BCI Glossary

A reference guide to the key terms, concepts, and technologies in the brain-computer interface field. 109 terms across 11 categories.

A

Action Potential

neuroscience

A brief electrical impulse (~1ms) generated by a neuron when it fires. Also called a 'spike.' The fundamental unit of neural communication that BCI systems detect and decode from brain tissue.

Active BCI

BCI types

A brain-computer interface that requires the user to consciously perform a mental task — such as imagining movement — to generate control signals. Contrast with passive BCI.

Adaptive Decoding

signal processing

Machine learning algorithms that continuously update their models based on changing neural signals, compensating for electrode drift, neural plasticity, and day-to-day variability in brain activity.

Afferent

neuroscience

Neural pathways that carry sensory information from the body toward the brain. In BCI, afferent signals are relevant to closed-loop systems that provide sensory feedback to the user.

Amyotrophic Lateral Sclerosis (ALS)

clinical

A progressive neurodegenerative disease that destroys motor neurons, eventually causing complete paralysis while leaving cognition intact. A key clinical population for BCI communication devices.

Assistive Technology

applications

Devices and systems designed to help people with disabilities perform tasks they would otherwise be unable to do. BCIs represent a new class of assistive technology that bypasses physical limitations entirely by reading brain signals.

B

Bandwidth (Neural)

performance

The amount of information a BCI can transmit per unit of time, often measured in bits per second. Higher bandwidth enables faster communication, more precise motor control, and richer data streams.

BCI Illiteracy

challenges

The phenomenon where a subset of users (estimated 15-30%) cannot produce reliably distinguishable brain signals to operate a BCI effectively. A significant challenge for universal BCI adoption.

Beta Oscillation

neuroscience

Brain rhythms in the 13-30 Hz frequency range, prominent in the motor cortex. Beta power decreases during movement planning (event-related desynchronization) and increases after movement (event-related synchronization), providing useful BCI control signals.

Bidirectional BCI

BCI types

A system capable of both reading neural signals and writing information back to the brain via stimulation. Enables sensory feedback — such as a prosthetic hand user feeling pressure through cortical microstimulation.

Biocompatibility

hardware

The ability of an implanted device to function in the body without causing adverse immune reactions, tissue damage, or chronic inflammation. A critical design constraint for long-term neural implants.

Biofeedback

applications

A technique that uses real-time monitoring of physiological signals — including brain activity — to help individuals learn to regulate their own bodily processes. A precursor concept to modern BCI neurofeedback.

Blood-Brain Barrier (BBB)

neuroscience

A selective membrane separating circulating blood from brain tissue. It protects the brain from pathogens but also complicates drug delivery and some BCI approaches. Endovascular BCIs like Synchron's Stentrode operate from within blood vessels without crossing the BBB.

Brain-Machine Interface (BMI)

BCI types

Synonymous with brain-computer interface (BCI). The term BMI is more common in academic literature, while BCI is preferred in clinical and commercial contexts.

Brain-Spine Interface

applications

A system that wirelessly bridges the brain and spinal cord, bypassing damaged neural pathways to restore movement. Pioneered by Gregoire Courtine at EPFL, enabling paralyzed patients to walk via thought-controlled spinal stimulation.

Breakthrough Device Designation

regulatory

An FDA program that provides expedited regulatory review for medical devices that offer significant advantages over existing treatments. Neuralink, Blackrock Neurotech, and Synchron have received this designation for their BCI systems.

Broca's Area

neuroscience

A brain region in the left frontal lobe critical for speech production. Targeted by speech BCI systems that decode intended words from neural activity in patients who have lost the ability to speak.

C

Calibration

signal processing

The process of training a BCI decoder to map a specific user's neural patterns to intended actions. Typically requires the user to perform or imagine tasks while the system records corresponding brain signals.

Chronic Implant

hardware

A neural device designed to remain in the body for months or years, as opposed to acute recordings used only during surgery. Long-term stability is a major engineering challenge due to glial scarring and electrode degradation.

Closed-Loop BCI

BCI types

A system that both reads neural signals and delivers stimulation back to the brain in real time, creating a feedback loop. Used in therapeutic applications like epilepsy management and adaptive deep brain stimulation.

Cognitive Liberty

ethics

The right to mental self-determination — including mental privacy, freedom of thought, and protection from unauthorized access to neural data. A legal and ethical framework advanced by Nita Farahany and Rafael Yuste.

Common Spatial Pattern (CSP)

signal processing

A signal processing technique that maximizes the difference in variance between two classes of EEG signals. Widely used in motor imagery BCIs to distinguish between imagined left-hand and right-hand movements.

Connectome

neuroscience

A comprehensive map of the neural connections in a brain or brain region. Efforts to map the human connectome inform BCI electrode placement strategies and decoder architecture.

Convolutional Neural Network (CNN)

signal processing

A deep learning architecture originally designed for image recognition, now applied to BCI signal decoding. CNNs can automatically extract spatial and temporal features from raw EEG or intracortical data.

Cortex

neuroscience

The outermost layer of the brain (cerebral cortex), responsible for higher cognitive functions including motor control, sensory processing, language, and decision-making. The primary target for most BCI electrode placements.

Cortical Column

neuroscience

A vertical unit of neural organization in the cortex, approximately 0.5mm in diameter, containing neurons that respond to similar stimuli. Understanding columnar organization helps optimize electrode placement in BCIs.

Cursor Control

applications

One of the earliest and most common BCI demonstrations — using decoded brain signals to move a computer cursor on screen. Serves as a benchmark for BCI performance and was first demonstrated by BrainGate in 2006.

D

Decoder

signal processing

The algorithm or model that translates raw neural signals into intended actions or commands. The core computational component of any BCI system, typically using machine learning techniques like Kalman filters or neural networks.

Deep Brain Stimulation (DBS)

applications

A neurosurgical procedure involving implanted electrodes that deliver electrical pulses to specific brain regions. Used to treat Parkinson's disease, essential tremor, and depression. Shares surgical techniques and hardware concepts with implantable BCIs.

Degrees of Freedom

performance

The number of independent parameters a BCI can control simultaneously — such as moving a robotic arm in three spatial dimensions plus grip. Higher degrees of freedom enable more natural and complex control.

Digital Twin (Neural)

signal processing

A computational model of an individual's brain activity patterns used to simulate and optimize BCI decoder performance before or between clinical sessions.

Dura Mater

neuroscience

The tough outermost membrane covering the brain and spinal cord. Epidural BCIs like Precision Neuroscience's Layer 7 sit on top of the dura, while subdural approaches like ECoG are placed beneath it.

E

Efferent

neuroscience

Neural pathways that carry motor commands from the brain to the body. Most BCIs decode efferent signals from the motor cortex to control external devices.

Electrocorticography (ECoG)

recording methods

A semi-invasive recording technique using electrodes placed directly on the brain's surface (subdural), beneath the skull but above the cortex. Offers better signal quality than EEG with lower surgical risk than intracortical arrays.

Electrode

hardware

A conductor used to detect or deliver electrical signals in neural tissue. BCI electrodes range from scalp-surface EEG pads to intracortical microelectrodes that penetrate brain tissue.

Electrode Array

hardware

A structured arrangement of multiple electrodes for simultaneous recording from many neural sites. Examples include the Utah Array (96-100 electrodes) and Neuralink's N1 (1,024 electrodes).

Electrode Impedance

hardware

The electrical resistance between an electrode and neural tissue. Changes in impedance over time — often due to glial scarring or electrode corrosion — indicate the health and longevity of an implanted BCI.

Electroencephalography (EEG)

recording methods

Non-invasive recording of electrical brain activity using electrodes placed on the scalp. The most accessible BCI recording method, used in consumer devices and research, though with lower spatial resolution than invasive approaches.

Electromyography (EMG)

recording methods

Recording of electrical activity in muscles. Used alongside BCI systems for hybrid control approaches and in prosthetic limbs that combine brain signals with residual muscle activity.

Endovascular BCI

BCI types

A brain-computer interface delivered through the body's blood vessels using catheter techniques, avoiding open brain surgery. Synchron's Stentrode is the leading example, inserted via the jugular vein into a cortical blood vessel.

Ensemble Recording

recording methods

Simultaneously recording from many neurons to capture population-level patterns of brain activity. Modern BCIs use ensemble recordings from hundreds to thousands of neurons to improve decoding accuracy.

Epidural

neuroscience

Located on or above the dura mater (the brain's outermost protective membrane). Epidural electrode placement avoids penetrating brain tissue, reducing surgical risk while sacrificing some signal resolution.

Event-Related Potential (ERP)

signal processing

A measurable brain response triggered by a specific stimulus or event. The P300 ERP — a positive voltage deflection occurring ~300ms after a rare or significant stimulus — is widely used in non-invasive BCI spellers.

F

FDA 510(k)

regulatory

A regulatory pathway requiring a manufacturer to demonstrate that a device is substantially equivalent to an existing legally marketed device. Precision Neuroscience received 510(k) clearance for its Layer 7 electrode array in 2025.

Feature Extraction

signal processing

The process of identifying and isolating informative patterns from raw neural data — such as frequency band power, spike rates, or spatial activation patterns — that the decoder can use to predict user intent.

Focused Ultrasound

recording methods

A non-invasive technique using converging ultrasound waves to stimulate or modulate brain activity at specific deep targets with millimeter precision. Pursued by Gestala and Merge Labs as a surgery-free BCI approach.

Functional Magnetic Resonance Imaging (fMRI)

recording methods

A neuroimaging technique that measures brain activity by detecting changes in blood oxygenation. Too slow and bulky for real-time BCI control but valuable for mapping brain regions to guide electrode placement.

Functional Near-Infrared Spectroscopy (fNIRS)

recording methods

A non-invasive optical technique that measures brain activity through changes in blood oxygenation using near-infrared light. Offers better spatial resolution than EEG and is being explored for wearable BCI devices.

G

Glial Scar

challenges

A layer of reactive glial cells (astrocytes) that forms around implanted electrodes over weeks to months. Increases electrical impedance and degrades signal quality, limiting the lifespan of intracortical implants.

Gray Matter

neuroscience

Brain tissue composed primarily of neuronal cell bodies, dendrites, and synapses. The outer cortical gray matter is the primary target for BCI electrodes because it contains the neurons that generate detectable electrical signals.

H

Hermetic Seal

hardware

An airtight enclosure protecting implant electronics from bodily fluids. Critical for long-term neural implant reliability — failure of the hermetic seal is a leading cause of chronic implant degradation.

Hybrid BCI

BCI types

A system combining multiple signal sources — such as EEG with eye tracking, or brain signals with EMG — to improve control accuracy and reliability beyond what any single modality can achieve.

I

Implantable Pulse Generator (IPG)

hardware

The battery-powered component of a neural stimulation system that generates the electrical pulses delivered to brain tissue. Similar in concept to a cardiac pacemaker but designed for neural applications.

Information Transfer Rate (ITR)

performance

A standardized metric for BCI performance measured in bits per minute, accounting for both speed and accuracy of communication. Enables comparison across different BCI paradigms and systems.

Intracortical

recording methods

Located within the cerebral cortex. Intracortical BCIs use penetrating electrodes inserted into brain tissue for the highest-quality neural recordings, but with greater surgical risk and foreign body response.

Invasive BCI

BCI types

A brain-computer interface requiring surgical implantation of electrodes in or on the brain. Offers the highest signal quality but carries surgical risks. Includes intracortical arrays (Neuralink, Blackrock) and subdural grids (ECoG).

Investigational Device Exemption (IDE)

regulatory

An FDA authorization allowing a medical device to be used in human clinical studies. Companies like Neuralink and Synchron operate under IDEs for their human BCI trials.

K

Kalman Filter

signal processing

A mathematical algorithm that estimates the state of a system from noisy measurements. Widely used in BCI decoders to smooth neural signal predictions and produce stable cursor or prosthetic control.

L

Latency

performance

The time delay between a user's neural intention and the BCI's output response. Low latency (under 100ms) is essential for real-time control applications like prosthetic limbs and cursor movement.

Local Field Potential (LFP)

neuroscience

The aggregate electrical potential generated by neural activity in a small brain region, reflecting the summed synaptic currents of thousands of neurons. Provides a more stable signal than individual spikes for long-term BCI use.

Locked-In Syndrome

clinical

A condition in which a person is fully conscious but unable to move or speak due to complete paralysis of voluntary muscles. One of the primary clinical motivations for BCI development, enabling communication through neural signals alone.

M

Magnetoencephalography (MEG)

recording methods

A non-invasive technique that detects the magnetic fields produced by neural electrical activity. Offers millisecond temporal resolution and better spatial localization than EEG but requires expensive, room-sized equipment.

Microelectrode

hardware

A miniature electrode (typically 10-100 micrometers) capable of recording from individual neurons or small neural populations. The building block of intracortical arrays like the Utah Array.

Motor Cortex

neuroscience

The brain region (precentral gyrus) that plans and executes voluntary movements. The primary target for motor BCIs, where neurons encoding movement direction, speed, and force are decoded to control prosthetics or cursors.

Motor Imagery

BCI types

The mental rehearsal of a movement without physically executing it. A foundational BCI paradigm — users imagine hand, foot, or tongue movements to generate distinct neural patterns that the system classifies into commands.

Mu Rhythm

neuroscience

An 8-13 Hz oscillation in the sensorimotor cortex that is suppressed (desynchronized) during both real and imagined movement. Used as a control signal in EEG-based motor imagery BCIs.

N

Neural Dust

hardware

A theoretical concept of wireless, millimeter-scale neural sensors powered by ultrasound that could be distributed throughout the brain. Proposed by UC Berkeley researchers as a future alternative to wired electrode arrays.

Neural Encoding

neuroscience

The process by which the brain represents information as patterns of neural activity — such as movement direction encoded by the firing rates of motor cortex neurons. Understanding encoding is essential for BCI decoder design.

Neural Interface

BCI types

Any device that establishes a direct communication pathway between the nervous system and an external system. An umbrella term encompassing BCIs, spinal cord stimulators, cochlear implants, and retinal prostheses.

Neural Network (Artificial)

signal processing

A machine learning architecture inspired by biological neural networks. Deep neural networks are increasingly used as BCI decoders, outperforming traditional methods like Kalman filters for complex decoding tasks like speech.

Neural Plasticity

neuroscience

The brain's ability to reorganize its structure and function in response to experience, injury, or training. BCI users exhibit neural plasticity as their brains learn to produce clearer control signals with practice.

Neuroethics

ethics

The study of ethical, legal, and social implications of neuroscience and neurotechnology. Central questions include neural data privacy, cognitive liberty, informed consent for brain implants, and equitable access to BCI technology.

Neurofeedback

applications

A form of biofeedback where users learn to modulate their own brain activity by observing real-time displays of their neural signals. Used therapeutically for ADHD, anxiety, and as a training method for BCI control.

Neuromodulation

applications

The alteration of nerve activity through targeted delivery of electrical, chemical, or ultrasound stimulation. Encompasses deep brain stimulation, transcranial stimulation, and the stimulation side of closed-loop BCIs.

Neuron

neuroscience

The fundamental signaling cell of the nervous system. The human brain contains approximately 86 billion neurons, each forming thousands of synaptic connections. BCIs detect electrical activity from populations of neurons to infer user intent.

Neuroprosthesis

applications

An artificial device that replaces or augments the function of a damaged nervous system. BCI-controlled prosthetic limbs, cochlear implants, and retinal implants are all examples of neuroprostheses.

Neurorights

ethics

Proposed human rights specific to the protection of brain data and mental integrity. The five neurorights — mental privacy, personal identity, free will, fair access to augmentation, and protection from bias — were formalized by Rafael Yuste's NeuroRights Foundation.

Neurostimulation

applications

The deliberate application of electrical currents or other energy to neural tissue to modulate activity. In BCI context, used to provide sensory feedback (e.g., touch sensation for prosthetic hands) or therapeutic effects.

Neurotransmitter

neuroscience

Chemical messengers (dopamine, serotonin, glutamate, GABA, etc.) released at synapses to transmit signals between neurons. Their concentrations influence neural circuit behavior detected by BCIs.

Non-Invasive BCI

BCI types

A brain-computer interface that does not require surgery, using external sensors like EEG caps, fNIRS headbands, or focused ultrasound. Lower signal quality but dramatically lower risk and broader accessibility.

O

Optogenetics

neuroscience

A technique using light-sensitive proteins to control genetically modified neurons with light. Enables extremely precise neural stimulation in research settings; potential future applications in high-precision BCI feedback.

P

P300

signal processing

A positive event-related potential occurring approximately 300 milliseconds after a rare or meaningful stimulus. The basis of the P300 speller — a non-invasive BCI that allows users to select letters by focusing attention on a flashing keyboard.

Passive BCI

BCI types

A brain-computer interface that monitors cognitive or emotional states without requiring conscious user effort. Applications include detecting drowsiness, stress, or attention levels in workplace and automotive settings.

Penetrating Electrode

hardware

An electrode designed to be inserted into brain tissue, typically reaching cortical layers 3-5 where large pyramidal neurons reside. The Utah Array and Neuralink's threads are penetrating electrode designs.

Percutaneous Connector

hardware

A physical port that passes through the skin to connect implanted electrodes to external electronics. Being replaced by wireless systems in newer BCI designs due to infection risk at the skin interface.

Population Coding

neuroscience

The principle that information in the brain is represented by the combined activity of many neurons rather than individual cells. BCI decoders exploit population coding by reading from large neural ensembles simultaneously.

PRIME Study

clinical

Neuralink's first human clinical trial (Precise Robotically Implanted Brain-Computer Interface), evaluating the safety and efficacy of the N1 implant in patients with quadriplegia.

Prosthetic Control

applications

Using BCI-decoded brain signals to operate artificial limbs or assistive devices. Ranges from basic grip commands to multi-degree-of-freedom arm and hand control with force and position feedback.

R

Readout

signal processing

The process of extracting usable information from neural signals. In BCI, readout refers to both the hardware (amplifiers, digitizers) and software (decoders) that convert brain activity into actionable data.

S

Sensorimotor Cortex

neuroscience

The combined motor and somatosensory cortex regions that control movement and process touch/proprioception. The primary target region for motor BCIs and bidirectional interfaces that both read and stimulate.

Signal-to-Noise Ratio (SNR)

performance

The ratio of desired neural signal power to background noise. A fundamental metric for BCI recording quality — invasive methods offer much higher SNR than non-invasive approaches, enabling more precise decoding.

Somatosensory Cortex

neuroscience

The brain region (postcentral gyrus) that processes touch, pressure, temperature, and proprioception. Targeted by BCI systems that deliver electrical stimulation to restore tactile feedback for prosthetic limb users.

Speech BCI

applications

A brain-computer interface designed to decode intended speech from neural activity, enabling communication for people who cannot speak. Recent systems have achieved over 90% word accuracy decoding from motor cortex and Broca's area.

Spike Sorting

signal processing

The computational process of attributing individual action potentials (spikes) to specific neurons from extracellular recordings where multiple neurons' signals overlap. Essential for extracting single-neuron data from multi-electrode arrays.

Spinal Cord Injury (SCI)

clinical

Damage to the spinal cord that results in loss of motor and/or sensory function below the injury level. One of the primary clinical targets for BCI technology, which can bypass the damaged spinal cord to restore communication or movement.

Steady-State Visual Evoked Potential (SSVEP)

signal processing

A brain response evoked by looking at a visual stimulus flickering at a specific frequency. Used in non-invasive BCIs where users select commands by fixating on different flickering targets, each at a unique frequency.

Stentrode

hardware

Synchron's endovascular electrode array, delivered through the jugular vein and deployed in a cortical blood vessel adjacent to the motor cortex. The first endovascular BCI implanted in a human, avoiding open brain surgery entirely.

Subdural

neuroscience

Located beneath the dura mater but above the brain surface. Subdural electrode grids (used in ECoG) offer high signal quality by being close to cortical tissue without penetrating it.

Synapse

neuroscience

The junction between two neurons where chemical or electrical signals pass from one cell to another. Synaptic activity generates the local field potentials and EEG signals detected by many BCI recording methods.

T

Telemetry

hardware

Wireless transmission of data from an implanted BCI to external receiving equipment. Modern BCIs increasingly use wireless telemetry to eliminate percutaneous connectors and reduce infection risk.

Tetraplegia

clinical

Paralysis of all four limbs and torso, typically resulting from cervical spinal cord injury. The primary patient population for current motor BCI clinical trials, as these patients have the greatest unmet communication and motor control needs.

Thalamus

neuroscience

A deep brain structure that relays sensory and motor signals to the cortex. Targeted in some BCI approaches for sensory feedback delivery and in deep brain stimulation for movement disorders.

Transcranial Direct Current Stimulation (tDCS)

applications

A non-invasive brain stimulation technique that delivers low-level electrical current through scalp electrodes. Explored as a method to enhance BCI performance by modulating cortical excitability during training.

Transcranial Magnetic Stimulation (TMS)

applications

A non-invasive technique using magnetic pulses to stimulate or inhibit neural activity in targeted brain regions. Used in research to map brain function and in clinical treatment of depression.

Transfer Learning

signal processing

A machine learning technique where a model trained on one task or user is adapted to work for another. Applied in BCI to reduce calibration time by transferring decoder knowledge across sessions or between users.

Tuning Curve

neuroscience

A graph showing how a neuron's firing rate varies with a stimulus parameter — such as the direction of arm movement. Motor cortex tuning curves are the basis for population vector algorithms used in early BCI decoders.

U

Utah Array

hardware

A silicon-based microelectrode array with 96-100 penetrating shanks, developed at the University of Utah and commercialized by Blackrock Neurotech. The most widely used intracortical recording platform in human BCI research for over two decades.

W

Wernicke's Area

neuroscience

A brain region in the left temporal lobe involved in language comprehension. Relevant to speech BCIs and reading comprehension interfaces targeting language processing circuits.

Wireless BCI

BCI types

A brain-computer interface that transmits data from implanted electrodes to external devices without physical cables passing through the skin. Reduces infection risk and increases user mobility compared to wired systems.

Stay Ahead of the Curve

Get the latest BCI news, breakthroughs, and market moves delivered to your inbox.

No spam. Unsubscribe anytime. We respect your privacy.