Sight restoration · Osaka, Japan
“The first thing I saw was a door. It was the wrong one”
After twenty years without sight, a former piano teacher was fitted with a retinal implant. What she sees now is not what she remembers, and she is clear about the difference.
Yoshida Haru taught piano in Sakai for twenty-two years. Retinitis pigmentosa, an inherited disease that kills the light-sensing cells of the retina, took her sight gradually through her thirties, and by thirty-six she could tell light from dark and nothing else. She kept teaching for another decade, by ear and by touch. She is fifty-seven.
In April 2025 she became one of the people the Foundation pays each year to have a retinal implant: 287 in the year before hers. The device is a thin array of electrodes laid on the retina, connected to a camera in a pair of glasses. It turns what the camera sees into small pulses of electricity, and the remaining nerve cells of the retina pass them to the brain as points of light.
Points of light
“People imagine you open your eyes and there is the world,” she says. “It is not like that. It is like a scoreboard in a stadium with most of the bulbs gone.” The implant has 1,500 electrodes. A healthy retina has about a hundred million light-sensing cells.
What the brain makes of the pattern has to be learned. For six months after the operation, Ms Yoshida went twice a week to a rehabilitation centre in Osaka to practise: finding the edge of a table, following a white line on the floor, telling a doorway from a wall. The centre’s measure of success, agreed before her operation, was whether she could find a doorway unaided at six months.
The first thing I saw was a door. It was the wrong one. I was so pleased I walked straight into it, and it was a cupboard.
Six months
At six months she found the right door eight times out of ten. She can see where a window is, and where people are standing if they are against a light background, though not who they are. She can follow the lines of a pedestrian crossing. She still uses her cane, because the implant does not show her steps, and she has no plans to stop.
Asked whether it was worth it, she takes a long time to answer. “It is worth it for me,” she says. “Ask me again in five years. And ask the people for whom it did not work. Not everyone can find the door.” Of the 287 people whose implants we funded in FY2024, 59.3% met the doorway measure at six months.
Whose device
Ms Yoshida’s implant is an AKARI, made by Kiroshi Opticals, the company the Foundation’s founder started and in which it is the largest shareholder. Seventy-one per cent of the retinal implants we funded last year were AKARI devices; only three companies make one approved in the countries where we work. Decisions on retinal implant funding are made by a clinical panel of five people, none of whom works for or owns shares in any implant maker, and they chose her centre, not her device.
The implant’s software can also describe what the camera sees in words, read aloud through an earpiece. She has turned this off, and finds the room’s piano by walking to it.