on soldier brain-interfaces) and analogous efforts in China’s military research.

Every year, the gap between thought and communication narrows. From primitive one-bit yes/no signals, we are now at the point of deciphering imagined words and sentences from brain or nerve activity. Should these trends continue, the coming decade could see practical systems for “thinking to text” in everyday life – fulfilling a dream long pursued by these public, private, and academic pioneers around the world. 

Sources: The information above is drawn from a range of reports and press releases, including NASA and MIT news on silent speech interfaces​spacenews.com​media.mit.edu, DARPA announcements and budget documents detailing programs like Silent Talk, NESD, and N³​wired.com​darpa.mil, academic publications and news from UCSF, Stanford, UCI, and others on BCI breakthroughs​ucsf.edu​med.stanford.edu, as well as international news via Reuters on China’s BCI trials​reuters.com​reuters.com and EU project descriptions for BrainCom​cogneuro.bio.lmu.de​cogneuro.bio.lmu.de. These illustrate the breadth of organizations and collaborative milestones pushing the frontiers of subvocal and neural communication.

Citations

Mind over mouth? Study could lead to communicating via thoughts – UC Irvine News

https://news.uci.edu/2008/08/13/mind-over-mouth-study-could-lead-to-communicating-via-thoughts/U.S. Army funds university research on mind reading – Los Angeles Timeshttps://www.latimes.com/socal/daily-pilot/news/tn-dpt-xpm-2008-08-16-dpt-telepathy081508-story.htmlU.S. Army funds university research on mind reading – Los Angeles Timeshttps://www.latimes.com/socal/daily-pilot/news/tn-dpt-xpm-2008-08-16-dpt-telepathy081508-story.htmlPentagon Preps Soldier Telepathy Push | WIREDhttps://www.wired.com/2009/05/pentagon-preps-soldier-telepathy-push/Pentagon Preps Soldier Telepathy Push | WIREDhttps://www.wired.com/2009/05/pentagon-preps-soldier-telepathy-push/Pentagon Preps Soldier Telepathy Push | WIREDhttps://www.wired.com/2009/05/pentagon-preps-soldier-telepathy-push/Pentagon Preps Soldier Telepathy Push | WIREDhttps://www.wired.com/2009/05/pentagon-preps-soldier-telepathy-push/Pentagon Preps Soldier Telepathy Push | WIREDhttps://www.wired.com/2009/05/pentagon-preps-soldier-telepathy-push/GAO-25-106952; Brain-Computer Interfaces: Applications, Challenges, and Policy Optionshttps://www.gao.gov/assets/gao-25-106952.pdfGAO-25-106952; Brain-Computer Interfaces: Applications, Challenges, and Policy Optionshttps://www.gao.gov/assets/gao-25-106952.pdfDARPA awards $65 million to develop the perfect, tiny two-way brain-computer interface | TechCrunchhttps://techcrunch.com/2017/07/10/darpa-nesd-grants-paradromics/DARPA awards $65 million to develop the perfect, tiny two-way brain-computer interface | TechCrunchhttps://techcrunch.com/2017/07/10/darpa-nesd-grants-paradromics/DARPA awards $65 million to develop the perfect, tiny two-way brain-computer interface | TechCrunchhttps://techcrunch.com/2017/07/10/darpa-nesd-grants-paradromics/DARPA awards $65 million to develop the perfect, tiny two-way brain-computer interface | TechCrunchhttps://techcrunch.com/2017/07/10/darpa-nesd-grants-paradromics/DARPA awards $65 million to develop the perfect, tiny two-way brain-computer interface | TechCrunchhttps://techcrunch.com/2017/07/10/darpa-nesd-grants-paradromics/Six Paths to the Nonsurgical Future of Brain-Machine Interfaceshttps://www.darpa.mil/news/2019/nonsurgical-brain-machine-interfacesSix Paths to the Nonsurgical Future of Brain-Machine Interfaceshttps://www.darpa.mil/news/2019/nonsurgical-brain-machine-interfacesSix Paths to the Nonsurgical Future of Brain-Machine Interfaceshttps://www.darpa.mil/news/2019/nonsurgical-brain-machine-interfacesSix Paths to the Nonsurgical Future of Brain-Machine Interfaceshttps://www.darpa.mil/news/2019/nonsurgical-brain-machine-interfacesOur Latest – Darpa Funding Computer-Mediated Telepathy | Digital Nation | FRONTLINE | PBShttps://www.pbs.org/wgbh/pages/frontline/digitalnation/blog/2009/05/darpa-funding-computer-mediated-telepathy.htmlPentagon Preps Soldier Telepathy Push | WIREDhttps://www.wired.com/2009/05/pentagon-preps-soldier-telepathy-push/BRAIN Initiative Funding Received for BCI Development and Testing – Regenerative Medicine at the McGowan Institutehttps://mirm-pitt.net/brain-initiative-funding-received-for-bci-development-and-testing/BRAIN Initiative Funding Received for BCI Development and Testing – Regenerative Medicine at the McGowan Institutehttps://mirm-pitt.net/brain-initiative-funding-received-for-bci-development-and-testing/BRAIN Initiative Funding Received for BCI Development and Testing – Regenerative Medicine at the McGowan Institutehttps://mirm-pitt.net/brain-initiative-funding-received-for-bci-development-and-testing/GAO-25-106952; Brain-Computer Interfaces: Applications, Challenges, and Policy Optionshttps://www.gao.gov/assets/gao-25-106952.pdfGAO-25-106952; Brain-Computer Interfaces: Applications, Challenges, and Policy Optionshttps://www.gao.gov/assets/gao-25-106952.pdfNASA Develops System to Computerize Silent, Subvocal Speech – SpaceNewshttps://spacenews.com/nasa-develops-system-to-computerize-silent-subvocal-speech/NASA Develops System to Computerize Silent, Subvocal Speech – SpaceNewshttps://spacenews.com/nasa-develops-system-to-computerize-silent-subvocal-speech/NASA Develops System to Computerize Silent, Subvocal Speech – SpaceNewshttps://spacenews.com/nasa-develops-system-to-computerize-silent-subvocal-speech/NASA Develops System to Computerize Silent, Subvocal Speech – SpaceNewshttps://spacenews.com/nasa-develops-system-to-computerize-silent-subvocal-speech/Software turns ‘mental handwriting’ into on-screen words, sentences | News Centerhttps://med.stanford.edu/news/all-news/2021/05/software-turns-handwriting-thoughts-into-on-screen-text.htmlSoftware turns ‘mental handwriting’ into on-screen words, sentences | News Centerhttps://med.stanford.edu/news/all-news/2021/05/software-turns-handwriting-thoughts-into-on-screen-text.htmlSoftware turns ‘mental handwriting’ into on-screen words, sentences | News Centerhttps://med.stanford.edu/news/all-news/2021/05/software-turns-handwriting-thoughts-into-on-screen-text.htmlSoftware turns ‘mental handwriting’ into on-screen words, sentences | News Centerhttps://med.stanford.edu/news/all-news/2021/05/software-turns-handwriting-thoughts-into-on-screen-text.htmlSoftware turns ‘mental handwriting’ into on-screen words, sentences | News Centerhttps://med.stanford.edu/news/all-news/2021/05/software-turns-handwriting-thoughts-into-on-screen-text.html“Neuroprosthesis” Restores Words to Man with Paralysis | UC San Franciscohttps://www.ucsf.edu/news/2021/07/420946/neuroprosthesis-restores-words-man-paralysis“Neuroprosthesis” Restores Words to Man with Paralysis | UC San Franciscohttps://www.ucsf.edu/news/2021/07/420946/neuroprosthesis-restores-words-man-paralysis“Neuroprosthesis” Restores Words to Man with Paralysis | UC San Franciscohttps://www.ucsf.edu/news/2021/07/420946/neuroprosthesis-restores-words-man-paralysis“Neuroprosthesis” Restores Words to Man with Paralysis | UC San Franciscohttps://www.ucsf.edu/news/2021/07/420946/neuroprosthesis-restores-words-man-paralysis“Neuroprosthesis” Restores Words to Man with Paralysis | UC San Franciscohttps://www.ucsf.edu/news/2021/07/420946/neuroprosthesis-restores-words-man-paralysisSpeaking Two Languages Again – Neurological Surgery – UCSFhttps://neurosurgery.ucsf.edu/news/speaking-two-languages-againMaryland Today | These Researchers Want to Read Your Mindhttps://today.umd.edu/these-researchers-want-to-read-your-mindMaryland Today | These Researchers Want to Read Your Mindhttps://today.umd.edu/these-researchers-want-to-read-your-mindMaryland Today | These Researchers Want to Read Your Mindhttps://today.umd.edu/these-researchers-want-to-read-your-mindMaryland Today | These Researchers Want to Read Your Mindhttps://today.umd.edu/these-researchers-want-to-read-your-mindMaryland Today | These Researchers Want to Read Your Mindhttps://today.umd.edu/these-researchers-want-to-read-your-mindSilent headset lets users quietly commune with computershttps://newatlas.com/silent-headset-mit-alterego/54077/Silent headset lets users quietly commune with computershttps://newatlas.com/silent-headset-mit-alterego/54077/AlterEgo: A Personalized Wearable Silent Speech Interface — MIT Media Labhttps://www.media.mit.edu/publications/alterego-IUI/AlterEgo: A Personalized Wearable Silent Speech Interface — MIT Media Labhttps://www.media.mit.edu/publications/alterego-IUI/Computer system transcribes words users “speak silently” | MIT News | Massachusetts Institute of Technologyhttps://news.mit.edu/2018/computer-system-transcribes-words-users-speak-silently-0404Computer system transcribes words users “speak silently” | MIT News | Massachusetts Institute of Technologyhttps://news.mit.edu/2018/computer-system-transcribes-words-users-speak-silently-0404Silent headset lets users quietly commune with computershttps://newatlas.com/silent-headset-mit-alterego/54077/Chinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/Chinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/Paradromics Raises $20MM in Seed Funding Led by Prime Movers …https://www.paradromics.com/news/paradromics-raises-20mm-in-seed-funding-led-by-prime-movers-lab-to-advance-brain-computer-interfacesFacebook is giving up on brain-typing as an AR glasses interfacehttps://www.theverge.com/2021/7/14/22577095/facebook-bci-ucsf-chang-lab-brain-typing-research-update-project-steno-ar-vrOpenBCI | Homehttps://openbci.com/Valve, OpenBCI & Tobii to Launch VR Brain-computer Interface …https://www.roadtovr.com/valve-openbci-immersive-vr-games/Our Latest – Darpa Funding Computer-Mediated Telepathy | Digital Nation | FRONTLINE | PBShttps://www.pbs.org/wgbh/pages/frontline/digitalnation/blog/2009/05/darpa-funding-computer-mediated-telepathy.htmlVenture Capital Funding for Brain-Computer Interfaces outstrips …https://www.from-the-interface.com/bci-funding-clinical-trials/BrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlBrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlBrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlBrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlBrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlBrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlBrainCom – Information and Press Releases – Cognition and Neural Plasticity – LMU Munichhttps://www.cogneuro.bio.lmu.de/events/braincom/index.htmlProf. Lin Lu： China Brain Project: from bench to bedside. the solid …https://mgv.pku.edu.cn/english/research/rp/369700.htmChinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/Chinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/Chinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/Chinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/Chinese brain chip project speeds up human trials after first success | Reutershttps://www.reuters.com/business/healthcare-pharmaceuticals/chinese-brain-chip-project-speeds-up-human-trials-after-first-success-2025-03-31/China developed its very own Neuralink — Neucyber brain interface …https://www.tomshardware.com/tech-industry/china-developed-its-very-own-neuralink-neucyber-brain-interface-comes-from-a-neurotechnology-firmChina’s new ethical guidelines for the use of brain–computer …https://pmc.ncbi.nlm.nih.gov/articles/PMC11085839/

All Sources

news.ucilatimeswiredgaotechcrunchdarpapbsmirm-pittspacenewsmed.stanforducsfneurosurgery.ucsftoday.umdnewatlasmedia.mitnews.mitreutersparadromicsthevergeopenbciroadtovrfrom-the-interfacecogneuro.bio.lmumgv.pku.edutomshardwarepmc.ncbi.nlm.nih