Vietnamese scientist's new eye chip helps blind patient read

One day in early June, a testing room at the Zhejiang Provincial People's Hospital in China fell silent as a 40-year-old woman sat before a computer screen wearing dark glasses fitted with a camera.

She had lost her sight two decades earlier to retinitis pigmentosa, or RP, an inherited disease that gradually destroys the retina's light-sensing cells. RP affects about two million people worldwide, or one in every 4,000. There is currently no cure capable of fully restoring vision lost to the disease.

On the first day of the eyesight test, technicians had repeatedly asked the woman whether she could see anything. Her answer remained no. The engineering team left the hospital in silence. Nhan and his colleagues stayed up through the night rewriting the software that controlled the electrical signals sent to her retinal implant.

At 1:50 p.m. on June 4, on the second day of the test, a new character appeared on the screen. The woman suddenly tilted her head and said: "I see a very bright letter L."

Nhan sank into his chair and clasped his hands together. "It felt better than winning the lottery," he says.

The patient had undergone surgery to implant the chip on May 19, and the device was switched on June 2. What she saw was not a conventional image of the letter L on the screen; instead, 320 microscopic electrodes implanted at the back of her eye generated electrical stimulation that produced artificial points of light.

After 20 years without sight, her brain was beginning to learn how to interpret those signals.

 

 

Vietnamese scientist's new eye chip helps blind patient read

Tiến sĩ người Việt chế tạo chip võng mạc giúp người mù 15 năm đọc chữ sau hai ngày

Designing sesame seed-sized implant

Nhan, 45, graduated from the Ho Chi Minh City University of Technology before moving to Australia for doctoral studies at the University of Melbourne.

In 2008, the Australian government provided AUD48 million (US$34 million) to five research institutes and universities for a national project to develop an artificial retina that could restore some vision to blind people.

Nhan and fellow researcher Yang Jiawei worked on the project as part of a team designing the integrated circuits needed for the technology. The project ran for eight years before ending due to technical and financial challenges. Nhan and Yang continued pursuing the technology.

In 2014, they moved from Melbourne to Hangzhou, China, and founded startup Nanochap to continue developing the retinal implant. Human vision depends on photoreceptor cells in the retina converting light into electrical signals that travel to the brain. In people with RP, these cells gradually deteriorate. Vision is lost even though other retinal nerve cells, the optic nerve and the brain can remain capable of transmitting and processing signals.

Nhan's system is designed to bypass the damaged photoreceptors. A camera mounted on a pair of glasses captures what is in front of the wearer. A processor converts the image into electrical signals, which are transmitted wirelessly to the chip implanted at the back of the eye. Its 320 tiny electrodes then stimulate surviving retinal nerve cells, allowing signals to travel along the eye's existing neural pathways to the brain.

The device does not restore normal eyesight. Patients instead perceive flashes or points of light known as phosphenes. Their brains must learn how to assemble those signals into recognizable outlines, letters, or obstacles. The goal is therefore not to restore natural vision, but to provide enough visual information for patients to recognize basic shapes and obstacles and regain some independence in everyday life.

One of the biggest challenges was size.

An incision for the eye surgery can be no wider than about four millimeters, meaning the implant itself had to fit within an area measuring just 3x3 mm, roughly the size of a sesame seed. Within that space, the engineers had to fit 320 independently controlled stimulation channels, a coil for receiving power and a two-way data transmission system. The implant also had to communicate with equipment outside the body through about two centimeters of eye tissue.

The team used 65-nanometer semiconductor technology to shrink the circuitry. That created another problem: the smaller chip technology limited its operating voltage to 3.3 volts, compared with the 10-15 volts considered preferable for older generations of implants.

"We had to go through hundreds of rounds of trial and error," Nhan says.

At one point, electrodes ordered from overseas failed to meet the team's requirements for thickness and biological durability. The researchers pooled their own money to buy equipment and set up a workshop in Hangzhou to manufacture the three-dimensional electrodes themselves.

 

 

Vietnamese scientist's new eye chip helps blind patient read

Tiến sĩ người Việt chế tạo chip võng mạc giúp người mù 15 năm đọc chữ sau hai ngày

Trial that almost never was

Technical difficulties were compounded by financial pressure. After the Covid-19 pandemic, Nanochap began running short of funds. During a funding round this May, investors imposed a condition: further financing would depend on the system demonstrating that it could work in a human patient.

Then, on the day scheduled for the first clinical implant, the volunteer patient unexpectedly withdrew and did not arrive at the hospital.

"That was the most stressful period," Nhan says. "If the trial failed or could not proceed, the company would go back to making conventional medical chips and the dream of restoring vision would be over."

Several weeks later, the team persuaded the 40-year-old woman in Zhejiang to participate. Two days after recognizing the letter L, she was able to distinguish more letters ... C, H, R. Over a month later, she could use a pen to write the Chinese characters for "China," locate a computer mouse and walk along a white guide line placed on the laboratory floor.

China's National Medical Products Administration approved Nanochap's device for clinical trials in early 2026. So far, the implant has been placed in two patients, with preparations underway for a third procedure.

Tran Quoc Huy, a senior application engineer at chip design software company Cadence Design Systems Vietnam, says integrating 320 stimulation channels into a wireless 3x3 mm module represented a notable achievement in microelectronics.

But the findings remain preliminary because they are based on a very small number of patients, and the researchers need to monitor the implants for possible biological rejection and infection and independently peer-reviewed scientific studies are also needed before the device's overall safety can be assessed, he adds.

Vietnamese scientist Tran Nhan, co-founder of Nanochap. Photo courtesy of Nhan

Researchers around the world are pursuing several approaches to artificial vision.

Elon Musk's Neuralink's Blindsight project aims to implant electrodes directly into the brain's visual cortex but has yet to be tested in humans. Science Corporation's PRIMA system uses a retinal implant with 378 electrodes and has been tested in 40 patients. Its patients undergo three to four months of visual rehabilitation.

Nhan's team uses direct electrical-pulse encoding, which it says can shorten the time patients need to begin interpreting the signals to several days.

However, the limited number of electrodes also restricts the level of visual detail the system can produce. Nanochap aims to complete 30 clinical implant procedures this year before seeking approval to commercialize the system in 2027.

Nhan says the company expects the procedure to cost $15,000-22,000 if the technology reaches the market. After a session at the hospital, he often stays behind in the corridor reflecting on how far the technology has come, and how much work remains.

"Seeing someone recognize their first flash of light makes me happy, but there is still a very long way to go before implanting this chip can become a routine procedure," he says.