NeuraViPeR · Project

Brain-Implant Technology That Could Restore Sight to Blind People

healthTestedTRL 4Thin data (2/5)

Imagine a tiny camera feeding what it sees directly into the brain's visual center through ultra-thin electrodes — basically a bionic eye that bypasses damaged eyes entirely. Current devices can only talk to a handful of brain cells and wear out in months. This team built flexible probes thinner than a human hair with 1024 channels that stay stable long-term, plus smart software that learns how each person's brain responds and adjusts the signal in real time. Think of it as upgrading from a fuzzy 8-pixel display to something that could let blind people recognize faces and walk through unfamiliar rooms.

By the numbers

1024

individually addressable electrode channels in the validated implant system

128

channels per stackable CMOS probe shaft

~10 µm

probe thickness — thinner than a human hair

< 50 µm

probe width for minimal tissue damage

maximum number of stackable probe assemblies

consortium partners across 6 countries

The business problem

What needed solving

Over 35 million people worldwide are blind, and current visual prosthetics can only activate a handful of brain cells through electrodes that degrade within months — delivering an image quality so poor it is barely functional. Companies developing brain-computer interfaces and neural implants are stuck at a technology ceiling where electrode count, longevity, and smart signal processing have not kept pace with demand.

The solution

What was built

The team built a 1024-channel flexible electrode implant system validated in living tissue with long-term stability, stackable CMOS probes with 128 channels each (up to 4 stackable), and a closed-loop deep learning system that converts camera footage into brain stimulation patterns — demonstrated in a working VR-based simulator.

Audience

Who needs this

Neural implant manufacturers (e.g., companies competing with Neuralink, Blackrock Neurotech)Assistive technology companies for the visually impairedCMOS/semiconductor firms seeking biomedical microelectronics contractsOphthalmology device companies exploring cortical bypassesNeuromorphic computing hardware companies needing validated use cases

Business applications

Who can put this to work

Medical Devices & Neurotechnology

enterprise

Target: Companies developing brain-computer interfaces or neural implants

If you are a neurotechnology company struggling with electrode longevity and channel count limitations — this project built a 1024-channel flexible probe system validated in vivo that remains stable over time. The probes are just 10 micrometers thick and under 50 micrometers wide, causing minimal tissue damage. This could directly feed into your next-generation implantable device pipeline.

Assistive Technology & Rehabilitation

mid-size

Target: Manufacturers of assistive devices for visually impaired people

If you are an assistive technology company looking beyond retinal implants for people with total vision loss — this project created a closed-loop system with deep learning algorithms that extract the most relevant visual information from a camera feed. The software recognizes objects, facial expressions, and navigation cues, translating them into brain stimulation patterns. This represents a platform you could license for cortical visual prosthetics.

Semiconductor & Microelectronics

any

Target: CMOS chip designers specializing in biomedical applications

If you are a microelectronics company seeking new markets in implantable devices — this project developed stackable CMOS multi-shaft 128-channel probes with flexible interconnects ready for implantation. Up to 4 probes can be stacked, reaching 512 channels per assembly. This micro-fabrication approach could open contract manufacturing opportunities in the growing neural interface market.

Frequently asked

Quick answers

What would it cost to license or adopt this technology?

The project was publicly funded under FET Open (frontier research), so IP terms would need to be negotiated with the consortium led by Universitat Zurich. Given the early-stage nature and academic-heavy consortium (5 universities, 2 research organizations), licensing terms are likely flexible. Expect significant additional R&D investment before a commercial product.

Can this scale to industrial manufacturing?

The CMOS-based stackable probe design (128 channels per probe, stackable up to 4) was specifically engineered with scalability in mind using standard semiconductor fabrication methods. However, the system has been validated in vivo in research settings, not in clinical manufacturing environments. Scaling to medical-grade production would require regulatory-compliant manufacturing processes.

What is the IP situation and licensing path?

As a Horizon 2020 RIA project, intellectual property is owned by the consortium partners who generated it. The consortium spans 6 countries (BE, CH, DE, ES, NL, SE) with 8 partners, so IP may be distributed across multiple institutions. Contact the coordinator at Universitat Zurich for licensing discussions.

What regulatory hurdles remain before market entry?

As a Class III implantable medical device stimulating the brain, this technology faces the most stringent regulatory pathway — FDA PMA in the US, MDR CE marking in Europe. Based on available project data, clinical trials in blind patients have not yet been conducted. Regulatory approval could take 5-10 years from the current stage.

How does this compare to existing visual prosthetics like Second Sight?

Existing cortical visual prosthetics use tens of electrodes and degrade within months. NeuraViPeR demonstrated a 1024-channel system with long-term in vivo stability, representing an order-of-magnitude improvement in channel count. The closed-loop deep learning approach that adapts to brain responses is also a significant differentiator over static stimulation systems.

What is the realistic timeline to a usable product?

The project ran from September 2020 to February 2025 and delivered working prototypes validated in animal models. Based on available project data, human clinical trials would be the next major milestone. A realistic timeline to a commercial product would be 7-12 years, factoring in clinical trials, regulatory approval, and manufacturing scale-up.

Can the algorithms and software be used separately from the implant hardware?

Yes. The project developed deep learning algorithms that transform camera footage into stimulation patterns, plus neuromorphic hardware for low-latency processing. The VR-based phosphene vision simulator can be used independently for research. These software components could potentially be licensed separately for other brain-computer interface applications.

Consortium

Who built it

The NeuraViPeR consortium is heavily academic, with 5 universities and 2 research organizations forming the core, plus just 1 industry partner (which is also the sole SME). The 12% industry ratio signals this is still a research-driven effort, not yet transitioning to commercialization. The geographic spread across 6 European countries (Belgium, Switzerland, Germany, Spain, Netherlands, Sweden) brings diverse regulatory and clinical expertise. The coordinator is Universitat Zurich, a top-tier research institution in neuroscience. For a business partner looking to engage, the key question is which specific partner owns the probe fabrication IP versus the algorithms — the value chain likely spans multiple consortium members.

UNIVERSITAT ZURICHCoordinator · CH
KONINKLIJKE NEDERLANDSE AKADEMIE VAN WETENSCHAPPEN - KNAWparticipant · NL
STICHTING RADBOUD UNIVERSITEITparticipant · NL
INTERUNIVERSITAIR MICRO-ELECTRONICA CENTRUMparticipant · BE
ALBERT-LUDWIGS-UNIVERSITAET FREIBURGparticipant · DE
UNIVERSIDAD MIGUEL HERNANDEZ DE ELCHEparticipant · ES
CHALMERS TEKNISKA HOGSKOLA ABparticipant · SE

How to reach the team

Universitat Zurich (Switzerland) — look for the principal investigator in computational neuroscience or neural engineering departments

Order a report on this consortium See UNIVERSITAT ZURICH profile →

Next steps

Talk to the team behind this work.

Want an introduction to the NeuraViPeR team to discuss licensing the 1024-channel probe technology or the adaptive deep learning algorithms? SciTransfer can arrange a targeted meeting with the right consortium partner for your specific interest.

Order a report on this topic More in Health & Biomedical