What is a brain-computer interface?
A brain-computer interface (BCI, also called a brain-machine interface) is a system of hardware and software that reads patterns in brain or nerve signals and uses them to operate an external device. Because raw brain signals are dense and hard to interpret, most BCI systems lean on AI, including machine learning and neural network models, to make sense of the data.
The FDA describes BCI-based neuroprostheses as devices that connect to the central or peripheral nervous system to restore motor or sensory function that’s been lost, either by stimulating neural tissue directly or by reading brain activity and translating it into commands for an external device.
Signals can be captured in three ways: non-invasively, through sensors placed on the scalp or skin (EEG or EMG); through a partially invasive method called electrocorticography, where electrodes sit on the brain’s surface beneath the skull; or through fully invasive intracortical recording, where electrodes penetrate brain tissue directly for the sharpest signal quality.
Applications
BCI-based neuroprosthetics already restore or enhance hearing, vision, mood, mobility, cognition, and communication. Cochlear implants, FDA-approved since 1995, convert sound into electrical signals sent to the auditory nerve. Visual prostheses like the Argus II, approved in 2013, do something similar for the visual cortex.
Limb prosthetics controlled by brain or nerve signals are advancing on two fronts: closed-loop systems that feed sensory information back to the user so a prosthetic limb can be “felt,” and wireless systems controlled remotely by brain signals. Researchers are also exploring brain-to-brain interfaces that transmit activity between two subjects, and hippocampal prostheses aimed at restoring or enhancing memory and cognition in people with neurodegenerative disease.
Market and Companies
Analysts projected the neurostimulation device market to hit $13 billion by 2023 and the neuroprosthetics market close behind at $14 billion by 2020. Established consumer products include Halo Neuroscience’s motor cortex stimulator, Emotiv, MUSE, and NeuroSky headbands for meditation and stress tracking, and Medtronic’s FDA-approved deep brain stimulators for Parkinson’s symptoms.
Startups in the space include Kernel (hippocampal prosthetics, $100 million raised), MindMaze (VR-based rehab tools, $108.5 million raised), NeuroPace (seizure-monitoring implants), BrainRobotics (low-cost EMG-controlled prosthetic hands), and Neuralink, which is developing implantable cognition-focused electrodes.
A Brief History
The field traces back to 1976, when UCLA researcher Jacques Vidal coined the term BCI after showing that visual evoked potentials could steer a cursor through a maze. Milestones followed steadily: the first high-quality invasive human brain implant in 1998, a BCI restoring limited hand movement to a quadriplegic patient in 1999, monkeys controlling a cursor in 2002, the first public BCI game demo in 2003, a monkey controlling a robotic arm in 2005, and a voiceless phone call demonstrated in 2008.
Open Questions
As the technology matures, it raises real ethical and legal questions around identity, privacy, consent, and security, particularly since BCI devices collect sensitive data tied to memory, emotion, and thought. Researchers have flagged risks including brain-signal-based lie detection, using brainwaves as a password, and pulling PINs or other sensitive information directly from recorded brain data.
Some people are already experimenting with brain stimulation outside any clinical setting, a practice sometimes called “brain zapping,” which several researchers have warned lacks adequate safety and efficacy data. One proposal, a Neuro Information Nondiscrimination Act modeled on the 2008 Genetic Information Nondiscrimination Act, would bar employers from requesting or acting on an employee’s neural data.