In 2020, scientists at Tufts University created tiny novel living forms called xenobots from frog cells, capable of traversing a watery environment, healing their own injuries, and even gathering other cells to build xenobot siblings.
Now, researchers at Tufts and the Wyss Institute have taken the quest to reimagine life forms a step further, adding nerve cells and observing how they self-organize and alter xenobot behavior. The resulting neurobots take on new shapes and show unique behaviors. The researchers reported the findings recently in Advanced Science .
The work, led by Michael Levin, Vannevar Bush Professor of Biology and Haleh Fotowat, senior scientist at the Wyss Institute, is part of a larger effort to understand the way cell collectives adapt and organize to build complex structures under new conditions – which could pave the way for advances in synthetic biology and regenerative medicine. The study of neurobots will help the researchers understand what those rules of self-organization are for the nervous system, and ultimately how to work with those rules to guide them to new structures in the lab or restore existing tissues in the body.
The research team started with cells from early embryos of the African clawed frog, Xenopus laevis . When precursor skin cells from these embryos are removed and allowed to develop in a dish, they spontaneously form small, spherical structures covered in tiny hair-like projections called cilia.
By coordinating the beating of these cilia, the Xenobots, a kind of biobot, swim through water. They are fully biological, formed without any scaffolding materials or genetic manipulation, capable of self-healing, and can survive for about 9 to 10 days on nutrients stored in the original embryonic cells. The team had extensively characterized the form and behavior of these bots, and wanted to know: how would these change if there were neurons present, and, what does a nervous system look like in a novel being whose neural structure and function had never had a chance to be shaped by evolution within ancient environments?
To create the neurobots, the researchers implanted clusters of neural precursor cells from frogs — cells destined to become neurons—into the center of the developing structure during the brief window when the spherical biobots were forming. The implanted cells matured into neurons and extended branching projections— axons and dendrites—throughout the interior and even toward the outer surface of the bots.
“We wanted to find out what would happen if we provided these biobots with the raw materials needed to build a nervous system,” said Levin, director of the Allen Discovery Center at Tufts . Making neurobots was a way to test how neurons organize on their own and perhaps modify the way the neurobot moves. Unlike studying neurons in lab dishes and in small samples of tissue, “this approach is different because you now have a system with a biological body that can exhibit behavior,” said Levin.
Understanding How Nervous Systems Form
For Fotowat, creating the neurobots was a way to examine the fundamental rules of how a nervous system forms: “I’ve tried to understand neuronal behavior in existing animals like zebrafish, and how they give rise to behavior, but neurobots are about reverse engineering. Can we build a nervous system from the start? What happens if you put neurons in a completely novel context? What are the basic, innate rules for them to organize and form networks?”
Microscopy revealed that the neurons in the bots developed hallmark features of real nervous systems, including axons and dendrites—long and short tree-like branches. The researchers identified protein markers typically associated with synapses, the contact points where neurons communicate. Calcium imaging—an established method for visualizing neural activity—showed that the neurons inside neurobots were electrically active and functioning in primitive neural networks.
Adding neurons changed the neurobots in visible and measurable ways. Compared to non-neural biobots, neurobots tended to grow larger and more elongated. They also moved differently. While both types could swim, neurobots were less likely to sit still and were more likely to display complex, repeating movement patterns rather than simple circles or straight lines.
To test whether neural signaling was influencing behavior, researchers exposed the bots to a drug pentylenetetrazole, which is known to affect brain activity and induce seizures. The drug altered the movement patterns of neurobots differently than it did those of non-neural biobots, suggesting that the newly formed nervous systems were actively shaping behavior. This provided further evidence that even a simple, self-organized neural network can influence how this novel creature moves.
“If you’re trying to build something new with biology, we first have to learn how cells themselves solve problems,” said Fotowat.
Surprisingly, in addition to seeing genes activated for major brain receptors, they also found activation of genes involved in visual perception, including those associated with light-sensitive cells in eyes. Could they be paving the way for the neurobots to perceive and respond to light cues?
“We don’t know, but my hypothesis is that these neurobots are up-regulating parts of the genome that could be useful for novel functions down the line,” said Levin. “If they lived longer, would they then also develop photoreceptors? It’s a fascinating question that we are actively studying.”
Advanced Science
Experimental study
Cells
Engineered Living Systems With Self-Organizing Neural Networks: From Anatomy to Behavior and Gene Expression
20-Feb-2026
Authors report no conflicts of interest