To boost the computing power of artificial intelligence, researchers have combined run-of-the-mill machine learning with a sophisticated 3D model of the human brain. Miniature models of the brain, known as cerebral organoids or "minibrains", have existed in various forms since 2013. The research uses traditional computing hardware to input electrical data into the organoid and then decipher the organoid's activity to produce an output. The method is far from mimicking the structure of the brain or how it works, but it may provide an early step toward creating bio computers, which would be more powerful and energy efficient than traditional computers. It could lead to more understanding of how the human brain works and how it is affected by disease such as Alzheimer's and Parkinson's. Scientists propose that lab-grown minibrains will be used as biological hardware.
The researchers used a technique called "reservoir computing" to do the new study. In that system, the reservoir stores information and responds to it with different inputs and outputs. The brain organoid was plugged into the system using the framework. According to the study co-author, the information can be converted into a temporal-spatial pattern of electrical stimulation. The organoid's electrical responses to stimulation are determined by timing and location. The brain organoid is simpler than an actual brain and can adapt and change in response to stimulation. Our brains change in response to electrical signals, but there are different types of brain cells. The researchers trained their hybrid algorithm to complete two types of tasks, one related to speech recognition and another to mathematics. It was fairly accurate in the math task, but less so than traditional types of machine learning. For example, scientists have interwoven brain tissue with a form of reinforcement learning, a type of machine learning that is similar to how humans and other animals learn. A commentary about the new study said that future research could attempt to combine brain organoids with reinforcement learning. Since our brains use less energy than today's advanced computing systems, it would make sense to create biocomputers. While organoids aren't close to replicating full-blown human brains, Smirnova hopes the technology will give scientists a better understanding of how the brain works. Researchers could better understand how the brain's structure is related to learning and cognitive functions by replicating it with organoids and computing. As with organoids in general, these computing systems could hopefully help replace drug testing in animals, but ethics issues and not always yield useful results, because animals differ so much from humans. Drug testing using organoids derived from human brain tissue could help close the gap.