Industries / Frontier science / Brain-computer interfaces
Implants
Implants are the electrodes and electronics placed in or on the brain to pick up the electrical signals of brain cells and carry them out of the body.
7
Milestones achieved, from 1991 to 2024.
3
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1991
A bed of 100 needles
A University of Utah team described how to make a silicon base 4.2 mm square with 100 needles, each 1.5 mm long and tipped with platinum, sharp enough to pass into the cortex, the brain's outer layer. This design became known as the Utah array. A 2021 study says the field of implanted brain-computer interfaces in people "currently relies on" it.
Credit
P. K. Campbell, K. E. Jones, R. J. Huber, K. W. Horch and Richard Normann
1998
Wires brain fibres grow into
Philip Kennedy and Roy Bakay implanted an electrode whose tip coaxes nerve fibres to grow into it in a patient with ALS (amyotrophic lateral sclerosis, a disease that paralyses the muscles) who was "locked in": aware but unable to move or speak. It recorded brain cell signals for several months, and the patient learned to turn them on and off. In 2000 the team reported a second patient who used the same kind of electrode to move a cursor and to type and produce synthetic speech. A 2024 review in Nature Reviews Bioengineering places that patient's implant at the Georgia Institute of Technology in 1998, after a brainstem stroke, and calls the patient "the first person with long-term implants who was able to control a computer cursor using brain signals".
Credit
P. R. Kennedy and R. A. E. Bakay
2016
Nothing crossing the skin
At UMC Utrecht, surgeons slid four strips of electrodes onto the surface of the brain of a patient with late-stage ALS through holes 1 cm across, and wired two strips to a recorder and radio transmitter placed under the skin below the collarbone. Both parts were Medtronic products made for other uses: the strips for spinal cord stimulation, the recorder for deep brain stimulation. With no wires leaving the body, the patient typed at home at about two letters a minute, 28 weeks after the operation.
Credit
Mariska Vansteensel, Nick Ramsey and colleagues at UMC Utrecht, with Medtronic
2019
Wireless recorders in the skull
At Clinatec in Grenoble, France, two wireless recorders, each with 64 electrodes resting on the dura (the tough membrane around the brain), were implanted over the movement areas of a 28-year-old man with tetraplegia, paralysis of all four limbs. Over 24 months he used them to move an avatar at home and an exoskeleton in the lab.
Lessonthe study recruited two people, and one was excluded because of a technical problem with the implants.
Credit
Alim-Louis Benabid, Stéphan Chabardès and colleagues, Clinatec and Grenoble University Hospital
2021
How long Utah arrays last
A study of 55 Utah arrays in 17 macaque monkeys and 2 people, drawn from over 6,000 recordings across almost nine years, found that usable recordings lasted 622 days on average. Some arrays lasted over 1,000 days and one lasted 9 years. Arrays in people lasted longer than in monkeys, and tips coated with iridium oxide kept more working electrodes than platinum tips.
Lessonthe authors conclude that recordings can be expected to last at least two years.
Credit
Caleb Sponheim, Jennifer Collinger, Nicholas Hatsopoulos and colleagues
2023
A film on the brain
At West Virginia University's Rockefeller Neuroscience Institute, the neurosurgeon Peter Konrad laid Precision Neuroscience's Layer 7 array, an electrode film thinner than a hair, on the brains of three patients who were already having tumour surgery. It stayed for 15 minutes in each and recorded about 1,000 channels at once while two patients were awake and speaking. Konrad told CNBC that standard surface electrodes are about 4 mm across and that the film fits 500 to 1,000 contacts into that space.
Credit
Precision Neuroscience; Peter Konrad
2024
Threads that pulled back
Neuralink's N1 implant records through 1,024 electrodes spread along 64 flexible threads, each thinner than a hair. Its sealed body sits in an opening in the skull, runs on a battery charged through the skin, and sends data by radio. The first one went into Noland Arbaugh at Barrow Neurological Institute in Phoenix in January 2024. In the weeks after, a number of threads retracted from the brain, which cut the number of working electrodes and the speed of his cursor, until Neuralink changed its software. Neuralink later said that after changes to surgery and aftercare, 18 of the next 20 participants had higher signal quality.
LessonNeuralink says it is now testing mechanical features to keep threads in place.
Credit
Neuralink; Barrow Neurological Institute
Watch for
Coin-sized implant, ultra-thin wires
The Center for Excellence in Brain Science and Intelligence Technology (CEBSIT) of the Chinese Academy of Sciences and Huashan Hospital in Shanghai say they implanted a device 26 mm across and under 6 mm thick in a patient in March 2025. Its flexible electrodes are about one hundredth the width of a hair, and the team says it has run stably with no infection or electrode failure. The report comes from the team and state media.
Achieved whenpeer-reviewed results show stable single-cell recordings in several people for a year or more.
Claimed by
CEBSIT and Huashan Hospital, June 2025
421 electrodes, fully implanted
University of Michigan Health surgeons implanted Paradromics' Connexus device in a woman as the first participant of a US trial. It has 421 microelectrodes, and their signals travel to a transceiver in the chest that sends them to a receiver outside the body. She will be followed for six years.
Achieved whenpublished trial results show how many electrodes are still recording after a year.
Claimed by
Paradromics and University of Michigan Health, June 2026
From 1,000 to 3,000 electrodes
Neuralink says it is raising its implant from about 1,000 to 3,000 electrodes and testing mechanical features that hold threads in place.
Achieved whena 3,000-electrode device is in people and outside researchers or a regulator report how many of its electrodes keep recording over time.
Claimed by
Neuralink, January 2026
Open
01
Open soak-test data for films
A bench rig that soaks test pieces of the thin plastic films used to insulate flexible brain electrodes (polyimide, parylene C and liquid crystal polymer) in warm salt water, some with dilute hydrogen peroxide added to mimic the body's chemical attack on implants. It logs each piece's electrical leakage every hour until the piece fails, and publishes every reading.
The briefCredit · unclaimed02
A table of implant lifetimes
A public spreadsheet with one row per published study or device report of a brain implant used for a brain-computer interface. Each row gives the device, electrode type and count, how long the implants stayed in, the share of electrodes still recording at 6, 12 and 24 months where the authors report it, and why any were removed, each figure taken only from the published report and linked to it. It is refreshed every three months.
The briefCredit · unclaimed
Open 01 · the brief
Open soak-test data for films
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A bench rig that soaks test pieces of the thin plastic films used to insulate flexible brain electrodes (polyimide, parylene C and liquid crystal polymer) in warm salt water, some with dilute hydrogen peroxide added to mimic the body's chemical attack on implants. It logs each piece's electrical leakage every hour until the piece fails, and publishes every reading.
Why it is still open
Soaking at raised temperature to speed up aging is standard. A US Food and Drug Administration (FDA) laboratory published a seven-day protocol in hot hydrogen peroxide in 2015. A 2023 study then found that the usual rule, that each 10 °C rise doubles the rate of aging, overstated polyimide lifetimes: it measured a 2.65-fold difference in median lifetime between 37 °C and 57 °C where the rule predicts 4-fold, and proposed a factor near 1.47 per 10 °C. Results appear as papers, usually one material from one lab. A 2021 University of Bremen study shared the mask design of its comb-shaped polyimide test piece, but tested one material at one temperature (70 °C) and published curves, not raw logs. This research found no shared, open dataset of raw soak-test failures across materials.
pmc.ncbi.nlm.nih.govpmc.ncbi.nlm.nih.govpmc.ncbi.nlm.nih.gov
Built with what exists
Comb-shaped test patterns of interleaved metal tracks, ordered from a commercial flex-circuit maker in polyimide, with parylene C added by a coating service; water baths held at 37 °C, 57 °C and 87 °C; phosphate-buffered saline and dilute hydrogen peroxide; a Raspberry Pi with a multiplexer and an analogue-to-digital converter to log leakage current; Python with SciPy to fit Weibull curves, a standard statistical model of time to failure. An AI coding agent can write and test the logging and fitting code.
First step
Draw a comb test coupon in KiCad, order 60 in polyimide, and start one 57 °C bath with 10 pieces under a fixed voltage.
Done looks like
At least 30 pieces per material at each of three temperatures, run until at least half have failed, with raw logs, Weibull fits and the measured speed-up factor per 10 °C published on Zenodo and compared with the 1.47 figure from the 2023 study.
Open 02 · the brief
A table of implant lifetimes
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public spreadsheet with one row per published study or device report of a brain implant used for a brain-computer interface. Each row gives the device, electrode type and count, how long the implants stayed in, the share of electrodes still recording at 6, 12 and 24 months where the authors report it, and why any were removed, each figure taken only from the published report and linked to it. It is refreshed every three months.
Why it is still open
The pooled counts that exist cover one device or stop years ago. A 2020 systematic review counted 48 people with Utah arrays, the longest implanted for 1,975 days, up to its 2019 search. The 2021 longevity study covers Utah arrays only. A 2025 review tabulates how long implants of many kinds stayed in people, but not how many electrodes kept working. Newer devices from Neuralink, Synchron, Paradromics and CEBSIT report mostly in press releases and company posts, outside any table.
chestekresearch.engin.umich.edupmc.ncbi.nlm.nih.govbioelecmed.biomedcentral.com
Built with what exists
The Europe PMC and PubMed search interfaces (APIs) and Crossref to find papers; an AI model that reads PDFs to draft each row for a person to check against the source; a public GitHub repository holding the CSV file and a log of changes.
First step
Enter the figures from the 2020 review, the 2021 longevity and explant studies, and the 2023 BrainGate and SWITCH safety papers as the first 20 rows.
Done looks like
At least 40 rows covering at least 6 device families, every figure with a working source link, two refreshes three months apart, and a chart of the share of electrodes still recording against time for each family that reports it.