Brain Organoids Learn to Decode Braille Letters

University of Bristol researchers taught human brain organoids to classify Braille letters, with three-organoid teams hitting 83% accuracy.

@sharedot

Brain Organoids Learn to Decode Braille Letters

By @sharedot · · 6 pages

  • AI
  • Organoid Intelligence
  • Biohybrid Computing
  • Research

University of Bristol researchers taught human brain organoids to classify Braille letters, with three-organoid teams hitting 83% accuracy.

Lab-grown brain tissue reads Braille

Researchers at the University of Bristol report, in a study published October 16 in Frontiers in Artificial Intelligence, that lab-grown human brain organoids can classify the 26 letters of the Braille alphabet.

The team used a tactile sensor to scan Braille dot patterns, converted that data into electrical stimulation patterns, and fed them to organoids smaller than a millimeter across, each containing about 10,000 neurons plus supporting cells, held alive on multielectrode arrays.

A support vector machine then decoded the organoids' neural activity to identify which letter had been presented.

Why three organoids beat one

According to 404 Media, a single organoid classified the letters about 61% of the time, while an ensemble of three organoids reached 83%.

Mirage News adds that the three-organoid team was also better at distinguishing letter signals from artificial noise than a decoder working directly on the obscured signals without the organoids' involvement.

The researchers' hypothesis is that different organoids respond to the same stimulation in similar but not identical ways, like three people looking at the same object and each noticing different details, so combining their responses yields complementary information.

No learning yet — just decoding

The researchers stress this was an open-loop experiment: the organoids received initial inputs but no feedback about their performance.

Per Mirage News, first author Tianyi Liu said the team showed tactile information could be encoded, processed through the organoids and decoded, but did not demonstrate learning or long-term changes in the organoids.

404 Media reports that future closed-loop systems, where a robot's interactions are fed back to the organoids, could begin to reveal whether these biological neural networks adapt to new tasks through ongoing experience.

The path to energy-efficient biohybrid computers

The work falls under organoid intelligence, an emerging field aiming to use living neural tissue as biological "wetware" for computing, with prior studies showing organoids recognizing Japanese vowels and playing Pong.

Liu told 404 Media that the brain performs complex processing with little energy, asking why not use real biological neural networks directly.

Limitations remain: Mirage News reports that organoids cannot be standardized, and Liu acknowledged that compared with established chips, organoid systems lag in reliability, reproducibility and scalability. Next steps include textures, edges and other touch tasks beyond Braille.

Sources

  1. 404media.co › Computers Made of Human Brain Tissue Identify Braille Letters In Breakthrough
  2. miragenews.com › Mini Brain Organoids Read Braille in Biohybrid Computing Breakthrough

More on AI