Scientists map part of a mouse's brain that's so complex it looks like a
galaxy
[April 10, 2025]
By LAURAN NEERGAARD
WASHINGTON (AP) — Thanks to a mouse watching clips from “The Matrix,”
scientists have created the largest functional map of a brain to date –
a diagram of the wiring connecting 84,000 neurons as they fire off
messages.
Using a piece of that mouse’s brain about the size of a poppy seed, the
researchers identified those neurons and traced how they communicated
via branch-like fibers through a surprising 500 million junctions called
synapses.
The massive dataset, published Wednesday by the journal Nature, marks a
step toward unraveling the mystery of how our brains work. The data,
assembled in a 3D reconstruction colored to delineate different brain
circuitry, is open to scientists worldwide for additional research – and
for the simply curious to take a peek.
“It definitely inspires a sense of awe, just like looking at pictures of
the galaxies,” said Forrest Collman of the Allen Institute for Brain
Science in Seattle, one of the project’s leading researchers. “You get a
sense of how complicated you are. We’re looking at one tiny part ... of
a mouse’s brain and the beauty and complexity that you can see in these
actual neurons and the hundreds of millions of connections between
them.”
How we think, feel, see, talk and move are due to neurons, or nerve
cells, in the brain – how they’re activated and send messages to each
other. Scientists have long known those signals move from one neuron
along fibers called axons and dendrites, using synapses to jump to the
next neuron. But there’s less known about the networks of neurons that
perform certain tasks and how disruptions of that wiring could play a
role in Alzheimer's, autism or other disorders.
“You can make a thousand hypotheses about how brain cells might do their
job but you can’t test those hypotheses unless you know perhaps the most
fundamental thing – how are those cells wired together,” said Allen
Institute scientist Clay Reid, who helped pioneer electron microscopy to
study neural connections.
With the new project, a global team of more than 150 researchers mapped
neural connections that Collman compares to tangled pieces of spaghetti
winding through part of the mouse brain responsible for vision.
The first step: Show a mouse video snippets of sci-fi movies, sports,
animation and nature.
A team at Baylor College of Medicine did just that, using a mouse
engineered with a gene that makes its neurons glow when they’re active.
The researchers used a laser-powered microscope to record how individual
cells in the animal’s visual cortex lit up as they processed the images
flashing by.
[to top of second column]
|
This photo provided by the Allen Institute shows, from left,
Associate Director of Informatics Forrest Collman, Data Analyst
Leila Elabbady and Senior Investigator Clay Reid reviewing neuron
reconstructions for the Machine Intelligence from Cortical Networks
project in Dec. 2024, in Seattle, Wash. (Jenny Burns/Allen Institute
via AP)
 Next, scientists at the Allen
Institute analyzed that small piece of brain tissue, using a special
tool to shave it into more than 25,000 layers, each far thinner than
a human hair. With electron microscopes, they took nearly 100
million high-resolution images of those sections, illuminating those
spaghetti-like fibers and painstakingly reassembling the data in 3D.
Finally, Princeton University scientists used artificial
intelligence to trace all that wiring and “paint each of the
individual wires a different color so that we can identify them
individually,” Collman explained.
They estimated that microscopic wiring, if laid out, would measure
more than 3 miles (5 kilometers). Importantly, matching up all that
anatomy with the activity in the mouse's brain as it watched movies
allowed researchers to trace how the circuitry worked.
The Princeton researchers also created digital 3D copies of the data
that other scientists can use in developing new studies.
Could this kind of mapping help scientists eventually find
treatments for brain diseases? The researchers call it a
foundational step, like how the Human Genome Project that provided
the first gene mapping eventually led to gene-based treatments.
Mapping a full mouse brain is one next goal.
“The technologies developed by this project will give us our first
chance to really identify some kind of abnormal pattern of
connectivity that gives rise to a disorder,” another of the
project's leading researchers, Princeton neuroscientist and computer
scientist Sebastian Seung, said in a statement.
The work “marks a major leap forwards and offers an invaluable
community resource for future discoveries,” wrote Harvard
neuroscientists Mariela Petkova and Gregor Schuhknecht, who weren’t
involved in the project.
The huge and publicly shared data “will help to unravel the complex
neural networks underlying cognition and behavior,” they added.
The Machine Intelligence from Cortical Networks, or MICrONS,
consortium was funded by the National Institutes of Health’s BRAIN
Initiative and IARPA, the Intelligence Advanced Research Projects
Activity.
All contents © copyright 2025 Associated Press. All rights reserved |
