Living Circuits: How Biology Is Becoming Computation

A groundbreaking study by researchers Panagiotis Mougkogiannis and Andrew Adamatzky has achieved what neuromorphic engineers have long pursued. Creating brain-inspired computing systems that don’t just replicate neural behavior, but actually emerge from biological materials themselves. By combining two naturally occurring biological components—proteinoids and myelin—the team has created hybrid microsystems that represent the next evolutionary leap in neuromorphic technology, performing complex calculations while consuming minimal energy through genuine bio-neural processes.

The research, published in September 2025, comes at a pivotal moment for the neuromorphic computing field. Leading researchers say the time is ripe to start building the first large-scale neuromorphic devices that can solve practical problems, and this bio-hybrid approach offers a fundamentally new pathway toward that goal. different from traditional silicon-based neuromorphic chips that simulate neural behavior, these biological systems generate authentic neural-like electrical signals, process information, and execute Boolean logic operations—the fundamental building blocks of all digital computation.

The Next Frontier of Neuromorphic Architecture

Standard neuromorphic computing has spent decades trying to imitate the brain’s extraordinary efficiency in silicon. By 2025, the neuromorphic computing market is projected to grow at 108% CAGR, with applications spanning robotics, healthcare, and edge devices, driven largely by the energy crisis facing conventional AI systems. Data centers consume 2% of global electricity—neuromorphic chips cut this by 1000x.

However, this new bio-hybrid approach represents a paradigm shift within neuromorphic technology itself. Rather than engineering silicon to behave like neurons, these researchers start with the biological building blocks that neurons themselves use. At the heart of this innovation lie two exceptional biological materials that form nature’s own neuromorphic architecture.

Proteinoids form when amino acids are heated to their melting point, creating polymer chains that self-assemble into microscopic spheres. These tiny structures, measuring between 1152 to 1889 nanometers in diameter, possess what neuromorphic engineers have long sought. The ability to generate their own electrical activity without external power sources—true event-driven, asynchronous processing at the molecular level.

Myelin, the second component, serves as nature’s most intricate electrical insulation system in biological neuromorphic networks—the nervous system. This lipid-rich material wraps around nerve fibers, enabling the rapid signal transmission that makes biological neural networks so efficient. In the hybrid system, myelin contributes its exceptional electrical properties while maintaining the biocompatibility essential for next-generation neuromorphic applications that must integrate with living tissues.

Biological Neuromorphic Processing in Action

The researchers observed these biological neuromorphic systems over an impressive 180,000 seconds—more than two full days—recording their electrical behavior with microscopic precision. What they discovered surpasses many artificial neuromorphic implementations. These biological materials spontaneously produce membrane potentials ranging from -90 mV to +70 mV, closely matching the electrical signatures of actual neurons and exhibiting the event-driven, spike-based communication that defines neuromorphic computing.

The hybrid structures exhibited complex neuromorphic phenomena including spontaneous spike generation, phase transitions, and oscillatory patterns—all hallmarks of effective brain-inspired computing systems. Beyond pure proteinoid microspheres, the myelin-enhanced versions showed dramatically improved electrical properties: higher capacitance (159.3 nF) and lower impedance (3.934 kΩ), making them more efficient at conducting electrical signals than many synthetic neuromorphic devices.

Compared to conventional chips, these processors act like living brains: firing only when needed, running in parallel, and sipping power rather than guzzling it. This biological implementation takes that principle to its logical extreme—these systems literally are biological neural elements.

Computing with Living Neuromorphic Matter

Perhaps most important for the neuromorphic field is the system’s authentic computational capability. The research team developed an innovative approach that bridges analog biological signals and digital neuromorphic processing. They established channel-specific voltage thresholds based on statistical analysis of the complete 50-hour recording period, with thresholds ranging from -42.1 mV to +15.2 mV across different channels.

By converting voltages above these thresholds to binary “1” states and those below to “0” states, the researchers successfully performed complete Boolean logic operations using the bioelectrical activity of these hybrid materials. This represents a breakthrough in neuromorphic implementation. The systems can execute the same fundamental logical operations that power neuromorphic chips—AND, OR, NOT, and XOR operations—but using authentic biological neural components instead of silicon approximations.

The timing characteristics of these bio-neuromorphic conversions varied dramatically across channels, providing multiple computational modes within a single system. Some showed rapid switching with transition frequencies of 0.0157 Hz, ideal for real-time neuromorphic processing, while others maintained stable states for over 2.5 hours (9000 seconds average duration), perfect for neuromorphic memory applications. This temporal diversity offers the kind of heterogeneous processing that neuromorphic systems need for complex cognitive tasks.

The energy consumption operates at nanowatt to microwatt levels (10-100 nW per channel), approaching the extraordinary energy efficiency of biological neurons (10 pJ per spike). This represents a substantial improvement over even the most advanced silicon neuromorphic devices, which typically consume 1-10 μW per synapse. Research using Loihi 2 processors has delivered orders of magnitude gains in the energy efficiency compared to traditional computing, but these biological systems push efficiency even further.

Beyond Silicon Neuromorphic Chips: A Living Computing Paradigm

This research addresses critical obstacles facing current neuromorphic technology. While gradient-based training of deep spiking neural networks is now an off-the-shelf technique for building general-purpose Neuromorphic applications, the hardware substrate has remained fundamentally artificial. Most neuromorphic devices are made from silicon, creating an inherent disconnect between the biological inspiration and synthetic implementation.

This bio-hybrid approach offers three distinct advantages for neuromorphic computing. First, it provides true biocompatibility—essential for brain-computer interfaces and implantable neuromorphic sensors that must integrate seamlessly with neural tissue. Current silicon neuromorphic chips like Intel’s Loihi, IBM’s NorthPole, and other processors face significant challenges when interfacing with biological systems.

Second, the ultra-low power consumption could solve the energy challenges that limit neuromorphic deployment in edge computing and Internet of Things applications. Real-Time Demands. Autonomous vehicles and other edge applications require neuromorphic processors that can operate continuously on minimal power—these biological systems deliver exactly that capability.

Third, the emergence of computational behavior through self-assembly reveals fundamental principles about how neuromorphic architectures can grow and adapt rather than being manufactured. This overcomes one of the key bottlenecks in scaling neuromorphic systems. How to create the massive, interconnected networks needed for complex cognitive processing.

Implications for Neuromorphic Technology Development

The implications extend throughout the neuromorphic computing ecosystem. For materials scientists developing neuromorphic substrates, this work illuminates how structure-function relationships in biological systems can inform next-generation neuromorphic materials. Without limiting themselves to CMOS-compatible processes, neuromorphic engineers can explore bio-hybrid approaches that combine the best of biological and synthetic neuromorphic elements.

For neuromorphic algorithm developers, these systems provide a novel computing substrate where logic operations emerge from material properties as opposed to programmed circuits. This could enable new approaches to neuromorphic learning and adaptation that more closely mirror biological neural plasticity.

Biomedical engineers developing neuromorphic medical devices gain access to truly biointegrated neuromorphic processors that can interface seamlessly with neural tissue. This is crucial for next-generation neuroprosthetics, brain-computer interfaces, and implantable neuromorphic sensors that must operate within biological environments.

Neuroscientists studying biological neural networks gain a simplified but authentic neuromorphic model for understanding how electrical signals and computational behaviors emerge from membrane dynamics and spatial organization—insights that can inform both biological understanding and neuromorphic design.

Methodology: Bio-Neuromorphic System Development

The methodology represents a new approach to neuromorphic system fabrication, combining thermal condensation to create proteinoid-myelin hybrid microspheres with advanced neuromorphic characterization techniques. Electrical characterization uses microelectrodes to record extracellular potentials, while electrochemical impedance spectroscopy reveals charge transport mechanisms—standard techniques in neuromorphic device evaluation.

The threshold-based signal processing that converts analog potentials into digital signals for logic operations follows established neuromorphic processing principles, but applies them to biological rather than silicon substrates. This comprehensive approach successfully links structural features to electrical behaviors across multiple spatial and temporal scales, providing a roadmap for bio-neuromorphic system development.

The Bio-Neuromorphic Revolution

As neuromorphic computing stands poised for widespread commercial adoption, this research opens an entirely new branch of the field. Neuromorphic chips have the potential to outpace traditional computers in energy and space efficiency as well as performance, but bio-neuromorphic systems could surpass even synthetic neuromorphic devices by leveraging billions of years of biological neural optimization.

The 50-hour continuous recordings captured behaviors ranging from rapid millisecond spikes to gradual phase transitions occurring over days—demonstrating temporal complexity that rivals the most sophisticated neuromorphic architectures. The fact that different regions naturally specialize for different computational functions suggests that self-organization principles could guide the development of increasingly sophisticated bio-neuromorphic systems.

This work fundamentally expands the definition of neuromorphic computing. While traditional neuromorphic technology seeks to emulate neural behavior in silicon, bio-neuromorphic systems use actual biological neural components to perform computation. This represents not just an incremental improvement, but a new category of neuromorphic technology that bridges artificial and biological neural networks.

The Living Neuromorphic Future

Rather than forcing synthetic materials to work within biological contexts, this bio-neuromorphic approach begins with biological neural building blocks and enhances them with computational capabilities. This fundamental shift could address key challenges facing neuromorphic technology: biocompatibility for implantable applications, ultra-low power consumption for edge devices, and scalability for large neuromorphic networks.

The research reveals that simple amino acids, when properly organized, can spontaneously generate the complex electrical behaviors necessary for neuromorphic information processing. These bio-neuromorphic systems truly bridge synthetic computing and biological intelligence, offering insights into how neural-like processing emerges from basic molecular interactions while providing practical advantages over conventional neuromorphic approaches.

As the neuromorphic field advances toward practical applications in robotics, healthcare, and edge AI, bio-neuromorphic systems point at a future where neuromorphic processors grow rather than being manufactured, where neuromorphic devices self-repair through biological processes, and where the distinction between biological neural networks and artificial neuromorphic systems becomes increasingly meaningless.

This bio-neuromorphic revolution challenges traditional boundaries not just between living and artificial systems, but within neuromorphic technology itself. As we witness the emergence of truly living neuromorphic computation, we may be seeing the birth of neuromorphic systems that seamlessly integrate the best of both biological neural networks and artificial computing architectures—representing the ultimate realization of brain-inspired computing.


Citation: Mougkogiannis, P.; Adamatzky, A. “Myelin–Proteinoids Interactions in Neural Signaling.” American Chemical Society, September 5, 2025. Licensed under CC-BY 4.0.

Related Articles

Responses

  1. Ciao! Questo articolo presenta delle informazioni molto interessanti sul ruolo dei circuiti neurali nella risposta allo stress. Vorrei chiedere se ci sono ulteriori studi o approfondimenti riguardo all’uso del propranololo e come esso influisce sui circuiti neurali in situazioni di stress elevato. Ho trovato un articolo in merito, che approfondisce la genetica della risposta allo stress e il ruolo delle sostanze farmacologiche: https://toddanthonylab.org/decoding-the-stress-response-neural-circuits-genetic-blueprints. Grazie!

    1. Hi there! Thanks so much for your thoughtful comment. That’s a really interesting question, and you’ve made a great connection between the research we talked about and pharmacology.

      The article “Living Circuits” looks at some of the basic biological building blocks behind neuromorphic systems. While the researchers didn’t specifically study propranolol, what they explored helps us start to understand how things like drugs might influence these bio-neural networks. The link you shared is a nice example showing how genetics and pharmacology, especially around stress responses, are important areas to explore.

      The big picture goal here is to build systems that can copy or connect with biological neural networks. Figuring out how drugs like propranolol affect those networks will be really important for future uses—like brain-computer interfaces or medical devices.

      Thanks again for jumping in with such a thoughtful comment. It’s exactly this kind of conversation that helps push things forward.

Comments are closed.

PHP Code Snippets Powered By : XYZScripts.com
★

Track This Company

Create a free account to add companies to your watchlist and get notified when patents, funding, or key developments change.