10 Companies Shaping the Brain-Computer Interface Landscape: Invasive vs Non-Invasive BCI Technology
The brain–computer interface (BCI) technology domain is constantly progressing as researchers and companies collaborate to devise inventive solutions that connect the human brain with external devices. This expanding sector, driven by the need for better communication, rehabilitation, and augmented cognitive capabilities, has given rise to both invasive and non-invasive BCI techniques. In this article, we will delve into the current status of brain-computer interface technology, explore the distinctions between invasive and non-invasive BCIs, and consider the potential consequences for the future development of this rapidly changing field.
Brain–computer interfaces (BCIs) are systems that read activity from the nervous system and translate it into control signals for external devices such as computer cursors, keyboards, wheelchairs, robotic limbs, or stimulation systems. In practical terms, they bypass muscles and peripheral nerves and create a more direct line between brain activity and hardware or software. Most current BCIs are built for people with severe motor or communication impairments, but similar signal decoding and feedback loops are also being explored for neurorehabilitation, mental state monitoring, and human–machine interaction in everyday environments.
Invasive BCIs
At a high level, any BCI has the same workflow. Sensors pick up brain-related signals, for example electrical activity from groups of neurons or blood-flow changes in active brain regions. Those raw signals are cleaned up and converted into numerical features, which are then fed into machine learning models that learn to map patterns of activity to specific commands such as “move cursor left,” “select this letter,” or “trigger this stimulation pattern.”
With practice, many users learn to modulate their brain activity in ways that make these decoders more reliable, turning the system into a closed loop between the brain, the algorithm, and the device being controlled.
The major fault line in today’s BCI landscape is how close the sensors get to the brain. Invasive BCIs involve implanted electrodes that sit on the brain surface or penetrate into cortical tissue. Because they are in direct contact with neurons, they can capture high-resolution signals with millisecond timing, which is useful for fine motor control, speech decoding, or precise neuromodulation.
These systems are being tested in people with paralysis, locked-in syndrome, or movement disorders, where the risk of surgery and long-term implants can be justified by potential gains in communication or motor function. Companies such as Neuralink, Blackrock Neurotech, and INBRAIN Neuroelectronics are all building around this implanted-electrode model, combining high-density recording hardware with decoding algorithms and sometimes closed-loop stimulation.
Despite the potential benefits, invasive BCI technology faces challenges, such as the risk of infection and the complexity of surgical implantation. Furthermore, the long-term stability and biocompatibility of these devices remain crucial concerns that researchers must address before widespread adoption.
Non-Invasive BCIs
Non-invasive BCIs sit at the other end of the spectrum. They use sensors outside the skull to pick up aggregate brain signals without surgery. The most common technique is electroencephalography (EEG), which records tiny voltage changes from electrodes placed on the scalp. Functional near-infrared spectroscopy (fNIRS) uses near-infrared light to track changes in blood oxygenation in the superficial cortex as a proxy for local neural activity. Electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) are two popular non-invasive techniques used by companies like Emotiv, Neurable, and Kernel.
These approaches are lower risk, easier to deploy, and more suitable for repeated use in clinics, home rehab, workplaces, and consumer devices, but the signals are weaker and blur together activity from larger brain areas, which limits spatial detail and sometimes speed.
Between those two poles, there is also a small but growing class of “partially invasive” systems such as stent-based electrode arrays that sit inside blood vessels rather than directly on the cortex. These devices aim to trade some signal quality for less invasive implantation, but the clinical and commercial experience is still early compared with fully invasive implants and established non-invasive EEG/fNIRS systems.
Our separate deep-dive in this series looks more broadly at how different BCI types, their underlying chips, AI stacks, and regulatory pathways are shaping the industry.
Read also: Emerging Brain-Computer Interface Industry Across Chips, AI, and Regulation
Invasive BCI developers
Neuralink
Founded in 2016 in the San Francisco Bay Area, Neuralink develops fully implantable brain–computer interfaces for people with severe paralysis, built around its N1 implant and R1 surgical robot. The first-in-human early feasibility trial (PRIME) runs under an FDA IDE cleared in May 2023.
Read also: Neuralink's Journey and the Broadening Horizons of Brain-Computer Interface Technology
In January 2024, Neuralink completed the first human implant and a live March demo where participant Noland Arbaugh used a cursor to play chess and navigate apps via thought. The company later disclosed that some electrode threads retracted in the weeks after surgery, reducing the number of effective channels.
Recently, Neuralink launched its first Great Britain trial (GB-PRIME) at UCLH and Newcastle to assess the N1 implants placed by the R1 robot safety and basic function for adults with severe paralysis; similar early-feasibility programs are running in the US, Canada, and the UAE.
Blackrock Neurotech
Founded in 2008 in Salt Lake City out of the University of Utah, Blackrock Neurotech develops invasive brain–computer interfaces based on intracortical microelectrode arrays and implantable electronics for people with paralysis and other severe neurological disorders.
The company markets the Utah Array intracortical electrode as a benchmark device for multi-channel recording and stimulation and supplies a full research stack including Cerebus and NeuroPort data acquisition systems, CereStim stimulators, and the CerePlex family of digital and wireless headstages that interface with Utah Arrays and related microelectrode arrays in animal models and humans.
Clinically, Blackrock is advancing MoveAgain, an implantable BCI system built around Utah Arrays and external processing hardware that decodes intended movement from neuronal activity so that people with tetraplegia or locked-in syndromes can control a mouse cursor, keyboard, wheelchair, prosthetic devices, and other assistive equipment through thought. MoveAgain received FDA Breakthrough Device designation in 2021.
INBRAIN Neuroelectronics
Founded in 2019 in Barcelona, Spain, INBRAIN Neuroelectronics develops invasive graphene-based brain–computer interface therapeutics and intelligent neural implants for neurological disorders such as Parkinson’s disease and epilepsy. Its platform uses ultra-thin (~10 µm) graphene cortical and deep-brain interfaces combined with machine learning to decode neural activity at high spatial resolution and deliver adaptive neuromodulation (BCI-Tx) during surgery and in chronic implants.
INBRAIN’s lead program is the Intelligent Network Modulation System, an implantable graphene-based interface that received FDA Breakthrough Device Designation in 2023 as an adjunctive therapy for Parkinson’s disease.
The same graphene platform underpins a cortical BCI used in a first-in-human procedure in 2024 in a brain-tumor resection patient at Salford Royal Hospital and an ongoing first-in-human study sponsored by the University of Manchester that has reported interim safety and high-fidelity mapping results.
Recently, INBRAIN entered a strategic collaboration with Microsoft to use Azure-based agentic AI and time-series models to adapt its graphene BCI-Tx platform in real time for precision neurology.
g.tec medical engineering
Founded in 1999 in Schiedlberg, Austria, g.tec medical engineering develops brain-computer interfaces and neurotechnology for both invasive and non-invasive recordings in research and clinical settings.
Its amplifier line includes the FDA and CE approved g.HIamp system, and the g.Nautilus family of wearable wireless amplifiers, with g.Nautilus PRO positioned as a CE-certified and FDA-cleared medical EEG system for clinical use.
g.tec runs two main BCI-driven clinical product lines: recoveriX, a closed-loop motor-rehabilitation system for stroke and multiple sclerosis, and mindBEAGLE, a portable EEG-based BCI used to assess command following and provide basic communication for patients with disorders of consciousness or locked-in syndrome.
An ongoing clinical trial is testing recoveriX-based treatment in Parkinson’s disease compared with standard FES+VR+motor-imagery protocols without EEG monitoring. recoveriX is also available in hospitals and large neurorehabilitation clinics in multiple countries.
Non-invasive BCI developers
Kernel
Founded in 2016 in Los Angeles, Kernel builds noninvasive neuroimaging headsets based on time-domain fNIRS (Flow/Flow2) to measure cortical hemodynamics for research and emerging clinical use.
Recently, Kernel has started moving Flow2, a whole-head time-domain fNIRS headset, into clinical-facing pilots. The company launched observational studies in major depressive disorder (PREDICT) with Bespoke Treatment to track pre/during/post-treatment signals; it also opened a mild cognitive impairment program with first patients measured at a Profound Research site. Peer-reviewed and preprint work reported test-retest reliability of Flow2 metrics across tasks, days, and headsets, positioning the system for multi-site use.
Neurable
Founded in 2015 in Boston as a spinout from the University of Michigan’s Direct Brain Interface Lab, Neurable develops non-invasive, EEG-based brain–computer interfaces embedded into everyday devices. The company’s current focus is consumer-grade over-ear headphones that integrate dry-electrode EEG and an analytics stack to estimate focus, fatigue, and cognitive load in real time, exposing those metrics through a companion app for work and wellness use cases.
The company is commercializing its technology through products such as its own Enten smart headphones and the MW75 Neuro, a Master&Dynamic active-noise-cancelling headset with embedded Neurable EEG sensors launched in September 2024 and positioned as a “brain-tracking” headphone for focus and mental health monitoring.
Neurable also reports work with the US Air Force Research Laboratory’s 711th Human Performance Wing on cognitive-state monitoring, reinforcing its role on the non-invasive, wearable side of the BCI landscape.
BrainCo
BrainCo was founded in 2015 in Somerville/Cambridge, Massachusetts by Bicheng Han while at Harvard’s Center for Brain Science, and later established a China headquarters in Hangzhou; it develops noninvasive BCI products across consumer EEG, STEM education, and prosthetics through three divisions—FocusCalm (EEG neurofeedback headset with companion app), NeuroMaker (classroom kits and curricula for BCI/biomedical projects), and BrainRobotics (AI-adaptive myoelectric prosthetics).
Recently, Hangzhou authorities reported full production of bionic prosthetics and NeuroMaker’s 2025 catalog added Hand 2.0, BioSensor Kit and BCI “NeuroRacing” with >100 hours of curriculum.
Emotiv
Founded in 2011 in San Francisco, Emotiv develops non-invasive, EEG-based brain–computer interfaces built around wireless headsets and earbud form factors for research, BCI prototyping, and workplace or wellness monitoring.
Recently, Emotiv has shifted more emphasis toward passive cognitive and affective monitoring in everyday environments using MN8 smart earbuds plus the Emotiv app, which estimate attention, cognitive load, and stress during work and daily activities. These devices sit on top of a software suite that includes EmotivPRO for EEG recording, EmotivBCI and Cortex SDK for mental-command BCIs and API access, BrainViz for 3D visualization, and the Emotiv Launcher as a unified entry point, with Virtual Brainwear providing simulated data streams for developers.
OpenBCI
Founded after a 2013 Kickstarter campaign by Joel Murphy and Conor Russomanno and now headquartered in Brooklyn, New York, OpenBCI develops open-source EEG and biosensing hardware and software used for non-invasive brain–computer interface research, neurofeedback, and physiological computing.
In recent years OpenBCI has extended into mixed reality and multi-sensor neurotechnology with Galea—a headset platform that combines EEG, EMG, EDA, PPG, and eye-tracking and integrates with XR headsets such as Varjo Aero and XR-3 for VR/MR user research and BCI prototyping.
In November 2025, OpenBCI and Pupil Labs announced Galea Neon—the first fully wireless, all-in-one brain, body, and eye-tracking headset.
MindMaze
Founded in 2012 in Lausanne, Switzerland, MindMaze develops digital neurotherapeutics for motor and cognitive recovery after stroke, traumatic brain injury, and other neurological conditions.
The company’s MindMotion line uses game-based VR/3D environments, motion capture, and sensor-based assessment rather than implants, with MindMotion PRO designed for early, hospital-based upper-limb rehab and MindMotion GO configured for outpatient and home telerehabilitation using depth cameras and controllers to track full-body movements.
MindMotion GO and MindMotion PRO are listed as FDA-regulated devices and carry CE marks. MindMotion GO is now deployed as a telerehab program in leading centers such as Mount Sinai’s Abilities Research Center and Johns Hopkins.
Balancing Advantages and Challenges
As both invasive and non-invasive brain-computer interface technology continue to evolve, researchers and companies weigh the benefits and drawbacks of each approach. Invasive BCIs may offer superior signal resolution, but their inherent risks and complex surgical procedures present significant barriers to entry. On the other hand, non-invasive BCIs provide a safer, more accessible alternative, albeit with reduced signal quality.
The field appears to be moving toward a portfolio of BCI options rather than a single “best” design. High-bandwidth invasive BCIs are likely to stay focused on severe paralysis, advanced neuromodulation, and neurosurgical use, while non-invasive and wearable BCIs expand into neurorehabilitation, occupational health, and consumer and research applications. Partially invasive devices like stent-based electrodes placed in blood vessels may eventually fill a middle ground for patients who need more detailed signals than non-invasive systems can deliver but are not candidates for open-brain surgery. Across all of these, progress in electrode materials, low-power electronics, wireless communication, and AI-based decoding is gradually improving reliability and usability.
Topic: NeuroTech
