"Neuron" observes living human brain cells in a petri dish and learns to play table tennis

This is a visual representation that simulates ping-pong, with neuronal activity reflected on
growing bricks.

Living biological neurons are more exemplary of how the brain works
than artificial intelligence.

For the first time, scientists have shown that 800,000 brain cells living in a dish can perform goal-directed tasks
.
In this case, they play a simple tennis-like computer game table tennis
.
The results of the Melbourne-led study were published today (12 October) in the journal Neuron
.
Now, researchers will investigate what happens
when their "brain organoids" are affected by drugs and alcohol.
(In 1972, Atari in the United States "PONG" simulated two people playing table tennis, that is, there was a point moving in the middle of the two lines
.
) ）

"We have shown that we can interact with living biological neurons in this way, forcing them to modify their own activities to produce something like intelligence," said
lead author Dr.
Brett Kagan.
He is the chief scientific officer of biotech startup Cortex Lab, which is dedicated to manufacturing a new generation of biocomputer chips
.
His co-authors are affiliated with Monash University, RMIT University, University College London and the Institute for Advanced Study of
Canada.

Microscopic images of nerve cells, with fluorescent markers showing different types of cells
.
Green represents neurons and axons, purple represents neurons, red represents dendrites, and blue represents all cells
.
When multiple markers appear, the colors are merged and usually appear yellow or pink
, depending on the scale of the markers.

"DishBrain provides an easier way to test how the brain works and gain insight into debilitating diseases like epilepsy and dementia," said
Dr.
Hon Weng Chong, chief executive of Cortex Labs.

While researchers have been able to mount neurons on multi-electrode arrays and read their activity, this is the first time cells
have been stimulated in a structured and meaningful way.

Kagan said: "In the past, brain models were developed
based on how computer scientists thought the brain might work.
This is often based on our current understanding
of information technology, such as silicon computing.
But in fact, we don't really understand how the brain works
.
”

The video shows that the game of ping pong is controlled
by a layer of neurons in a petri dish.

By constructing a living brain model from basic structures in this way, scientists will be able to experiment
using real brain function rather than flawed similar models like computers.
For example, Kagan and his team will conduct the next experiment to see what effect
alcohol has when it is introduced into brain organoids.
"We tried to create a dose-response curve with ethanol — basically getting them 'drunk' and seeing if they played worse, like people do when they drink
," Kagan said.

This could pave the way
for entirely new ways to understand what's happening in the brain.

Scanning electron microscope images
of neural cultures grown over 6 months on a high-density multi-electrode array.
Some nerve cells grow on the periphery, forming a complex network that covers the electrodes
in the center.
Credit Corticulation Laboratory

Dr Adeel Razi said: "This new ability to teach cell cultures to perform a task – to control the return of the racket through induction – opens up new possibilities of discovery that will have profound implications
for technology, health and society.
" He is the Director
of the Laboratory of Computational and Systems Neuroscience at Monash University.
"We know that our brains have the evolutionary advantage of being tuned
over hundreds of millions of years to survive.
Now, we seem to have mastered how to harness this incredibly powerful and inexpensive biological intelligence
.
”

When studying how new drugs or gene therapies respond in these dynamic environments, the findings also raise the possibility of
creating an alternative to animal testing.

"We also showed that we can modify the stimulus based on how cells change their behavior and do this in a real-time closed loop
," Kagan said.

To carry out the experiment, the team of scientists collected mouse cells extracted from embryonic brains and some human brain cells extracted from stem cells
.
They placed them on arrays of microelectrodes, which both stimulated them and read their activity
.

Electrodes on the left and right sides of an array are lit to tell the organoids which side the brain ball is on, while the distance from the racket is indicated
by the signal frequency.
Feedback from the electrodes tells the dish organoids how to put the ball back into the ball by making the cells move
like a racket.
Kagan said: "We have never seen how cells work
in a virtual environment before.
We have succeeded in creating a closed-loop environment that can read what's happening in the cells, stimulate them with meaningful information, and then change the cells in an interactive way so they can really change each other
.
”

Professor Karl Friston, a theoretical neuroscientist at UCL and co-author, said: "The beauty and groundbreaking of this work is that it gives neurons sensory – feedback – and, crucially, the ability to
act on the world around them.
Remarkably, these cultures learned how to make their world more predictable
through action.
This is notable because you can't teach this kind of self-organization; That's just because — unlike pets — these mini-brains don't have a sense of
reward and punishment.
The translational potential of this work is really exciting: it means we don't have to worry about creating 'digital twins' to test therapeutic interventions
.
" In principle, we now have the ultimate biomimetic 'sandbox' where we can test the effects of drugs and genetic variants – a sandbox made up of
the exact same computational (neuron) elements found in your and my brains.
”

This study also supports Professor Friston's "free energy principle"
.

"When we were studying how to direct cells along a specific pathway, we faced a challenge
.
We don't have direct access to the dopamine system or anything else we can use to provide concrete real-time incentives, so we have to dig deeper into Professor Freeston's research: information entropy—a fundamental level of information about how systems self-organize and interact with the environment on a physical level
.
The free energy principle states that cells at this level minimize unpredictability
in the environment.
”

One exciting finding is that organoids don't behave
like computer-based systems.
"When we present structured information to non-solid neurons, we see a change in their activity, which is very consistent
with their actual behavior as a dynamic system," he said.
For example, over time, a neuron's ability to change and adapt to its own activities increases with experience, consistent with
what we see at cellular learning rates.
”

Chong said he's excited about the discovery, but it's just the beginning
.
"It's a whole new virgin land
.
We hope that more people will join and collaborate with them to further explore this new area of
science using the systems we have built.
As one of our collaborators put it, it's not every day that you wake up to create a new field of
science.
”

Reference: In vitro neurons learn and exhibit sentience when embodied in a simulated game-world