For generations, the eye was understood as a fixed instrument — a biological lens that simply gathered light and handed it off to the brain for interpretation. A new study from King's College London and the Friedrich Miescher Institute, conducted in zebrafish, quietly dismantles that assumption: the retina itself, it turns out, listens to the world and reshapes itself accordingly. In the earliest days of visual life, the eye is not merely a receiver but a learner — and what it learns may stay with an animal long after the lesson ends.
Early visual experience reshapes retinal cells, challenging decades of neuroscience dogma
The retina had literally reshaped itself
Mark
So the retina was supposed to be locked in by the time an animal is born. What changed that assumption?
Mimi
The assumption was based on what we knew about spontaneous activity—waves of neural firing that happen before an eye ever sees light. Those waves seemed to do all the developmental work. But this study shows that actual visual experience, after birth, can still reshape the retina's cells. The cells aren't just receiving instructions from activity patterns; they're responding to what the animal is actually seeing.
Mark
The amacrine cells got longer or shorter depending on the stripes. That's a physical change, not just a functional one.
Mimi
Exactly. And that's what makes it so striking. It's not just that the cells fire differently. The cells themselves change shape. A cell oriented parallel to vertical stripes becomes elongated; one oriented perpendicular becomes rounder. The retina is literally reorganizing its own real estate.
Mark
But why would that matter? Why would a cell need to change shape?
Mimi
That's the open question now. Changing shape likely changes how the cell connects to other cells, how signals flow through it, what information it can encode. A longer cell might reach farther, make different connections. The shape change is probably a window into something deeper about how the cell is rewiring itself.
Mark
The fish that were raised with horizontal stripes didn't prefer horizontal stripes later. That seems backwards.
Mimi
It does seem counterintuitive. The researchers aren't sure why yet. But the fish that were raised with vertical stripes did prefer vertical stripes. So there's something asymmetrical about how the plasticity works, at least in this setup. That's another mystery to solve.
Mark
What happens if you delete the gene that makes this plasticity possible?
Mimi
The fish can still see the stripes—they respond to them normally. But they lose the ability to adapt. They don't develop the preference, and presumably their retinal cells don't reshape. The gene isn't about vision itself; it's about the capacity to learn from what you see.
Mark
Does this matter for humans?
Mimi
We don't know yet. But the basic architecture of the retina is shared across vertebrates. If it's true in zebrafish, it might be true in us. And if the retina is plastic during early development, that could change how we think about visual disorders, about critical periods, about what shapes how we see the world.
The Pulse
- A foundational belief in neuroscience — that the retina is hardwired and only the brain adapts to experience — has been directly contradicted by experimental evidence in zebrafish.
- Amacrine cells in the retina physically changed shape depending on whether larvae were raised seeing horizontal or vertical stripes, with some cells elongating and others rounding in response to their visual environment.
- The structural changes were not cosmetic: they biased the fish's visual output toward the orientations they had experienced, and the bias held even after the animals were moved to a neutral setting.
- Behavioral tests confirmed the effect — fish raised in vertical stripe environments showed measurable orientation preferences, while those with a deleted plasticity gene lost the ability to adapt entirely.
- Researchers and outside experts now face an open question: if zebrafish retinas reshape themselves through experience, the same mechanism may operate in mice, rabbits, and humans — with implications for how visual systems across vertebrates are formed.
For generations, the eye was understood as a fixed instrument — a biological lens that simply gathered light and handed it off to the brain for interpretation. A new study from King's College London and the Friedrich Miescher Institute, conducted in zebrafish, quietly dismantles that assumption: the retina itself, it turns out, listens to the world and reshapes itself accordingly. In the earliest days of visual life, the eye is not merely a receiver but a learner — and what it learns may stay with an animal long after the lesson ends.
For decades, neuroscientists told the story of visual development as one that begins in the eye but unfolds almost entirely in the brain. The retina, it was assumed, was fixed — shaped before birth by spontaneous waves of neural activity, then locked in place once real vision began. Its job was to process incoming light and pass the signal along. The learning happened elsewhere.
A new study in zebrafish challenges that picture at its foundation. Researchers at King's College London and the Friedrich Miescher Institute raised zebrafish larvae in channels lined with either horizontal or vertical black-and-white stripes during the first five days after fertilization. They then examined amacrine cells — retinal interneurons known to play a key role in detecting oriented visual features, particularly those expressing a protein called teneurin-3.
What they found was unexpected: the cells physically changed shape based on what the fish had seen. Amacrine cells oriented parallel to the stripes the fish experienced became more elongated; those oriented perpendicular became rounder. The retina had literally reorganized itself, and the effect was functional — the fish's visual output to the brain became biased toward the orientations they had been raised with. That bias persisted for at least two days after the animals were moved to a neutral environment.
Behavior tracked the biology. Fish raised in vertical stripe environments preferred swimming toward stripes that ran parallel to their bodies. Fish raised in horizontal stripes showed no such preference. When the teneurin-3 gene was deleted entirely, fish raised in horizontal environments behaved normally but lost the ability to adapt — the cellular machinery for plasticity had been removed.
Marla Feller, a neuroscientist at UC Berkeley not involved in the study, described the finding as significant precisely because it overturns a widespread assumption: that visual activity has no effect on the retina itself. This is the first demonstration that what an animal actually sees can physically alter retinal interneurons and shift behavior downstream.
Because vertebrates share the same fundamental retinal architecture, the finding opens the possibility that experience-dependent retinal plasticity is not unique to zebrafish. Whether the same mechanisms operate in mammals — and what cellular signals translate visual patterns into structural change — are now the questions the field must pursue.
For decades, neuroscientists have understood visual development as a story told mostly in the brain. Light enters the eye, photoreceptors fire, and then the real work of learning happens downstream—in the cortex, where circuits rewire themselves based on what an animal sees. The retina itself was thought to be locked in place, a fixed biological camera that simply delivered its signal and nothing more.
A new study in zebrafish is upending that assumption. Researchers at King's College London and the Friedrich Miescher Institute have shown that the retina itself is plastic, reshaping its own cells in response to early visual experience. The finding challenges a foundational belief in neuroscience and suggests that even the eye's most basic sensory structures adapt to the world they encounter.
The work builds on decades of elegant research. Hubel and Wiesel's Nobel Prize-winning experiments showed that kittens raised seeing only vertical stripes develop cortical neurons tuned to vertical orientations—and those raised with horizontal stripes tune to horizontals instead. The brain learns what it sees. But the retina, everyone assumed, was different. It was hardwired before birth, shaped by spontaneous waves of neural activity that ripple through developing eyes before an animal ever opens them to light. Once vision began, the retina's job was simply to process and pass along information.
The new research suggests otherwise. Zebrafish larvae were raised in V-shaped channels with either horizontal or vertical black-and-white stripes on the walls for the first five days after fertilization. The researchers focused on amacrine cells—interneurons in the retina that play a crucial role in detecting oriented features of the visual world. These cells come in many varieties, and the ones expressing a protein called teneurin-3 are known to be essential for establishing orientation selectivity, the ability to prefer certain angles of visual input.
What happened was striking. Amacrine cells oriented parallel to the stripes the fish saw became noticeably more elongated, while cells oriented perpendicular to the stripes became rounder. Since these cells are distributed evenly across the retina, the shape changes meant that some cells were now occupying more physical space than others. The retina had literally reshaped itself. And the effect was functional: the fish's visual output to the brain became biased toward the stripe orientation they had been raised with. Even after the animals were moved to a neutral environment, this bias persisted for at least two days.
Behavior followed. Fish given a choice between stripes of different angles preferred swimming toward stripes that ran parallel to their bodies—but only if they had been raised in a vertical stripe environment. Fish raised in horizontal stripes showed no such preference. And when researchers deleted the teneurin-3 gene entirely, fish raised in a horizontal environment behaved normally, suggesting they could still see the stripes but had lost the ability to adapt to them. The cells had been disturbed in a way that prevented plasticity.
Marla Feller, a neuroscientist at UC Berkeley who was not involved in the work, called the finding significant because it overturns a widespread assumption in the field. "The field in general doesn't think that activity has any effect on the retina," she said. "It's always downstream." This study is the first to show that what an animal actually sees can prompt retinal activity that physically changes interneurons and alters behavior downstream.
The implications ripple outward. Vertebrates share fundamental retinal architecture—the same layers of interneurons, the same basic logic of how photoreceptors transform into feature-detecting channels. If zebrafish retinas reshape themselves based on experience, the same may be true in mice, rabbits, and potentially humans. The question now is what cellular mechanisms read the pattern of activity and decide to elongate a cell or round it out. And whether this kind of early retinal plasticity matters more in some animals than others—perhaps especially in fish, which begin exploring their environments far earlier in development than mammals do.
Notable Quotes
The field in general doesn't think that activity has any effect on the retina, and it's always downstream.— Marla Feller, UC Berkeley neuroscientist
You can raise animals in a very specific environment, and their retinal circuits have now changed so that they're better for that environment.— Alexandre Tiriac, Vanderbilt University
Want the full story?
Read the original at The Transmitter ↗
Education