Mouse brain cells in a lab dish learn to play Pong

Playing Pong, an old video game similar to tennis, is what they have achieved a group of neurons in a lab dish, showing that they can display inherent intelligence and modify their behavior over time.

Magazine Neruon publishes the details of this experiment collected in an international study led by the University of Melbourne (Australia).

“We have shown that we can interact with living biological neurons in such a way that we force them to modify their activity, which leads to something that resembles intelligence”said lead author Brett Kagan of the biotech company Cortical Labs.

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

To carry out the experiment, the team took mouse cells from embryonic brainsas well as some human brain cells derived from stem cells, and grew them on microelectrode arrays, which could both stimulate them and read their activity.

The neurons were connected to a computer so that they received information about whether their in-game paddle was hitting the ball. In addition, it allowed their activity and responses to this feedback to be monitored using electrical probes that recorded “spikes” on a grid.

The spikes grew stronger the more a neuron moved its paddle and hit the ball, but when they failed their play was criticized by a software program, showing that neurons could adapt activity to a changing environment in a game-oriented way. the objectives, in real time, explains the magazine.

The theory behind this learning is based on the free energy principle, whereby the brain adapts to its environment by changing its world view or its actions to better fit the world around it.

“Surprisingly, the crops learned to make their world more predictable by acting on it. This is remarkable because this kind of self-organization cannot be taught; simply because – unlike a pet – these mini brains have no sense of reward and punishment.” noted Karl Friston of University College London.

“Never before have we been able to see how cells act in a virtual environment” and now it has been achieved by building a closed-loop environment that can read what is happening in cells, stimulate them with meaningful information and “then change them interactively so they can actually be altered,” Kagan said.

In the past, models of the brain have been developed based on how computer scientists think the brain might work, but we don’t really understand how it works, the researcher was quoted as saying by Cortical Labs.

By building a model of brain alive from basic structures in this way, scientists will be able to experiment using real brain function.

This ability to teach cell cultures to perform a task for which they show sensitivity – by controlling the paddle to return the ball through sensing – opens up new possibilities for discovery that will have far-reaching consequences for technology, health and society. , according to another of the signatories, Adeel Razi, from Monash University (Australia),

Future direction of this work has potential in disease modeling, drug discovery, and expanding the current understanding of how the brain works and how intelligence arises.

The findings also raise the possibility of creating an alternative to animal testing when investigating how new drugs or gene therapies respond in these dynamic settings.