Biological neural networks playing the Hunting Game
Technical demo showing how we interface biological neural networks with a game environment. We train them to score, by learning to browse the environment and target moving prey.
Harnessing
the
Computational Power
of
Living Neurons
Highlights
Watch our technical demos:
Training biological networks to process visual information
We demonstrate how biological neural networks can form short-term visual memories. They reconstruct digital images from memory with striking detail. We further showcase their ability in image classification and few-shot learning.
Stay tuned for more!
Company
About
At Neurorium, we train Biological Neural Networks (BNNs) to function as highly energy-efficient
processing units. We envision a future in which BNNs complement, and in some cases replace,
silicon-based computing systems.
Neurorium is built on the idea of harnessing biological neural networks for complex computational
tasks and is rapidly evolving into a sustainable, real-world solution. The immense potential of
biocomputers as ultra-low-power processors is only beginning to be realized, and accelerating this
transition is central to our mission.
As pioneers in the emerging field of biocomputing, we aim to establish ourselves as leading
providers of commercial biocomputing solutions.
Story
Neurorium is the result of a bold leap from an idea in a research lab to a startup for developing
the technology.
In 2023, a joint research project between the Swiss company FinalSpark and the
AI4Science Group at the Free University of Berlin, conducted under the
supervision of our co-founder Mohsen Sadeghi, demonstrated that training Biological Neural Networks
(BNNs) is a feasible task. Founding Neurorium in 2025 was the natural next step
toward developing the required technology for it and unlocking this potential.
For us, this marks only the beginning. Harnessing the computational power of living neurons at
Neurorium lays the foundation for a greener, more sustainable path to intelligence. We have embarked
on a journey to push the boundaries of what is possible through what we call Post-Silicon
Intelligence.
Technology
Our mission is to develop novel methods and approaches to harness the computational power of
biological neural networks (BNNs). Brain organoids, i.e. three-dimensional clusters of living
neurons,
serve as the physical embodiment of these networks.
Communication with brain organoids is achieved by sending and receiving electrical signals via
Multi-Electrode Arrays (MEAs). Signals transmitted through an MEA consist of extremely sparse
spikes
generated by synaptic activity. This is where our core technology comes into play.
We develop methods that translate digital information, such as numerical or visual data, into
spike patterns that are intelligible to neurons. In addition, we design automated procedures for
training
BNNs to respond to this information. These responses are shaped in a closed-loop
manner to achieve well-defined, quantifiable computational objectives. Once training is complete,
BNNs can be deployed as functional substitutes for conventional computational systems.
Through our API, which encapsulates this entire toolchain, users can explore
Organoid Intelligence (OI)
by invoking capabilities analogous to those found in
traditional AI and machine-learning frameworks. Our solutions for training Biological Neural
Networks thus pave the way toward a new era of Post-Silicon Intelligence.
Team
Elly Keramati
Co-founder and Executive Director
elly@neurorium.ai
Elly has a background in computer science, cognitive science and philosophy. She also has experience with startup ecosystem as a contributor to an incubator and leading fresh academic postgraduates toward the industry.
Mohsen Sadeghi
Co-founder and Research Director
mohsen@neurorium.ai
Mohsen is a computational biophysicist with a specialization in the modeling and simulation of biological processes, as well as the application of AI/ML techniques to life sciences. In a recent endeavor, he has been engaged in research into training methods for biological neural networks, out of which, the startup Neurorium is born.
Börries von Seggern
Researcher
Börries is a computational sciences researcher, where he is most interested in modeling processes of biological and chemical interest. At Neurorium, Börries focuses on developing computational techniques to understand and manipulate the behavior of biological neural networks.
Join us
If you are interested in joining us on this journey, whether through collaboration or as an
investor, feel free to reach out at
contact@neurorium.ai.
If you would like to request a demo for a pilot project, or are curious about how our technologies
can be integrated into your existing tech stack, we would be more than happy to start the
conversation. Assisting you in venturing into this new technological frontier is our top priority.
Contact
Follow our latest developments:
contact@neurorium.ai
Altensteinstraße 40, 14195 Berlin
Copyright © 2025-2026, Neurorium. All Rights Reserved.