Engineers at Northwestern University in the United States have developed artificial neurons that successfully interact with living brain cells. The results of the study were published in the prestigious journal Nature Nanotechnology.
The Invention
These new devices consist of artificial neurons manufactured using aerosol jet printing. This technology involves depositing "electronic ink"—specialized materials for printing electrical circuits—precisely onto designated areas of a flexible polymer base. This process results in devices that are soft and possess properties similar to biological tissue.
The key distinction of this development is its ability to generate complex electrical signals resembling those used by actual neurons. Unlike most artificial counterparts that produce simple pulses, these new neurons replicate various types of activity, including single spikes, bursts, and intermittent patterns.
How It Works
The scientists turned a characteristic of the polymer into an advantage. Normally, this material is removed from neuromorphic systems because it hinders the flow of current. In this instance, the polymer is partially decomposed, and as current passes through, it continues to break down unevenly. This creates a narrow conducting channel that generates a sharp electrical response, mimicking the behavior of a real neuron.
Proof of Concept
To verify the artificial neurons' compatibility with living tissue, researchers tested them on slices of a mouse cerebellum. Electrical signals from the artificial neurons triggered a response in real neurons, matching not only the timing but also the shape of the pulses. This indicates that the devices can genuinely stimulate activity within neural circuits.
Technological Advantages
These printed neurons are highly energy-efficient. Due to the diversity of their signals, a single neuron can encode more information than the standard artificial neurons used in today’s computing systems. This reduces the number of required components and energy consumption compared to the latest AI models that demand massive processing power.
The printing technology also minimizes waste, as materials are only applied where necessary. The resulting devices are relatively inexpensive and simple to manufacture.
Future Applications
The study’s authors believe these printed neurons could serve as the foundation for:
- novel neural interfaces;
- neuroprosthetics for restoring hearing, vision, or movement;
- computing systems that operate on principles similar to the human brain.
Technologies capable of interacting directly with neurons could accelerate the convergence of living tissue and electronic systems. Devices will cease to be "external" to the body and will instead function as extensions of the nervous system. This will transform the approach to treating neurological diseases and restoring bodily functions.
These new neurons can encode more information within a single element, potentially reducing the total number of components in the system. This paves the way for more compact and affordable devices. If the technology scales, complex computing could become more accessible to small businesses and the medical field. Consequently, AI could expand more rapidly beyond the reach of major tech corporations.
Context: Why This Matters Now
Artificial intelligence requires an ever-increasing amount of energy: the growth of models and data is straining data centers, their cooling systems, and power grids. This has evolved into both a technological and an environmental challenge.
In classical electronics, energy efficiency is improved by increasing transistor counts and optimizing chip architecture, but this approach is gradually hitting physical and economic limits. Scientists are seeking alternatives by looking to the biological brain—one of the most energy-efficient "computing devices," capable of processing complex information with extremely low power consumption. The attempt to replicate brain function in electronics is known as neuromorphic computing.
Such approaches are already moving beyond the laboratory. In February 2026, a center opened in Texas that utilizes systems mimicking neuronal activity for computing.
This development marks a significant milestone in the evolution of neural interfaces and next-generation energy-efficient computing, bridging advancements in bioengineering, electronics, and neuroscience.
