Live brain cells playing pong in a bowl could shed light on the mechanics of the mind

Scientists have created a gamer – from cells, in a laboratory.

An Australian-led research team placed 800,000 live human and mouse brain cells in a dish, connected them to electrodes and a simulation of the classic game of pong. The scientists then watched as the mini-mind quickly taught itself the game and improved with practice. They were able to participate by converting the cellular reactions into a visual representation of the game that looks very similar to the original.

They call their system DishBrain and say it proves that neurons can learn in a bowl and show basic signs of intelligence. The team describes the new setup, dubbed synthetic biological intelligence, or SBI, in a study published Wednesday in the journal Neuron.

Finally, the authors say, SBI could help solve long-standing problems Secrets of Brain Mechanics and lead to better treatments for certain neurological disorders. “DishBrain offers a simpler approach to testing how the brain works and gaining insight into debilitating conditions like epilepsy and dementia,” said Hon Weng Chong, chief executive officer of biotech start-up Cortical Labs.

SBI could also offer an alternative to animal testing, which scientists often use to investigate the feasibility of new drugs and therapies.

“We now have essentially the ultimate biomimetic ‘sandbox’ in which to test the effects of drugs and genetic variants – a sandbox made up of exactly the same (neural) computational elements found in your brain and mine,” adds co-author Professor Karl Friston, a theoretical neuroscientist at University College London.

Artificial vs. biological intelligence

The study team found that biological intelligence, also known as living brain cells, behaves quite differently than a computer when it comes to AI.

“In the past, models of the brain have been developed that were aimed at how computer scientists think the brain might work,” says Brett Kagan, chief scientific officer of Cortical Labs and co-author of the study. “That’s usually based on our current understanding of information technology, like silicon computing… But the truth is, we don’t really understand how the brain works.”

Interestingly, DishBrain learned to play Pong out of an apparent tendency to interact with its surroundings in ways that make it more predictable and less random. In other words, this system behaves much more like a real living brain than the AI.

For example, when DishBrain successfully returned the “ball” in Pong, the system was better able to predict where it would move next. If DishBrain failed, it would lose the point and a new point would begin with the computer releasing a ball from a random starting spot, and so on. Since DishBrain uses a feedback loop, it seems to get better and better the more it plays.

“This is remarkable because you can’t teach this kind of self-organization simply because these mini-brains – unlike a pet – don’t have a sense of reward and punishment,” adds Friston.

Now, Cortical Labs, an Australian biotech startup, is working on a new generation of biological computing chips to create a generalized form of SBI that the team writes in their study, “coming ahead of artificial general intelligence because of inherent efficiency and evolution could benefit biological systems.”

“We know that our brains have the evolutionary advantage of being primed for survival for hundreds of millions of years,” explains co-author Adeel Razi of Monash University. “Now we seem to have it in our hands where we can harness this incredibly powerful and cheap biological intelligence.”

The researchers also tried the system on other simple games.

“You know when the Google Chrome browser crashes and you get this dinosaur you can jump over obstacles (Project Bolan),” says Kagan. “We’ve done that and we’ve seen some nice preliminary results, but we still have more work to do to create new environments for custom purposes.”

Next, the team plans to make DishBrain have a good time.

“We’re trying to do a dose-response curve with ethanol — basically you get them ‘drunk’ and see if they play the game worse, just like when people drink,” says Kagan.

While we look forward to the results of the drunken DishBrain study, perhaps we should keep these drunken neurons far from any self-driving car code.