The Nobel That Turned Neurons Into Light Switches

On October 5, the Nobel Assembly at Karolinska Institutet awarded the 2026 Nobel Prize in Physiology or Medicine to Karl Deisseroth of the Howard Hughes Medical Institute and Stanford, 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 12 million Swedish kronor is shared equally.

Deisseroth, reached by the Associated Press an hour after the call, summed up his night: "No real sleep was achieved."

For anyone who thinks about computers, the prize is easy to translate. For most of the 20th century, neuroscience could observe the brain and damage it, but it could not cleanly write to it. Optogenetics gave researchers a way to send an on or off command to a chosen type of nerve cell, in a living animal, on a millisecond clock. That is the closest biology has come to treating a circuit as something you can address.

A curious question about algae

The story starts with a question that had no obvious payoff. In the early 1990s at the Max Planck Institute for Biochemistry in Martinsried, Hegemann wanted to know how Chlamydomonas, a single-celled green alga, reacts so fast to light. According to the Nobel committee's popular account, the alga fired an electrical impulse just half a millisecond after light hit its eyespot. Human vision runs through a chain of chemical steps that takes at least 10 milliseconds.

Hegemann proposed that a single protein both caught the light and opened the channel. Colleagues were sceptical. Around the turn of the millennium, Japanese researchers published a large set of Chlamydomonas gene sequences, and Hegemann's group spotted two that resembled known light-capturing proteins.

He sent them to Nagel at the Max Planck Institute for Biophysics in Frankfurt. Nagel injected the genes into frog eggs, which produced the proteins on their surface. Lit up, they opened as ion channels. The hypothesis was right. The pair named them channelrhodopsin-1 and channelrhodopsin-2, and found the second opened within 0.2 milliseconds. In a 2003 paper they put channelrhodopsin-2 into human and hamster kidney cells, which became light-sensitive, and proposed it as a tool for generating electrical signals in cells with light.

The engineer's move

Deisseroth supplied the leap from curiosity to instrument. A psychiatrist as well as a neuroscientist, he had been searching for a protein that could trigger nerve impulses on command. He wrote to Nagel asking for the DNA, put it into cultured rat neurons, and found they fired when hit with blue light. The 2005 paper, in Nature Neuroscience, had a title that reads like a spec sheet: "Millisecond-timescale, genetically targeted optical control of neural activity."

The two key words are genetically targeted. An electrode stimulates whatever tissue is near its tip. Channelrhodopsin can be delivered only to the cell types a researcher chooses, so light switches on those cells and leaves their neighbours alone. In a brain of around 90 billion neurons, where cells with unrelated jobs sit side by side, that selectivity is the whole game.

The method was named optogenetics in 2006. In 2007, Deisseroth's group put channelrhodopsin-2 into a specific cell type in the mouse motor cortex, fed a thin optical fibre through a small hole in the skull, and moved the animal's whiskers with light. They titled that paper "An optical neural interface." The same year, with collaborators, they woke sleeping mice by lighting up a cell type suspected of controlling wakefulness.

The experiment people remember came in 2012, with Susumu Tonegawa. The team tagged the neurons that formed a fear memory in mice, then reactivated them later. The mice showed fear in a safe setting. The Nobel committee calls it the first demonstration of exactly which nerve cells are needed for a specific memory.

From observation to causation

That is the real shift being honoured. Older methods linked brain regions to functions, but, as the committee puts it, they could not prove that a particular cell type caused a feeling or behaviour. Optogenetics turned correlation into intervention. Researchers have since mapped circuits for pain, thirst, eating, reward, attention and social behaviour, and found that even parenting in mice splits into separate circuits for gathering pups and grooming them.

Where it meets neurotech

The obvious question for a BCI-curious reader is whether this reaches people. The honest answer: slowly, because every use requires gene therapy first.

The clearest human result is in vision. In a 2021 Nature Medicine paper led by José-Alain Sahel and Botond Roska, a man blinded by retinitis pigmentosa received a gene for ChrimsonR, a channelrhodopsin variant that responds to amber light, in his retina's ganglion cells. Camera goggles turned the scene into amber light projected onto the retina. With the goggles on, he touched a large notebook in 92% of trials and counted glass tumblers correctly 63% of the time, according to Sorbonne University's release. Without them, he could do neither. An EEG decoder detected when a tumbler was present with 78% accuracy. One patient is a proof of concept, not a treatment, and the authors said as much.

Hearing is next in line. A Göttingen spin-off, OptoGenTech, co-founded by Tobias Moser, is preparing the first clinical study of an optical cochlear implant that would make auditory nerve cells light-sensitive and stimulate them with a planned 64-channel device. The bet, in the Nobel committee's words, is that light could activate the auditory nerve "with greater precision" than today's electrical implants. We found no confirmation that a human trial has begun.

Then there is the BCI frontier proper. Science Corporation, founded by Neuralink co-founder Max Hodak, describes a biohybrid brain interface in which embedded neurons are modified with optogenetics "so they can be activated with light," flanked by "hundreds of thousands of microLEDs." It is early-stage, but it shows where the idea leads: an interface that writes to tissue through light, not current.

Why it matters now

Electrode-based BCIs have become very good at reading the brain. Writing to it with precision remains much harder. Optogenetics is now used in laboratories around the world, and the committee's own account says it has sharpened understanding of depression, anxiety, schizophrenia, Alzheimer's and Parkinson's.

The 2026 prize is a reminder of where the capability came from: not a moonshot brain programme, but a researcher asking why a green alga swims toward light. The switch was sitting in pond water. It took three people and more than a decade to wire it into a neuron.