Brain Organoids Decode Braille Signals in Biohybrid Computing Breakthrough

Scientists show that brain organoids can receive and interpret Braille signals.

Image credit:©iStock.com, alejomiranda

The human brain carries out extraordinarily complex computations on roughly 20 watts of power—a small fraction of the energy required by conventional computing systems. Inspired by this efficiency, engineers have created neuromorphic chips, brain-inspired processors that unite memory and computation in the same local architecture, mirroring biological synapses. Unlike traditional processors that handle data continuously, neuromorphic chips activate only during relevant events, making them significantly more energy efficient.

Scaling neuromorphic systems to rival the complexity of the human cerebral cortex remains difficult, leading researchers to investigate biohybrid computing, which merges living neural tissue with electronic hardware. In research published in Frontiers in Artificial Intelligence, a team at the University of Bristol showed that human brain organoids can identify Braille characters. Their experimental platform converted input from a tactile sensor into signals that the organoids could process, suggesting a path toward using brain organoids as biohybrid computational units for low-power computing.

Tianyi Liu, a graduate student at the University of Bristol and study coauthor, said in a press release, “We can take information from a real tactile sensor, convert it into a form that a living neural network can receive, and extract meaningful information from the organoid response.” She added, “This provides a practical way to link artificial sensors with biological neural systems, a promising avenue toward more energy-efficient computing.”

Testing Whether Brain Organoids Can Recognize Braille

To evaluate the system, Liu and her colleagues mounted a neuromorphic tactile sensor on a robotic arm to scan a board containing the 26 letters of the Braille alphabet. The sensor gathered tactile data from 1,300 recordings, which were translated into stimulation patterns delivered to the brain organoids. These patterns differed in pulse strength, timing, and quantity. Electrodes positioned beneath the organoids recorded the resulting neural activity, and a standard machine-learning algorithm analyzed the responses to determine the corresponding Braille letters.

In initial tests, the team used a single organoid. Each organoid contained mature neurons, astrocytes, and oligodendrocytes with strong network activity. When electrical stimulation matching a specific Braille letter was delivered, the organoid’s neural activity could be decoded to identify the letter with approximately 61 percent accuracy.

To see whether multiple organoids would improve performance, the researchers expanded the setup to three organoids.

“Different organoids may have slightly different network connections, developmental states, and background activity, so identical inputs can produce similar yet distinct responses,” Liu explained. “Combining those responses may provide complementary information unavailable from a single organoid. At the same time, if every organoid behaves differently, creating a predictable and standardized system becomes more difficult.”

As the saying goes, two heads are better than one. The multi-organoid approach raised letter-recognition accuracy to 83 percent. The team also ran a stress test on single and three-organoid groups by introducing artificial noise. Once again, the organoid cluster decoded and classified Braille letters more reliably under noisy conditions. This indicates that integrating information across multiple neural networks may improve the dependability of biological computing systems.

A Step Toward Biohybrid Computing

The researchers stressed that this proof-of-concept study demonstrated that brain organoids can receive and respond to electrical signals representing Braille letters. While the findings point to the potential of brain organoids as low-power biohybrid computing systems, more work is needed to understand how performance varies with different organoid counts and whether the method applies beyond Braille recognition. The team aims to extend the system to handle more complex tactile information.

“We are interested in textures, edges, and different types of touch, and eventually in closed-loop tasks where the organoid receives feedback based on its actions,” Liu said. “That would allow us to ask whether the organoid can adapt through experience.”

Source link