A team of researchers from the US and the UK has assembled an artificial neuron that behaves almost like a living brain cell, copying its impulses and adapting to new signals. This could become the basis for robots that better sense and understand the world around them.
A team of researchers from the US and the UK has assembled an artificial neuron that behaves almost like a living brain cell, copying its impulses and adapting to new signals. This could become the basis for robots that better sense and understand the world around them.
According to Interesting Engineering, the new element, called a transneuron, is capable of operating in modes similar to neurons from the brain regions responsible for vision, movement planning, and movement itself. Unlike classic artificial neurons, which are hard-wired for a single task, this device can be «switched» between roles by simply fine-tuning its electrical parameters.
To test the idea, the researchers fed the transneuron electrical signals recorded from the brains of macaques and compared its responses to the activity of living cells. The device reproduced characteristic pulsations from three different zones with 100% accuracy: from uniform firing of pulses to chaotic bursts. When the input signal changed, the frequency of «firings» changed, as in biological neurons, and when two signals were applied simultaneously, the response depended on their time offset.
The key to this flexibility is the memristor. This is a nanoelement whose resistance remembers previous states. Inside the transneuron, silver atoms form and destroy tiny «bridges,» generating electrical pulses. Changes in temperature, voltage, or resistance switch the system to different operating modes, so the same chip can perform the functions of different neurons without rewriting the program code.
The authors emphasize that their device does not simply play back recorded activity, but processes information. This brings neuromorphic systems closer to truly «brain-like» computing, where the hardware, like in the human nervous system, simultaneously stores and processes data.
In the future, the team wants to assemble such elements into a «cortex on a chip,» a network that could form the basis of an artificial nervous system for robots, energy-efficient computing devices, or interfaces to the human nervous system. Unlike purely software-based brain models, such hardware solutions could combine high speed, low power consumption, and adaptability, which is critical for next-generation autonomous robotics.
Previously, dev.ua wrote about how Australian startup Cortical Labs introduced the commercial CL1 biocomputer, which combines living human neurons with silicon chips. This development is capable of processing information in real time and potentially opens a new era in brain research, drug development, and AI.
