Like a title from a sci-fi horror movie – scientists have succeeded in bridging the critical gap between the biological and electronic worlds. A study published in the journal Nature Electronics describes in detail “hybrid biocomputer” combining lab-grown human brain tissue with conventional circuitry and artificial intelligence. A system called Brainoware learned to identify voices with 78% accuracy. This could one day lead to the creation of silicon microchips linked to neurons.
Brainoware combines brain organelles (clusters of human stem cells transformed into “mini brains”) with common electronic circuits. To make it, scientists placed a single organoid on a plate containing thousands of electrodes. The circuits that communicate with the brain organoid convert the information they want to input into a pattern of electrical impulses.
Brain tissue then teaches and communicates with technology. A sensor in the electronic field detects the response of the mini-cerebellum, which is decoded by a trained machine learning algorithm. In other words, with the help of artificial intelligence, neurons and electronics combine into a single (yet extremely basic) problem-solving biomacro.
The researchers taught the computer-brain system recognize human voices. Brainoware trained on 240 recordings of eight people speaking and translated the sound into an electrical form, which they then presented to the organoid. The organic part responded differently to each voice, creating a pattern of neural activity that the AI learned to understand. That’s how Brainoware learned identify voices with 78% accuracy.
The team sees this work as a proof of concept rather than something with any imminent practical use. Although previous studies have shown that two-dimensional cultures of neuronal cells can handle similar tasks, this is first trial run using trained human brain cells. It could be about foundation for biological computerswhere the speed and efficiency of the human brain will be used to create super-powerful artificial intelligence.