The miracle of "brain in a jar": brain cells in a petri dish learned to play games by themselves!

In 1981, the American philosopher Hilary Putnam put forward the idea of "brain in a jar" in the book "Reason, Truth and History": if someone cuts off a person's brain, puts it in a tank filled with nutrient solution, and connects the nerve endings of the brain to the computer, and the computer continues to send normal information feedback to the brain - in this way, all the movements and environments felt by the brain are the same , so can it still be aware of the situation it is in? And will you be a "brain in a jar"?

Fortunately, this thoughtful hypothesis is only a thought experiment
at this stage.
However, the latest breakthrough from the Australian research team is enough to surprise us, even a little scary
.
When 800,000 brain cells from human stem cells or mice gathered in a petri dish, they learned to play simple video games
.
The effort, dubbed DishBrain (meaning "brain in a plate"), attracted attention
when it went live late last year on a preprint site.
Now, the progress is officially published in the prestigious academic journal Neuron
.

Dr Brett Kagan, Chief Scientific Officer of biotech startup Cortical Labs, who led the study, commented: "Our study shows that we can communicate with active biological neurons, forcing them to change their behavior, resulting in behavior similar to intelligence
.
The researchers hope that the DishBrain model can make it easier to detect how the brain works, improve the understanding of symptoms such as epilepsy and dementia, and help the development of
related drugs and therapies.

In the study, a team of researchers from Cortical Labs and universities obtained brain cells from two sources: human induced pluripotent stem cells or mouse embryos
.
The team placed brain cells on an array of microelectrodes in a dish that could stimulate these cells with electrical signals and read neural activity
.

In the electrical stimulation signals applied to brain cells, there is a hidden main task: to teach these brain cells to play a video game
called Pong.

This game, born nearly 50 years ago, has very simple rules and can be understood as playing table tennis against the wall: by moving a line from side to side, that is, a "racket", hitting the flying ball and bouncing it back
.

▲ The neuron in the petri dish controls the racket and plays the Pong game (video source: Reference [1]).

Of course, the first thing the researchers have to do is to translate the rules into the "language" that brain cells can receive, and train this DishBrain system: by controlling the discharge stimulation of the left or right electrode, it can tell the direction of movement of the brain cell glomeruli; The distance between the ball and the racket is represented
by the signal frequency.
The signals emitted by DishBrain are monitored by the electrodes and translated into control
of the racket with the same rules.
In this way, these brain cells can learn rules in this closed-loop system and generate perception through
training.

"We've never seen these cells work
in a virtual environment.
We have successfully built a closed-loop system that reads cell activity
.
We activate them with meaningful information to change these cells in an interactive way so that they can change
each other.
Dr.
Kagan said
.