In Vitro Biological Neuronal Networks Offer Blueprint for Low-Power, High-Speed AI Computing
20 May 2026
A research team led by Prof. CAI Xinxia, Assoc. Prof. LUO Jinping and Assoc. Prof. WANG Mixia from the Aerospace Information Research Institute of the Chinese Academy of Sciences (AIRCAS) has demonstrated that in vitro biological neuronal networks can achieve both low-power consumption and high-speed communication through predictable electrical stimulation, providing a promising new route for next-generation brain-inspired computing systems. The study shows that regular stimulation increased neural communication speed by about 1.79 times without adding extra energy cost.
Supported by the Foundation for Innovative Research Groups of the National Natural Science Foundation of China, the research was recently published in ACS Sensors under the title In Vitro Biological Neuronal Networks Achieve Low-Power Consumption and High-Speed Communication through Predictable Stimulation.
With the rise of artificial intelligence (AI), traditional computing architectures are facing major challenges in energy efficiency and parallel processing capacity. By contrast, biological neuronal networks (BNNs) can process complex information with remarkably low power consumption. Understanding how living neurons compute is therefore regarded as an important frontier for developing efficient brain-inspired processors.
To support this research, the team developed a multi-channel in vitro microelectrode array (MEA) featuring a platinum nanoparticle and conductive polymer (PtNPs/PEDOT:PSS) composite coating for large-scale neural recording. The device achieved an impedance of just 15.33 ± 0.63 kΩ at 1 kHz, while also delivering high charge storage capacity and excellent biocompatibility. These properties provide a robust hardware foundation for long-term neural recording and precise stimulation.
The team also improved the way neuronal communication velocity is calculated. Traditional straight-line measurements often underestimate actual signals transmission speed. To address this, the team developed a new algorithm based on actual propagation distance, significantly improving measurement accuracy.
Using hippocampal neural networks, the researchers then applied regular and predictable electrical pulses. They found that this predictable stimulation significantly increased the network's communication speed, enabling faster signal transmission without increasing power demand.
According to the researchers, the findings not only deepen understanding of the computational principles of living neuronal networks, but also provide a foundation for in vitro brain-computer interface devices that enable low-power, high-speed communication. The work could support future advances in biohybrid processors, neuromorphic computing, and energy-efficient AI systems.
From living neurons to intelligent chips: a brain-on-chip platform enabling low-power, high-speed neural computing. (Image by AIRCAS)
Research News
In Vitro Biological Neuronal Networks Offer Blueprint for Low-Power, High-Speed AI Computing
A research team led by Prof. CAI Xinxia, Assoc. Prof. LUO Jinping and Assoc. Prof. WANG Mixia from the Aerospace Information Research Institute of the Chinese Academy of Sciences (AIRCAS) has demonstrated that in vitro biological neuronal networks can achieve both low-power consumption and high-speed communication through predictable electrical stimulation, providing a promising new route for next-generation brain-inspired computing systems. The study shows that regular stimulation increased neural communication speed by about 1.79 times without adding extra energy cost.
Supported by the Foundation for Innovative Research Groups of the National Natural Science Foundation of China, the research was recently published in ACS Sensors under the title In Vitro Biological Neuronal Networks Achieve Low-Power Consumption and High-Speed Communication through Predictable Stimulation.
With the rise of artificial intelligence (AI), traditional computing architectures are facing major challenges in energy efficiency and parallel processing capacity. By contrast, biological neuronal networks (BNNs) can process complex information with remarkably low power consumption. Understanding how living neurons compute is therefore regarded as an important frontier for developing efficient brain-inspired processors.
To support this research, the team developed a multi-channel in vitro microelectrode array (MEA) featuring a platinum nanoparticle and conductive polymer (PtNPs/PEDOT:PSS) composite coating for large-scale neural recording. The device achieved an impedance of just 15.33 ± 0.63 kΩ at 1 kHz, while also delivering high charge storage capacity and excellent biocompatibility. These properties provide a robust hardware foundation for long-term neural recording and precise stimulation.
The team also improved the way neuronal communication velocity is calculated. Traditional straight-line measurements often underestimate actual signals transmission speed. To address this, the team developed a new algorithm based on actual propagation distance, significantly improving measurement accuracy.
Using hippocampal neural networks, the researchers then applied regular and predictable electrical pulses. They found that this predictable stimulation significantly increased the network's communication speed, enabling faster signal transmission without increasing power demand.
According to the researchers, the findings not only deepen understanding of the computational principles of living neuronal networks, but also provide a foundation for in vitro brain-computer interface devices that enable low-power, high-speed communication. The work could support future advances in biohybrid processors, neuromorphic computing, and energy-efficient AI systems.
From living neurons to intelligent chips: a brain-on-chip platform enabling low-power, high-speed neural computing. (Image by AIRCAS)