10/3/2025 Ben Libman
A team led by bioengineering professor and Dean of The Grainger College of Engineering Rashid Bashir, with lead author Hyegi Min, has created the first biohybrid robots powered by neurons activating muscle tissue mimicking natural neuromuscular function. By integrating genetically modified neurons, muscle strips, and wireless micro-LEDs, the robots crawled forward and even showed signs of learning, continuing to move after stimulation stopped. The advance extends robot lifespan and opens doors to applications in disease modeling, drug testing, and future autonomous biohybrid machines.
Written by Ben Libman
A team of Illinois Grainger engineers have taken a major step forward in the field of biohybrid robotics.
Biohybrid robots combine the mechanical elements of familiar robotics with real animal cells. This means using muscle tissue to facilitate movement instead of a more traditional robotic mechanism, like hydraulics.
Earlier models of these robots relied on muscle tissue alone. Researchers would use electricity to stimulate muscle tissue directly or in case of muscles that were transfected with special proteins, light could also be used to actuate the muscles. Connecting neurons to muscles to produce coordinated motion of the robots has been a grand challenge. No one had been able to get the muscles to move by modulating neurons, the way they do in animals. “There was a lot of exciting work on neural training and programming,” said Hyegi Min, postdoctoral researcher in bioengineering and the Holonyak Micro and Nanotechnology Lab, “but no one had shown a clear link between neuromodulation and mechanical performance.”
That is, until now. PI, bioengineering professor and Dean of The Grainger College of Engineering Rashid Bashir, lead author Hyegi Min, and an interdisciplinary team at Grainger and collaborators at Northwestern University have made a significant advancement in the field of biohybrid robotics, detailed in a paper recently published in Science Robotics. The team used a 3D printer to make a simple two-legged skeleton and filled it with living mouse muscle cells, which grew into strips capable of contracting. These mouse muscle cells were attached to specialized neurons, which the team had genetically modified to respond to blue light. The muscle tissue and neurons combined to form neuromuscular junctions (NMJs), which allow the communications between nerves and muscles in animals. Finally, wireless microscale LED lights, prepared by professor John A. Rogers’ group at Northwestern University, were placed on top of the robots to remotely deliver pulses of blue light. The dynamic behavior of the robots were analyzed using finite element analysis (FEA) by professor Yonggang Huang’s group at Northwestern university.
The results constituted a significant move forward in the field of robotics. By pulsing the blue light, the team was able to compel the neurons to fire. These neurons in turn activated the muscle, moving the crawler forward. The team pulsed blue light at varying speeds. They found the crawling velocity was greater with a slower pulse-light, while faster pulses caused the robot to have weaker movements.
Interestingly, the crawlers continued to move even after the light had turned off, demonstrating that the neurons were learning and adapting as they do in animals. “This was one of the most exciting findings,” said Min. “That shows us that the system can be tuned and operate autonomously once neurons are trained or modulated to a specific mode. It’s a promising step toward creating skeletal-muscle-driven biohybrid robots that can exhibit automated behaviors.”
Creating a system in which neurons activate muscles is mutually beneficial, as Min explains: “When neurons innervate muscles, the muscle releases chemicals that can enhance neural firing activity. Conversely, neurons also provide the chemicals and stimulation necessary for muscle development.” This may be the reason behind the robot’s enhanced lifespans- while earlier, muscle-only robots had a lifespan of only 10 days, these new models lasted over two weeks and showed muscle activity for a full week after that. Though a mechanism for this longevity has not been verified, the symbiotic relationship between muscle tissue and neurons is a likely answer. “This is significant because extending lifespan means these robots can be studied and applied over longer periods,” said Min. “It opens the door to more complex experiments, as well as potential applications in drug screening or disease modeling, where stability over time is critical.”
The potential for these robots is limitless. Though small today, if the technology continues to advance, we could one day have fully formed machines that run themselves with no need for external commands. Their flexible scaffolding is more adaptable than most rigid machines, making them better on uneven terrains. These robots also could be useful models for neuromuscular diseases such as ALS, or test subjects for novel pharmaceutical therapies.
Min is continuing to push these robots forward. “We’re working on integrating more complex architecture. One project involves combining two muscle tissues with a single neuronal unit in the same scaffold. Another we’re pursuing is building more complicated neural circuits. Both directions are steps toward biohybrid robots that can not only move forward, but also turn, adapt, and eventually perform tasks with much greater functionality.”
Hyegi Min emphasizes that this work is only beginning. “We’ve shown that it’s possible to control and tune neuromuscular actuation through neuromodulation, but the broader vision is much bigger. Our long-term goal is to create biohybrid robots that can embody higher-level neural functions such as learning, memory, and programmed responses, while still controlling mechanical outputs like in animals and the human body.”
Whatever the future holds, Min’s crawlers are exemplary of the groundbreaking research Grainger can accomplish via interdisciplinary collaboration. “This research is spanning the interface of neuroscience, tissue engineering and robotics,” said Min. “It demonstrates not only a proof-of-concept for autonomous, neuron-controlled robots but also provides a foundation for systems that could combine biological intelligence with robotic function.”
Rashid Bashir is Dean of The Grainger College of Engineering, professor of bioengineering and Grainger Distinguished Chair in Engineering at the University of Illinois Urbana-Champaign.