Scientists have combined a machine Xi, an artificial intelligence system, with tiny 3D models of the brain. （*blackjack3d via getty images）
To improve the computing power of artificial intelligence (AI), researchers combine ordinary machine Xi with a complex 3D model of the human brain made from different types of brain tissue grown in the lab.
These miniature models of the brain, known as cerebral organoids or "mini-brains," have been around in various forms since 2013. But they have never been used as a way to augment artificial intelligence.
The new study uses more traditional computational hardware to feed electrical data into organoids, which then decipher the organoids' activity to produce outputs – so organoids only serve as an "intermediate layer" of the computational process.
While this approach is far from mimicking the real structure or workings of the brain, it may provide an early step in creating biological computers that would draw on the tricks of biology to make them more powerful and energy-efficient than traditional computers. It also provides a deeper understanding of how the human brain works and how it is affected by neurodegenerative diseases such as Alzheimer's and Parkinson's.
In the new study, published Monday (Dec. 11) in the journal Nature Electronics, researchers used a technique called reservoir calculations;In this case, organoids act as a "reservoir". In such a system, the reservoir stores information and reacts to the incoming information. The algorithm Xi recognizes changes triggered in the reservoir by different inputs and then converts those changes into their outputs.
Using this framework, the researchers inserted brain organoids into the system by providing them with electrical input delivered through electrodes.
Basically, we can encode information, such as image or audio information, into temporal and spatial patterns of electrical stimuli," said Feng Guo, co-author of the study and associate professor of intelligent systems engineering at Indiana University Bloomington.
In other words, the response of organoids depends on the temporal and spatial distribution of the current from the electrodes. The algorithm learned to interpret the organoid's electrical response to this stimulus.
While the cerebral organoid is much simpler than the actual brain – it is essentially a cerebellar cell ball – it has some adaptability and can change in response to stimuli. The responses of different types of brain cells, cells at different stages of development, and brain-like structures in organoids roughly mimic the way our brains respond to electrical signals. This change in the brain stimulates our ability to learn and Xi.
Using this unconventional hardware, the researchers trained their hybrid algorithm to accomplish two types of tasks: one related to speech recognition and the other to mathematics. In the former, the computer recognizes Japanese vowels from hundreds of audio samples with about 78% accuracy. It is fairly accurate at solving mathematical tasks, but slightly inferior to traditional types of machine learning Xi.
This study marks the first time that brain organoids have been used with artificial intelligence, but previous studies have used simpler types of lab-grown neural tissue in a similar way. For example, scientists have intertwined brain tissue with a form of strong chemical Xi that may Xi more similar to the way humans and other animals Xi than reservoir calculations.
Lena Smirnova, an assistant professor of environmental health and engineering at Johns Hopkins University, said future research could try to combine brain organoids with intensive chemical Xi, co-authoring a review of the new study.
One of the advantages of creating a biocomputer is energy efficiency, as our brains use much less energy than today's advanced computing systems. But Smirnova said it could be decades before technology like this could be used to make a general-purpose biocomputer.
While organoids can't replicate mature human brains yet, Smirnova hopes the technology will give scientists a better understanding of how the brain works, including for diseases like Alzheimer's. For example, replicating the structure (organoids) and functions (computational) of the brain can give researchers a better understanding of how the structure of the brain relates to Xi and cognition.
As with organoids in general, these computing systems are expected to help replace drug testing in animals, which Smirnova added raises both ethical questions and not always yielding useful results, as animals are very different from humans. The inclusion of organoids derived from human brain tissue in drug testing may help close this gap.