University of California, Irvine
UC Irvine Biorobotics Lab
Biorobotics
Dr. David J. Reinkensmeyer is a Professor in the Departments of Mechanical and Aerospace Engineering, Anatomy and Neurobiology, and Biomedical Engineering at the University of California, Irvine. He earned his PhD and MS in Electrical Engineering from UC Berkeley and a BS in Electrical Engineering from MIT.
Rebuilding Movement: How Robots Are Teaching the Brain to Heal After Injury
A rehabilitation robot can help a patient move, but can it teach the brain to move again? University of California, Irvine biomedical engineer Dr. David Reinkensmeyer is uncovering how robotic systems can harness neuroplasticity, transform rehabilitation, and engineer new pathways for recovery after neurological injury.
By BioBuilt Editorial
A patient recovering from a stroke may still have functioning muscles, intact joints, and the physical structures needed to move. Yet a simple action, such as reaching for a cup or taking a step, can become extraordinarily difficult. The challenge is not always the body itself. It is the brain’s ability to communicate with and control that body again.
For decades, rehabilitation after neurological injury has relied on repeated physical practice to encourage recovery. Patients perform movements again and again, hoping that the brain will gradually reorganize itself. But as our understanding of neuroplasticity has advanced, engineers have begun asking a deeper question: can technology actively shape the way the brain relearns movement?
At the University of California, Irvine, biomedical engineer Dr. David Reinkensmeyer is exploring this question through rehabilitation robotics and human movement engineering. His work focuses on designing machines that do not simply assist patients, but interact with the nervous system to encourage recovery. Rather than replacing the work of the brain, his goal is to understand how robotic systems can create the right conditions for the brain to adapt.
"The goal is not simply to make movement look normal," Dr. Reinkensmeyer explained. "The challenge is designing a system that promotes neuroplasticity."
This idea has guided decades of research in his laboratory, where engineers develop robotic devices, computational models, and rehabilitation technologies designed around one central principle: recovery occurs when the brain is challenged, engaged, and given meaningful feedback.
Rehabilitation as an Engineering Problem
The human brain is one of the most adaptive systems in nature. Every time someone learns a new skill, neural circuits reorganize based on experience. This ability, known as neuroplasticity, allows the brain to strengthen useful connections and adjust its behavior over time.
After neurological injuries such as stroke or spinal cord injury, however, this process becomes more complicated. Some pathways responsible for movement may be damaged, leaving the brain dependent on remaining neural circuits. Rehabilitation attempts to strengthen these surviving pathways through repeated training, but the effectiveness of that training depends on how the brain receives and interprets information.
This creates a unique engineering challenge. A rehabilitation robot can make movement easier, but easier movement does not always lead to better learning. If a machine performs too much of the task, the patient may no longer need to actively engage the neural pathways necessary for recovery.
Dr. Reinkensmeyer’s research revealed this unexpected problem through a phenomenon he describes as "slacking." When a robotic device rigidly assists movement, patients often gradually reduce the amount of force they contribute.
"The patient can essentially fall asleep, like Wallace in The Wrong Trousers," he explained. "The robot helps their movement look normal, but that is not necessarily good for plasticity."
The reason lies in how the brain naturally controls movement. Rather than simply trying to move accurately, the brain balances multiple goals, including minimizing effort. Dr. Reinkensmeyer’s team has shown that the brain effectively solves a type of optimization problem, weighing movement accuracy against the energy required to complete the task.
"Your brain tries to minimize a cost function that has both kinematic error and effort terms," he explained. "If error is small because the robot is assisting, then the brain works on minimizing effort."
This discovery changed the way researchers think about rehabilitation robotics. The purpose of a robotic device is not to eliminate difficulty. It is to create the ideal amount of difficulty that encourages the nervous system to adapt.
Designing Robots That Teach
Preventing neural slacking requires engineers to rethink what assistance means. A successful rehabilitation robot cannot simply take control of movement. Instead, it must provide support while keeping the patient actively involved.
One strategy is adaptive assistance, where the robot provides help only when the patient needs it. If the patient can complete part of a movement independently, the device allows that effort rather than replacing it.
Researchers have also explored intentionally making robotic assistance less predictable. While it may seem beneficial for a robot to perform a movement perfectly every time, flawless assistance can reduce the challenge required for learning. Small variations encourage the brain to continue adjusting.
Another approach is requiring patients to initiate movements before robotic assistance begins. This ensures that the brain remains engaged from the beginning of the task.
"The device has to present a learning challenge to the person," Dr. Reinkensmeyer said. "Not too difficult, not too easy."
This principle mirrors how humans naturally learn. Whether mastering a sport, learning an instrument, or recovering from injury, improvement occurs when a person operates at the boundary between ability and challenge.
The best rehabilitation robot is therefore not the machine that does the most work. It is the machine that understands exactly when to help and when to allow the brain to struggle.
The Hidden Sense That Teaches the Brain to Move
Although movement is often associated with muscles and strength, successful movement depends on another system that is less visible: proprioception.
Proprioception is the body’s internal sense of position and motion. It allows someone to know where their hand is without looking at it, or adjust their balance while walking without consciously thinking about each movement.
For Dr. Reinkensmeyer, proprioception may represent one of the most important signals guiding motor learning.
"Our working hypothesis is that proprioception is the key teaching signal that the brain leverages to shape motor commands," he explained.
Every movement produces a pattern of sensory information. The brain compares what it intended to do with what actually happened, then adjusts future movements based on that feedback.
"When you move, the movement causes a pattern of proprioceptive input, and your brain looks at that pattern to judge how good that movement was and try the next thing," he said.
This explains why robotic rehabilitation can be so powerful. A robot does not only assist movement. It can amplify the sensory information returning to the brain by allowing patients to move farther, faster, and more accurately.
However, this benefit depends on maintaining the brain’s active participation. Assistance must enhance learning, not replace it.
Bringing Rehabilitation Beyond the Laboratory
While robotic rehabilitation technologies have shown promise in research settings, one of the greatest challenges is translating these systems into tools that patients will consistently use in their daily lives.
Dr. Reinkensmeyer’s laboratory has developed several technologies aimed at expanding access to rehabilitation, including the ArmeoSpring, MusicGlove, and FitMi. These devices use different approaches to encourage repetitive practice, from supporting arm movements to transforming hand exercises into interactive experiences.
Yet creating a successful medical technology requires more than advanced engineering. A device can only improve recovery if patients actually use it.
"A key barrier to adoption is complexity," Dr. Reinkensmeyer explained. "If the device is too complex, then people won’t use it."
This creates a difficult balance. Engineers must maximize therapeutic benefit while minimizing the burden placed on patients, clinicians, and healthcare systems.
"Right now, for almost any level of device complexity, you seem to get about the same therapeutic benefit, which is modest," he said. "If we could increase the slope of that curve, then I think people would be more willing to tolerate complexity to get a bigger benefit."
The challenge is not simply building more sophisticated machines. It is designing technologies that provide meaningful improvements while remaining practical enough to become part of everyday rehabilitation.
Combining Robotics With Regenerative Medicine
The future of neurological recovery may not rely on robotics alone. As regenerative medicine advances, researchers are beginning to explore how biological therapies and engineered systems can work together.
Stem cells and other regenerative approaches may eventually provide new neural tissue or restore damaged biological systems. In neuro-rehabilitation, this biological framework—the physical brain tissue and nerve pathways that carry movement signals—is known as the neural substrate. However, rebuilding biological structures is only the beginning. The nervous system must still learn how to use those new connections.
"Robotic devices can help people optimize their remaining neural substrate, but if the substrate is limited, then even an optimized substrate will still be limited," Dr. Reinkensmeyer explained.
"Stem cells have the potential to provide new substrate. But raw substrate is no good by itself. It needs to be shaped by practice."
This idea represents a broader principle in biomedical engineering. Biological repair creates possibility, but engineering provides the structure needed to transform that possibility into function.
A regenerated neural pathway may provide the foundation for recovery. Robotic training may provide the instructions.
The Future of Personalized Neurorehabilitation
Today, many rehabilitation programs rely on standardized approaches. However, every patient’s injury, remaining neural capacity, and ability to recover are different.
Dr. Reinkensmeyer believes the future of rehabilitation will become increasingly personalized through computational modeling. By combining measurements of a patient’s nervous system with mathematical models of learning, engineers may eventually predict the most effective training strategies for each individual.
"I believe we have a good chance to mathematically model, on a person-specific basis, how much a person’s remaining neural substrate can be optimized, and what they should train and how much they should train," he said.
This could transform rehabilitation from a generalized process into an individualized engineering system. Instead of every patient following the same exercises, future technologies may determine precisely what movements to practice, how much assistance to provide, and when to increase difficulty.
The goal is not simply recovery; it is optimized recovery.
Engineering the Future of Movement
The future of neurorehabilitation will not come from choosing between biology and technology. It will come from understanding how the two interact.
Robotic systems cannot replace the brain’s ability to learn, but they can create environments where learning becomes possible. Regenerative therapies may provide new biological opportunities, but those opportunities must still be shaped through experience and practice.
Dr. Reinkensmeyer’s work represents a shift in how engineers approach medicine. The most advanced rehabilitation technology will not be the machine that does the most for the patient. It will be the machine that understands when to assist, when to challenge, and when to let the brain take over.
By designing robots that communicate with the nervous system, biomedical engineers are beginning to redefine what recovery means. The future of rehabilitation may not simply be restoring lost movement, but engineering new pathways between the brain, the body, and the technologies that help reconnect them.