Neuroprosthetics: revision 2 to 3

Revision 2

Revision 3

categories: ["cybernetics", "bodies"]

tags: ["neurotechnology", "implants", "sensory restoration", "paralysis", "neuromodulation", "assistive technology"]

summary: "The engineering field that builds devices interfacing directly with the nervous system to restore lost sensory, motor, cognitive, or autonomic function."

summary: "The engineering field that builds devices interfacing directly with the nervous system to restore lost sensory, motor, cognitive, or autonomic function."

updated: "2026-10-27"

updated: "2026-10-29"

humanEvidence: "Cochlear implants, deep brain stimulators and spinal cord stimulators are in routine clinical use in large patient populations; every bidirectional cortical result rests on single-digit numbers of participants."

humanEvidence: "Cochlear implants, deep brain stimulators and spinal cord stimulators are in routine clinical use in large patient populations; every bidirectional cortical result rests on single-digit numbers of participants."

access: "Hearing, movement-disorder, pain and bladder devices are approved and reimbursed in high-income countries; cortical, spinal and sensory-feedback systems are available only inside research studies."

access: "Hearing, movement-disorder, pain and bladder devices are approved and reimbursed in high-income countries; cortical, spinal and sensory-feedback systems are available only inside research studies."

reversibility: "difficult"

{ "label": "Scope", "value": "Sensory, motor, cognitive, autonomic" },

{ "label": "Most deployed device", "value": "Cochlear implant", "link": "/wiki/cochlear-implant" },

{ "label": "Interface sites", "value": "Cortex, deep nuclei, spinal cord, peripheral nerve" },

{ "label": "Field established", "value": "1960s–1970s" },

{ "label": "Field established", "value": "1960s-1970s" },

{ "label": "Main bottleneck", "value": "Writing information in, not reading out" },

{ "label": "Chronic failure mode", "value": "Foreign-body encapsulation" },

{ "label": "Readiness", "value": "TRL 9 (hearing) to TRL 3 (memory)" }

**Neuroprosthetics** is the engineering discipline concerned with devices that exchange signals

directly with the nervous system in order to replace a function the nervous system has lost. It

spans devices in routine clinical use — the [[cochlear-implant]], the spinal cord stimulator,

spans devices in routine clinical use, such as the [[cochlear-implant]], the spinal cord stimulator

[[deep-brain-stimulation]] — and devices that exist only in single-patient research studies, such as

and [[deep-brain-stimulation]], as well as devices that exist only in single-patient research

cortical arrays that restore a sense of touch. The unifying problem is not electronics but

studies, such as cortical arrays that restore a sense of touch. The problem that unites them is the

interface: how to place a durable transducer against excitable tissue that reacts to its presence,

interface rather than the electronics: how to place a durable transducer against excitable tissue

and how to encode information in a form neurons will accept.

that reacts to its presence, and how to encode information in a form neurons will accept.

## Scope and taxonomy

Four families are usually distinguished, though real devices increasingly cross the boundaries.

**Sensory prostheses** convert an external stimulus into neural activity. The cochlear implant and

the [[retinal-implant]] are the archetypes; vestibular implants for bilateral vestibular loss have

the [[retinal-implant]] are the archetypes, and vestibular implants for bilateral vestibular loss

reached small human trials. These devices must solve an encoding problem, because the natural code

have reached small human trials. These devices have to solve an encoding problem, because electrodes

of the sense organ is not reproducible with electrodes.

cannot reproduce the natural code of the sense organ.

**Motor prostheses** read intent and act on it. Cortical systems for people with paralysis fall

under [[brain-computer-interface]]; peripheral systems that read muscle activity fall under

under [[brain-computer-interface]], and peripheral systems that read muscle activity fall under

[[myoelectric-prosthetics]]; functional electrical stimulation drives the user's own paralysed

[[myoelectric-prosthetics]]. Functional electrical stimulation drives the user's own paralysed

muscles or spinal circuitry rather than an external actuator.

muscles or spinal circuitry instead of an external actuator.

**Autonomic and organ-directed devices** are numerically the largest category and receive the least

**Autonomic and organ-directed devices** are the largest category by number and receive the least

attention. Sacral neuromodulation for bladder dysfunction, vagus nerve stimulation for epilepsy and

for rehabilitation after stroke, phrenic pacing for ventilator dependence, and spinal cord

stimulation for chronic pain together account for far more implants than every cortical device

combined.

**Cognitive prostheses** attempt to restore a computation rather than a signal path. The only

serious example is the hippocampal [[memory-prosthesis]], and it remains at the proof-of-concept

serious example is the hippocampal [[memory-prosthesis]], which remains at the proof-of-concept

stage.

## How a neuroprosthesis works

Every device is a loop with four parts, and each part fails differently.

Every device is a loop with four parts, and each part fails in its own way.

The *transducer* couples electrical or optical energy to tissue. Choices range from penetrating

The *transducer* couples electrical or optical energy to tissue. The options range from penetrating

microelectrode arrays such as the [[utah-array]], through surface arrays used in

microelectrode arrays such as the [[utah-array]], through the surface arrays used in

[[ecog-interfaces]], to nerve cuffs, intrafascicular electrodes, and the endovascular approach of

the [[stentrode]]. Closer coupling means more information per channel and more tissue damage.

the [[stentrode]]. Closer coupling yields more information per channel and more tissue damage.

The *front end* amplifies microvolt signals in the presence of stimulation artefacts thousands of

times larger, and does so within a power budget that will not heat the surrounding tissue by more

times larger, within a power budget that keeps the surrounding tissue from heating by more than a

than a fraction of a degree.

fraction of a degree.

The *algorithm* converts signals into commands or converts sensor data into stimulation patterns.

The *algorithm* converts signals into commands, or sensor data into stimulation patterns. On the

On the read side this is [[neural-decoding]], which has improved faster than the hardware. On the

read side this is [[neural-decoding]], which has improved faster than the hardware. On the write

write side it is an encoding problem with no equivalent body of theory.

side it is an encoding problem with no comparable body of theory.

The *actuator* is a robotic limb, a synthesized voice, a stimulator on the user's own spinal cord,

or a pattern of current in a sensory nucleus. Systems that both read and write are called

bidirectional, and they are where the field's hardest problems live.

bidirectional, and the field's hardest problems are concentrated in them.

## Development history

```timeline

[

  { "year": "1957–1961", "title": "Direct stimulation of a sense", "text": "Djourno and Eyriès stimulate the auditory nerve of a deaf patient; William House begins implanting single-channel cochlear devices in Los Angeles." },

  { "year": "1957-1961", "title": "Direct stimulation of a sense", "text": "Djourno and Eyriès stimulate the auditory nerve of a deaf patient; William House begins implanting single-channel cochlear devices in Los Angeles." },

  { "year": "1968", "title": "Cortical visual prosthesis", "text": "Brindley and Lewin evoke patterned phosphenes from an electrode array over the occipital cortex of a blind volunteer." },

  { "year": "1968", "title": "Cortical visual prosthesis", "text": "Brindley and Lewin evoke patterned phosphenes from an electrode array over the occipital cortex of a blind volunteer." },

  { "year": "1987–1997", "title": "Deep brain stimulation becomes clinical", "text": "Chronic high-frequency thalamic stimulation for tremor moves from a Grenoble case series to regulatory approval." },

  { "year": "1987-1997", "title": "Deep brain stimulation becomes clinical", "text": "Chronic high-frequency thalamic stimulation for tremor moves from a Grenoble case series to regulatory approval." },

  { "year": "2004–2006", "title": "Human cortical motor decoding", "text": "A participant with tetraplegia in the BrainGate pilot controls a cursor and a prosthetic hand with a penetrating array." },

  { "year": "2004-2006", "title": "Human cortical motor decoding", "text": "A participant with tetraplegia in the BrainGate pilot controls a cursor and a prosthetic hand with a penetrating array." },

  { "year": "2014–2016", "title": "Touch written back in", "text": "Peripheral nerve cuffs in amputees and intracortical microstimulation of somatosensory cortex in a participant with spinal cord injury evoke localized tactile percepts." },

  { "year": "2014-2016", "title": "Touch written back in", "text": "Peripheral nerve cuffs in amputees and intracortical microstimulation of somatosensory cortex in a participant with spinal cord injury evoke localized tactile percepts." },

  { "year": "2018–2023", "title": "Spinal interfaces restore walking", "text": "Targeted epidural stimulation restores stepping after spinal cord injury; a 2023 system links a cortical implant to a spinal stimulator to give volitional control of walking." },

  { "year": "2018-2023", "title": "Spinal interfaces restore walking", "text": "Targeted epidural stimulation restores stepping after spinal cord injury; a 2023 system links a cortical implant to a spinal stimulator to give volitional control of walking." },

  { "year": "2020s", "title": "Commercial implants enter trials", "text": "Multiple companies run early feasibility studies of implanted interfaces, and adaptive sensing-and-stimulating devices reach the market for movement disorders." }

  { "year": "2020s", "title": "Commercial implants enter trials", "text": "Multiple companies run early feasibility studies of implanted interfaces, and adaptive sensing-and-stimulating devices reach the market for movement disorders." }

]

```

## Current state

The clinically settled devices restore hearing, suppress movement-disorder symptoms, manage pain,

The clinically settled devices restore hearing, suppress the symptoms of movement disorders, manage

and control bladder and bowel function. They are unglamorous, well reimbursed in wealthy countries,

pain, and control bladder and bowel function. They are unglamorous, well reimbursed in wealthy

and improving incrementally.

countries, and are improving incrementally.

The research frontier is bidirectional restoration after paralysis and amputation. Three results

The research frontier is bidirectional restoration after paralysis and amputation, and three results

define it. Intracortical microstimulation of human somatosensory cortex produces tactile sensations

that a participant localizes to specific fingers of a paralysed hand, and adding that feedback

roughly halved the time needed to complete an object-transfer task with a robotic

arm.[^flesher2016][^flesher2021] Implanted electrodes in residual peripheral nerves evoke graded

touch in amputees and, combined with skeletal attachment through

touch in amputees, and combined with skeletal attachment through [[osseointegration]] they have

[[osseointegration]], have supported prostheses used at home for years.[^ortizcatalan2020] And

supported prostheses used at home for years.[^ortizcatalan2020] Epidural spinal stimulation,

epidural spinal stimulation, targeted to the dorsal roots that recruit specific leg muscle groups,

targeted to the dorsal roots that recruit specific leg muscle groups, restores stepping in people

restores stepping in people with severe spinal cord injury.[^wagner2018] A cortical implant wired to

with severe spinal cord injury.[^wagner2018] A cortical implant wired to such a stimulator gave one

such a stimulator gave one participant volitional, thought-driven walking, and produced neurological

participant volitional, thought-driven walking, and produced neurological recovery that persisted with

recovery that persisted with the system switched off.

the system switched off.

Each of these involves single-digit numbers of participants. A device cleared in the United States

in late 2024 delivers non-invasive transcutaneous spinal stimulation to improve hand and arm

strength after cervical injury, and represents the first of this generation to reach the market.

strength after cervical injury, and it is the first of this generation to reach the market.

## The feedback problem

Reading intent out of the nervous system is comparatively tractable, because a decoder can be

trained on labelled attempts and improved offline without touching the tissue. Writing information

in is harder for reasons that are structural rather than temporary.

in is harder, for structural rather than temporary reasons.

An electrode excites every axon within its field indiscriminately, including axons of passage

An electrode excites every axon within its field indiscriminately, including axons of passage that

belonging to unrelated circuits. Natural sensory input arrives as spatiotemporally patterned

belong to unrelated circuits. Natural sensory input arrives as spatiotemporally patterned activity

activity across specific cell types; electrical stimulation produces synchronous activation of a

across specific cell types. Electrical stimulation instead activates a sphere of tissue

sphere of tissue, which the brain has never encountered and has no code for. Percepts evoked this

synchronously, a pattern the brain has never encountered and has no code for. Percepts evoked this

way are describable but unnatural: pressure, tingling, or a vague spot of light rather than texture,

way can be described but feel unnatural: pressure, tingling or a vague spot of light, rather than

temperature, or shape. Repeated stimulation drives adaptation, so amplitude must be modulated to

texture, temperature or shape. Repeated stimulation causes adaptation, so amplitude has to be

maintain a stable percept. And every stimulation pulse saturates the recording amplifiers, which is

modulated to keep a percept stable. Every stimulation pulse also saturates the recording amplifiers,

why bidirectional devices need artefact-rejection schemes that discard part of the signal they were

which is why bidirectional devices need artefact-rejection schemes that discard part of the signal

built to collect.

they were built to collect.

Two approaches attack the specificity limit. Biomimetic stimulation shapes pulse trains to imitate

the dynamics of natural afferent firing rather than delivering constant-amplitude trains, which

the dynamics of natural afferent firing instead of delivering constant-amplitude trains, which makes

improves the naturalness of evoked touch. [[optogenetics]] achieves genuine cell-type specificity in

evoked touch more natural. [[optogenetics]] achieves genuine cell-type specificity in animals but

animals but requires gene delivery and implanted light sources; no human neuroprosthesis uses it.

requires gene delivery and implanted light sources, and no human neuroprosthesis uses it.

> [!key] Why this asymmetry matters

> A motor prosthesis that reads well and writes badly is still useful, because vision substitutes for

> A motor prosthesis that reads well and writes badly is still useful, because vision substitutes

> missing touch. A sensory prosthesis has nothing to substitute for a bad write channel, which is

> for missing touch. A sensory prosthesis has nothing to compensate for a poor write channel. That

> why hearing and vision restoration have plateaued at low information rates while motor decoding

> is why hearing and vision restoration have plateaued at low information rates while motor decoding

> has kept improving.

## Limitations and failure modes

Chronic implants fail in predictable ways. Penetrating electrodes provoke a foreign-body response:

microglia activate, astrocytes encapsulate the shank, and neurons retreat from the recording tip, so

signal yield declines over months to years. Insulation absorbs water and delaminates; thin metal

signal yield declines over months to years. Insulation absorbs water and delaminates, thin metal

traces corrode; hermetic feedthroughs leak. Percutaneous connectors, still standard in academic

traces corrode, and hermetic feedthroughs leak. Percutaneous connectors, still standard in academic

systems, are an infection route and tether the user to a laboratory.

Beyond the hardware, three limits recur. Decoders drift as the recorded population changes,

Three further limits recur beyond the hardware. Decoders drift as the recorded population changes

requiring recalibration. Almost all headline performance is measured with a technician present in a

and need recalibration. Almost all headline performance is measured with a technician present in a

controlled setting, and degrades outside it. And the number of independent control or perceptual

controlled setting, and it degrades outside that setting. Finally, the number of independent control

dimensions a person can use saturates well below the electrode count, for the same reason in every

or perceptual dimensions a person can use saturates well below the electrode count, for the same

modality: current spread makes nearby channels non-independent.

reason in every modality: current spread makes nearby channels non-independent.

## Ethics and access

Implanted neurotechnology raises questions that ordinary medical devices do not. Recorded neural

activity supports inferences the user did not intend to disclose, the concern formalized under

[[mental-privacy]] and addressed legislatively under [[neurorights]]. Devices that modulate mood or

[[mental-privacy]] and addressed in legislation under [[neurorights]]. Devices that modulate mood or

motivation raise questions about authorship of action. And a device implanted for therapy in a

motivation raise questions about who authors an action. A device implanted for therapy in a person

person who cannot easily have it removed creates an asymmetric dependence on the manufacturer's

who cannot easily have it removed also creates an asymmetric dependence on the manufacturer's

continued existence — an exposure demonstrated when retinal implant support was withdrawn from

continued existence. That exposure became concrete when retinal implant support was withdrawn from

several hundred patients.

Access follows wealth rather than need. Cochlear implantation rates differ by more than an order of

magnitude between high- and low-income countries, and every newer device is more expensive. The

magnitude between high- and low-income countries, and every newer device costs more. The

distributional pattern examined in [[access-and-inequality]] is not hypothetical here; it is the

distributional pattern examined in [[access-and-inequality]] is already the current state of the

current state of the most successful neuroprosthesis ever built.

most successful neuroprosthesis ever built, not a hypothetical.

Where restoration shades into [[human-enhancement]] is a live boundary rather than a settled line.

The point where restoration shades into [[human-enhancement]] is contested rather than settled.

Devices such as those catalogued in [[sensory-augmentation]] use the same interfaces to add

capacities rather than restore them, and speculative uses such as

capacities rather than restore them. Speculative uses such as [[brain-to-brain-interface]]

[[brain-to-brain-interface]] communication borrow the vocabulary of clinical neuroprosthetics while

communication borrow the vocabulary of clinical neuroprosthetics while resting on far weaker

resting on far weaker demonstrations.

demonstrations.

## Outlook

Two trends look likely to shape the next decade. Materials and packaging are improving faster than

algorithms — thin-film flexible arrays, carbon-fibre electrodes, and fully implanted wireless

algorithms. Thin-film flexible arrays, carbon-fibre electrodes and fully implanted wireless systems

systems address the specific failure modes that have limited chronic recording, and commercial

address the specific failure modes that have limited chronic recording, and commercial programmes at

programmes at [[neuralink]], [[synchron]], and others are funding that engineering at a scale

[[neuralink]], [[synchron]] and others are funding that engineering at a scale academic laboratories

academic laboratories never could. Miniaturized wireless nodes of the kind described under

never could. Miniaturized wireless nodes of the kind described under [[neural-dust]] would change

[[neural-dust]] would change the placement calculus if they can be made to work at cortical depth.

the placement calculus if they can be made to work at cortical depth. Further from the clinic,

Further from the clinic, [[artificial-neurons|bio-interfacing artificial neurons]] aim to make the

[[artificial-neurons|bio-interfacing artificial neurons]] aim to make the stimulating electronics

stimulating electronics themselves fire like nerve cells; printed devices of this kind have

themselves fire like nerve cells. Printed devices of this kind have stimulated Purkinje neurons in

stimulated Purkinje neurons in slices of mouse cerebellum, but none has been reported implanted in a

slices of mouse cerebellum, but none has been reported implanted in a living brain.[^hadke2026]

living brain.[^hadke2026]

The competing trajectory is biological. Regenerative approaches under

The competing trajectory is biological. The regenerative approaches covered under

[[limb-regeneration]] and cell replacement therapies would remove the need for an interface

[[limb-regeneration]], along with cell replacement therapies, would remove the need for an interface

entirely in some indications, and external assistive devices such as the

entirely in some indications, and external assistive devices such as the [[exoskeleton]] avoid

[[exoskeleton]] avoid implantation altogether. Which path wins depends on an unresolved empirical

implantation altogether. Which path prevails depends on an unresolved empirical question: whether

question: whether the write-channel problem yields to better encoding and better electrodes, or

the write-channel problem yields to better encoding and better electrodes, or whether artificial

whether artificial stimulation is permanently limited to coarse, unnatural percepts that the brain

stimulation is permanently limited to coarse, unnatural percepts that the brain can learn to use but

can learn to use but never to interpret as its own.

never to interpret as its own.

## See also