Scientists Successfully Print Artificial Neurons That Can Communicate with the Brain

This research opens up important avenues for electronic devices capable of communicating with the nervous system. The technology holds promise for the development of brain-machine interfaces and neuroprosthetic devices, including implants aimed at restoring hearing, vision, or motor skills. Simultaneously, the research results point towards a future of more efficient computing. By replicating the way neurons send signals—a core feature of the brain as the most energy-efficient computing system known—next-generation hardware could consume far less energy than current systems when processing complex tasks.

The study was published on April 15th in the journal *Nature Nanotechnology*. Mark C. Hersam, Walter P. Murphy Professor of Materials Science and Engineering at Northwestern, and the study’s lead author, stated: "The world we live in today is dominated by artificial intelligence. To make AI smarter, we need to train it with more and more data. This data-intensive training leads to enormous energy problems. Therefore, we must develop more efficient hardware to process big data and AI. Since the brain is five orders of magnitude more energy-efficient than digital computers, it is reasonable to look to the brain for inspiration for the next generation of computing."

As computing demands grow, traditional systems cope with these challenges by adding more of the same components. Modern chips contain billions of transistors arranged on rigid silicon wafers, each performing the same function. Once manufactured, these systems are immutable. The brain, however, operates very differently. It is composed of diverse types of neurons, each with a specialized role, organized in a soft, three-dimensional network. These networks constantly adapt, forming new connections and reshaping existing ones as learning occurs. Hersam explains: "Silicon achieves complexity by having billions of identical devices. Everything is the same, rigid, and fixed once manufactured. The brain is the opposite. It’s heterogeneous, dynamic, and three-dimensional. To move in that direction, we need new materials and new ways to build electronics."

Although artificial neurons have been created before, most produce signals that are too simplistic. To generate more complex behaviors, engineers typically rely on large networks, which increases energy consumption. To better match the behavior of real neurons, the researchers designed their devices using soft, printable materials. They fabricated a specialized electronic ink using flakes of molybdenum disulfide, which acts as a semiconductor, and graphene as a conductor. These inks were deposited onto a flexible polymer surface using a technique called aerosol jet printing.

Previously, the polymer component in these inks was considered a drawback, as it interfered with current flow, and was therefore typically removed after printing. In this case, the research team instead leveraged its advantages. The researchers stated: "We didn’t completely remove the polymer, but partially decomposed it. Then, when we passed current through the device, we further drove the decomposition of the polymer. This decomposition occurs in a spatially non-uniform way, leading to the formation of conductive filaments, confining all the current to a narrow region in space." This narrow conductive path produced a sudden electrical response, similar to a neuron firing. Consequently, the artificial neurons can generate a variety of signals, including single spikes, steady discharges, and burst patterns, closely mimicking real neural activity. Because each device can handle more complex signal transmission, fewer components are needed overall, which could significantly improve the efficiency of future computing systems.

To determine whether these artificial neurons could interact with real biological systems, the research team collaborated with Indira Raman, a professor of neurobiology at the Weinberg College of Arts and Sciences. Her team applied the artificial signals to slices of mouse cerebellum. The results showed that these electrical spikes matched key characteristics of natural neuronal activity, including timing and duration. The signals reliably activated real neurons and triggered neural circuits in a manner similar to natural brain signals. Hersam said: "Other labs have tried to make artificial neurons from organic materials, but they fire too slowly. Or they use metal oxides, which are too fast. We’re in a time window that hasn’t been demonstrated before in artificial neurons. You can see the live neurons responding to our artificial neurons. So, we’ve shown signals that are not only the right timescale, but also the right spike shape, that can directly interact with live neurons."

This new approach also offers environmental and practical advantages. The manufacturing process is simple and cost-effective, and the additive printing method efficiently uses materials by placing them only where needed, reducing waste. As AI systems continue to expand, improving energy efficiency is particularly crucial. Large data centers already consume vast amounts of electricity and require significant water for cooling. Hersam points out: "To meet the energy demands of AI, tech companies are building gigawatt-scale data centers powered by dedicated nuclear power plants. It’s clear that this enormous power draw will limit further expansion of computing, as it’s hard to imagine the next generation of data centers needing 100 nuclear power plants. Another issue is that dissipating terawatts of power generates a lot of heat. Because data centers use water for cooling, AI is putting a severe strain on water supplies. Either way, we need to develop more energy-efficient hardware for AI."

The research was supported by the National Science Foundation.