Journal · Science

Karl Deisseroth and optogenetics: the 2026 Nobel explained

For a century neuroscience could watch the brain closely and change it only bluntly. Optogenetics made it possible to switch one kind of nerve cell on and off in milliseconds. The 2026 Nobel Prize in Physiology or Medicine recognised the discoveries behind it, which began with a green alga that swims toward light.

Highland Willow · Boone, North Carolina · 2 November 2026

Optogenetics is a method for switching chosen nerve cells on or off with light. A gene for a light-sensitive protein is delivered only to the cells you want to study. When light reaches them, the protein moves charged particles across the cell's surface: one kind opens a channel that makes the cell fire, another pumps chloride in and silences it. Cells without the protein do nothing.

On 5 October 2026 the Nobel Assembly at Karolinska Institutet announced its decision to award the 2026 Nobel Prize in Physiology or Medicine jointly to Karl Deisseroth of the Howard Hughes Medical Institute and Stanford University, Peter Hegemann of Humboldt University of Berlin and Georg Nagel of the University of Würzburg, "for their discoveries concerning light-gated ion channels and optogenetics". The prize is 12 million Swedish kronor, shared equally.

The short version, that scientists can now control brain cells with light, is true but undersells it. Optogenetics did not let researchers see anything new. It let them prove things they had only been able to suspect, and that is a different kind of progress.

A century of blunt instruments

Twentieth century neuroscience was very good at finding out which regions of the brain were involved in what. Electrodes recorded which cells fired during a task. Damage to a region showed what was lost without it. Drugs nudged whole systems up or down. Scanners showed which areas lit up. The committee's scientific background is clear about the limit of all of it: these methods established correlations, and lacked the precision to act on one type of cell, inside an intact circuit, at the speed the brain actually works.

The difficulty is how the brain is packed. An adult human brain has around 90 billion nerve cells, each with thousands of connections. Cells that do opposite jobs sit side by side, and a cell's long fibre can run to a completely different region. Any method that acts on a region acts on all of those cells at once, opposites included, and averages them. The committee's popular background describes the resulting picture of the brain as a blurry photograph.

Francis Crick, of the DNA double helix and a 1962 Nobel Prize, spent his later career on the brain and wrote down what would be needed. You would want to activate or silence one type of nerve cell and leave the rest alone, in an awake animal, quickly. In 1999 he guessed that "the ideal signal would be light", and admitted the idea sounded far-fetched.

Why the early attempts fell short

People tried. In 1971 a laser was used to stimulate sea slug neurons directly, and in 1983 leech neurons were stained with a dye that responded to laser light. Both approaches were abandoned because the intense light and the dyes damaged the cells.

In 2002 Gero Miesenböck's laboratory made mammalian neurons light sensitive with genes for the first time, by installing three proteins borrowed from the fruit fly's eye. It worked, as a proof of concept. But it needed several genes delivered together and a light-catching molecule supplied alongside them, and it was slow, because the light signal passed through a chain of intermediate chemical steps inside the cell before anything opened, and each step blurred the timing. Another approach, from Ehud Isacoff, Richard Kramer, Dirk Trauner and colleagues, bolted a synthetic light-sensitive chemical onto an engineered channel. That one was fast, but the chemical had to be delivered into the tissue from outside.

The committee lists the properties a practical tool needed: deliverable as a gene, compact, reliably placed in the cell membrane, fast, reversible, and driven by light directly. Read that list against the failures and the pattern is plain. Light strong enough to work alone damaged cells, every extra component was something else to deliver, and every extra step between the light and the response was time in which the signal could smear. What was wanted was one protein that did the whole job.

A green alga that swims toward light

The protein came from a question nobody had asked for medical reasons. In the early 1990s Peter Hegemann, at the Max Planck Institute for Biochemistry in Martinsried, was studying Chlamydomonas, a single-celled green alga that swims toward light. It senses light with an eyespot, and when Hegemann measured the electrical response he found it arrived about half a millisecond after the light did.

That is fast. In the human eye, light sets off a chain of chemical steps that ends at an ion channel, and the whole process takes at least 10 milliseconds. The alga was more than twenty times quicker. Hegemann could think of only one explanation: the process in the alga had to be far simpler. He proposed that a single protein complex both captured the light and acted as the channel. No such thing was known, and the idea was met with scepticism.

Proving it took years, because the eyespot proteins fell apart when extracted. The break came around 2000, when Japanese researchers published a catalogue of the alga's genes. Hegemann's group found two that looked like the genes for known light-capturing proteins. The committee's scientific background notes that two other groups, one of them John Spudich's, identified the same sequences independently.

To find out what the proteins did, Hegemann turned to Georg Nagel at the Max Planck Institute for Biophysics in Frankfurt, who specialised in a useful trick: put a protein's genetic instructions into frog eggs and they will manufacture the protein and place it in their surface, where it can be measured. In 2002 Nagel showed that the first protein was itself an ion channel that opened when lit. Hegemann's hypothesis, a decade old by then, was right. It was named channelrhodopsin-1.

The second, published in 2003, turned out to matter more. Channelrhodopsin-2 opened to blue light within 0.2 milliseconds and let positively charged ions flow into the cell. Nagel and Hegemann put its gene into human kidney cells as well, and those became light sensitive too. They proposed it as a tool for triggering electrical signals in cells. Even a fragment made of just its first 315 amino acids does the job.

From frog eggs to a living brain

Karl Deisseroth was a psychiatrist starting his own laboratory at Stanford. The committee's account says his clinical training convinced him that understanding the illnesses he saw meant studying nerve cells in a living brain, not in slices in a dish, and that the poor treatments he could offer made an idea others thought unrealistic worth the risk of failing. He wrote to Nagel and asked for the DNA.

He worried that the foreign gene would harm the delicate rat neurons. It did not appear to. They produced channelrhodopsin-2 without apparent problems, and pulses of blue light made them fire, reliably, on a millisecond timescale. The 2005 paper reporting it was written by Edward Boyden, Feng Zhang, Ernst Bamberg, Georg Nagel and Deisseroth. One detail turned a promising result into a practical tool: Nagel observed that the light-catching molecule the protein needs, retinal, did not have to be added. The neurons already had enough of their own. One gene was the whole kit.

After that it moved fast. The method was named optogenetics in a 2006 review. In 2007 Deisseroth's group fed a thin optical fibre into the motor cortex of living mice and moved their whiskers with light. The same year, with Luis de Lecea, they showed that lighting up one type of cell suspected of controlling wakefulness made sleeping mice more likely to wake. Two groups, one in Deisseroth's lab and one led by Xue Han and Ed Boyden, added an off switch the same year: a chloride pump from an archaeon that silences cells under yellow light, which meant one circuit could be switched on with blue light and off with yellow.

Writing, not reading

The committee's scientific background ends by describing the change as a move from reading brain activity to writing into it. That is the right summary, and the reason it matters is logical rather than technical.

If a cell fires whenever a mouse is afraid, you know the two go together. You do not know whether the cell causes fear, responds to fear, or merely shares a cause with it. With optogenetics you can ask both questions that settle it. Switch the cell on in a calm animal: if fear appears, the cell's activity is sufficient. Switch it off in a frightened one: if the fear goes, it is necessary.

The toy below runs all three approaches on one invented patch of brain. Record first, then stimulate the whole patch, then switch to optogenetics and run both tests on each type of cell.

The best known example came in 2012. Working with Susumu Tonegawa's laboratory, researchers labelled the specific cells that were active while a mouse formed a fear memory, then later switched just those cells on with light in a different, harmless place. The mouse showed fear. A follow-up in 2013 went further: by reactivating the memory of a safe place while the mouse was being frightened elsewhere, researchers made mice afraid of a place where nothing bad had ever happened to them. Switching on a particular set of cells, it turned out, is enough to bring a particular memory back.

The other recurring discovery is that regions are not units. Dopamine cells in one small area of the midbrain were shown to drive reward or aversion depending on which other area was sending them signals: the same neighbourhood, opposite effects. In mice, caring for young turned out to be several separate circuits, one for gathering pups into the nest and another for grooming them. Region-level methods could never have separated those, because they average them.

Six names, three medals

The Nobel Prize has a hard ceiling. Section 4 of the Statutes of the Nobel Foundation says: "In no case may a prize amount be divided between more than three persons." Optogenetics had more contributors than that, and the Nobel documents themselves name many of them.

Ernst Bamberg, under whom Nagel did his doctorate, is an author on the 2002, 2003 and 2005 papers that the popular background lists as key publications. Edward Boyden is the first author of the 2005 paper. Gero Miesenböck made neurons light sensitive with genes first, in 2002, and is a co-author of the review that named the field. In 2013 the Brain Prize, from the Grete Lundbeck European Brain Research Prize Foundation, recognised optogenetics and was shared six ways: Bamberg, Boyden, Deisseroth, Hegemann, Miesenböck and Nagel.

We do not know why the line fell where it did. The decision belongs to the Nobel Assembly at Karolinska Institutet, and section 10 of the statutes keeps the material behind it closed for at least 50 years. Our own reading, for what it is worth: a prize is a summary of a body of work, and summaries have a word limit. The scientific background, linked below, names many of the people involved, and it is worth reading for that alone.

What it has done for medicine so far

Mostly, optogenetics is a research tool. The committee credits it with sharpening the understanding of depression, anxiety, schizophrenia, Alzheimer's disease and Parkinson's disease. Understanding is not treatment, and the committee does not claim it is.

The first human results are in the eye. In 2021 a team led by José-Alain Sahel and Botond Roska reported a single patient, blind from retinitis pigmentosa, a disease that destroys the light-sensing cells of the retina. One eye was injected with a modified virus carrying the gene for ChrimsonR, a red-shifted channelrhodopsin, and the patient wore goggles that project pulses of light onto the retina. With the goggles, using the treated eye, the patient could perceive, locate, count and touch objects. Without them, the patient could not detect objects at all.

On 8 October 2026, three days after the prize was announced, the same group reported a larger study in the New England Journal of Medicine. It treated 10 people blind from advanced retinitis pigmentosa, one injection each in the worse-seeing eye, in an open-label study whose primary outcome was safety. Measured with the goggles, light sensitivity increased in 7 of the 10, and by enough to count as clinically meaningful in 6; the paper notes that this threshold came from the literature and was not set in the protocol in advance. There were 34 eye-related adverse events among 9 of the participants. One was severe: a blockage of the retina's central artery straight after the injection, which cleared within minutes with eye drops. The authors conclude that, within the limits of a small study, the treatment was safe, and that more research is needed on safety and efficacy. The study was funded by GenSight Biologics, which is developing the treatment, and it is not approved anywhere.

It is probably not a coincidence that the first human use is in the eye. Our reading, not the committee's: the retina is the one part of the nervous system that light already reaches. Everywhere else, optogenetics needs two deliveries, a gene by virus and light by an implanted fibre, which is routine in a laboratory mouse and a very different proposition in a person. The committee also mentions hopes for cochlear implants that stimulate the hearing nerve with light rather than electricity, for greater precision. Those are hopes, not results.

What a business-systems firm takes from this

We are not neuroscientists. We work on the operations and ERP systems of owner-operated businesses, and the part of this story that transfers to that work is not the light. It is why the early tools fell short.

The committee gives several reasons. Light strong enough to act on its own damaged cells. Some systems needed extra ingredients delivered alongside the genes, or a synthetic chemical delivered from outside. And the slower ones relayed the signal through intermediate steps, which loosened the link between the light and the response until the timing could not be trusted. Two of those have close equivalents in changes to how a business runs. A change that only works if three other things also happen is a multi-component system. A change whose effect arrives through several departments, and is judged weeks later in a margin figure, is a slow relay. In our experience, which is not a count, both are common, and when either is present something may move without anyone being able to say what moved it.

Optogenetics is a standard for an intervention you can learn from, and it fits in four words: targeted, single, fast, reversible. Change one thing, for one defined part of the operation, measure something that responds within days rather than quarters, and be able to switch it back.

An invented example, with round numbers. A distributor suspects that the order in which its pick lists are printed is causing late shipments. The dashboard shows the two moving together, which is a correlation. So it changes the sort order in one warehouse only, for two weeks, and counts late shipments each day. Then it switches the old order back for two weeks, then the new one on again. If late shipments fall, rise and fall again in step with the switch, that is evidence no report can give you: the change is sufficient, and removing it undoes it. If nothing moves, that is weaker evidence, because a small effect can hide in a noisy fortnight, but it was cheap to find out, in one building, before rolling anything out to all of them.

What this does not establish

Every statement about the science here is read from the Nobel Assembly's three documents for 2026 (the press release, the popular science background and the scientific background) and from the Statutes of the Nobel Foundation. The two vision studies are described from their published abstracts. We did not read the full papers, nor the 2002, 2003 and 2005 papers, which are described here as the Nobel documents describe them.

This is not medical advice. Optogenetics is not an approved treatment for any condition in any of these sources, and on 8 October 2026 Retinal Physician reported that the vision therapy has no marketing authorisation in any jurisdiction. The vision results are one patient in 2021 and ten people in a small, open-label safety study in 2026. Anyone living with retinitis pigmentosa or any other condition mentioned here should be talking to their own specialist, not reading a business firm's journal.

The section on credit reports who wrote which paper and whom another prize included, and marks its one piece of interpretation as ours. It does not say the Assembly chose wrongly, and we have no way of knowing its reasons.

The circuit toy is invented. Its three types of cell, and the numbers behind them, illustrate the logic of the two tests and describe no real circuit.

The business section is an analogy, and an analogy is not evidence. A warehouse is not a brain. People notice when their process changes and adjust to it, which mice with an optical fibre do not, so an on and off test in a business is noisier than one in a laboratory and needs longer periods to read. The narrow claim is that two of the reasons the early tools fell short, components that had to arrive together and steps that blurred the timing, also make changes to a business hard to learn from.

Questions people ask

What is optogenetics?

Optogenetics is a method for switching specific nerve cells on or off with light. A gene for a light-sensitive protein is delivered only to the chosen cells. When light reaches them, the protein moves charged particles across the cell's surface: a channelrhodopsin from algae opens a channel that makes the cell fire, and a light-driven chloride pump silences it. Cells without the protein are unaffected, which is what makes it precise.

Who won the 2026 Nobel Prize in Physiology or Medicine?

Karl Deisseroth of the Howard Hughes Medical Institute and Stanford University, Peter Hegemann of Humboldt University of Berlin and Georg Nagel of the University of Würzburg, announced on 5 October 2026, for their discoveries concerning light-gated ion channels and optogenetics. Hegemann and Nagel discovered channelrhodopsin in a green alga; Deisseroth turned it into a light switch for nerve cells. The prize of 12 million Swedish kronor is shared equally.

What is channelrhodopsin?

A protein from the single-celled green alga Chlamydomonas that is both a light sensor and an ion channel. When blue light hits channelrhodopsin-2, a channel through it opens within about 0.2 milliseconds and positively charged ions flow into the cell. Because one protein does the whole job, and nerve cells already hold the light-catching molecule it needs, one gene is enough to make a nerve cell respond to light.

Is optogenetics used to treat people?

Not as an approved treatment. The furthest along is vision. A 2021 report described one blind patient with retinitis pigmentosa, and on 8 October 2026 the New England Journal of Medicine published a 10-person open-label safety study in which light sensitivity, measured with light-projecting goggles, rose in 7 participants and by a clinically meaningful amount in 6. The authors say more research is needed. In the brain itself, optogenetics remains a research tool. This is not medical advice.

Who else contributed to optogenetics?

Many people, and the Nobel documents name them. Ernst Bamberg co-wrote the key 2002, 2003 and 2005 papers. Edward Boyden was first author of the 2005 paper. Gero Miesenböck first made neurons light sensitive with genes, in 2002, and co-wrote the 2006 review that named the field. John Spudich's group and a group led by Takahashi found the algal genes independently. The 2013 Brain Prize for optogenetics was shared by six people. A Nobel Prize can be shared by at most three, and the material behind each decision stays closed for at least 50 years.

What is the difference between optogenetics and chemogenetics?

Both switch chosen cells on or off. Optogenetics uses light and acts within milliseconds, matching the speed of nerve signals. Chemogenetics uses engineered receptors that respond only to a designer drug, which can be given by injection and acts more slowly over a wider area. The Nobel committee describes them as complementary.

Sources

1. The Nobel Assembly at Karolinska Institutet. Press release: The 2026 Nobel Prize in Physiology or Medicine, 5 October 2026. Read 2 November 2026. https://www.nobelprize.org/uploads/2026/10/press-medicineprize2026.pdf
2. The Nobel Assembly at Karolinska Institutet. Popular science background: A light-sensitive algal protein energised neuroscience. Source of the timings, the alga's story, the 2012 memory experiment and the list of key publications. Read 2 November 2026. https://www.nobelprize.org/uploads/2026/10/popular-medicineprize2026.pdf
3. The Nobel Committee for Physiology or Medicine. Scientific background 2026: Optogenetics, discovery of a neuronal switch. Source of the early attempts, the 2006 naming, the circuit findings and the Crick quotation. Read 2 November 2026. https://www.nobelprize.org/uploads/2026/10/advanced-medicineprize2026.pdf
4. Crick F. The impact of molecular biology on neuroscience. Philosophical Transactions of the Royal Society B 354, 2021 to 2025 (1999), as quoted in the scientific background. https://pubmed.ncbi.nlm.nih.gov/10670022/
5. Statutes of the Nobel Foundation, sections 4 and 10, read 2 November 2026. https://www.nobelprize.org/about/statutes-of-the-nobel-foundation/
6. Federation of European Neuroscience Societies. Six eminent scientists share the world's largest brain research prize, on the 2013 Brain Prize. https://fens.org/news-activities/news/six-eminent-scientists-share-the-worlds-largest-brain-research-prize
7. Sahel JA, Boulanger-Scemama E, Pagot C, et al., Roska B. Partial recovery of visual function in a blind patient after optogenetic therapy. Nature Medicine 27, 1223 to 1229 (2021). PubMed record read 2 November 2026. https://pubmed.ncbi.nlm.nih.gov/34031601/
8. Sahel JA, Futterknecht S, Audo I, et al., Roska B. Optogenetic therapy for restoring aspects of visual function. New England Journal of Medicine 395, 1399 to 1408 (8 October 2026), PIONEER, NCT03326336. Abstract read on PubMed 2 November 2026. https://pubmed.ncbi.nlm.nih.gov/42842903/
9. Retinal Physician. GS030 shows potential to restore visual function in retinitis pigmentosa, 8 October 2026, for the marketing authorisation status. https://retinalphysician.com/news/2026/gs030-shows-potential-to-restore-visual-function-in-retinitis-pigmentosa/