Virtual worlds, ChatGPT, and AI seem to be all the rage in 2024, with all sorts of developments shaking up not just the technology space, but having wider implications for medicine, politics, and even the judicial system. Now, researchers from Swiss startup FinalSpark have combined a virtual world with tiny human mini brains and built a two-way connection.
The team has been working on a virtual world controlled by brain organoids using a system called Neuroplatform. It’s thought to be the world’s first wetware cloud platform and allows researchers to interact with the brain organoids remotely. These organoids, dubbed “mini brains”, are only pea-sized, but are made up of around 10,000 neurons derived from human induced pluripotent stem cells.
The 16 organoids that the system runs on are kept in incubators at 37 °C (98.6 °F) and can respond to stimuli through a multi-electrode array (MEA); MEAs are often used in neuroscience to measure the electrical activity of neurons. It
To demonstrate that the mini brains can respond to stimulation, the team created a virtual world containing a butterfly. If a living human clicks in the virtual world, the software works out whether the click was in the butterfly’s field of view, then the mini brains respond and tell the butterfly to fly towards the location where the click occurred within the virtual world. Otherwise, the butterfly flies randomly in the virtual space.
“It’s crucial to emphasize that while these movement functions are implemented in software, the decision to use one or the other is driven by the brain organoid’s response to stimulation,” explained Daniel Burger, a research and development engineer working with FinalSpark on the Neuroplatform project, in a blog post.
Instead of running software on traditional computing, Burger also said that you could run software on the mini brains as biological neural networks (BNN). There are several potential advantages to this idea, including that BNNs have “significantly lower energy consumption” compared to supercomputers.
“We currently compare CPU vs. organoid processing directly,” Burger told The Register. “We still have energy consumption from supporting hardware like incubators and electrical stimulation systems. We don’t yet have an exact 1:1 comparison of CPU vs. organoid including all supporting hardware, but this is something we’re working on quantifying in the near future.”
The paper detailing Neuroplatform is published in Frontiers In Artificial Intelligence.
[H/T: The Register]
