When the body does the thinking
By studying how the sea slug reconfigures its body to reject or bite food, CMU researchers uncovered new evidence that intelligent behavior emerges not only from the brain, but also from the body’s mechanics.
Made entirely of muscle, with no rigid skeleton, muscular hydrostats like octopus arms, elephant trunks, and human tongues can move in ways traditional mechanical systems cannot. Between crawling across the sea floor, picking up a peanut, and enabling speech, scientists have long wondered how soft structures can perform so many different functions without becoming impossible to control.
Researchers in Vickie Webster-Wood’s lab have uncovered part of the answer. A new study published in iScience found that the muscular hydrostat feeding organ of the sea slug, Aplysia californica, doesn’t rely on the nervous system to solve its every movement, rather it physically reconfigures its anatomy depending on the task at hand. This allows the sea slug to efficiently bite or reject food.
These findings, built on past research from the lab, offer new evidence in the growing field of integrated mechanical and computational intelligence, which presents the idea that intelligence emerges not only from computation in the brain, but also from the design and mechanics of the body itself.
“This idea of the body being important in computation and control is not new, and there are many names for different aspects of this phenomenon, from mechanical intelligence to physical intelligence,” said Webster-Wood, an associate professor of mechanical engineering. “While there are slight nuances to each term, the ideas boil down to the concept that the dynamics and mechanics of the physical system, or its interaction with the environment, can take the place of a higher-level controller. Essentially, instead of having to control every minute detail of motion with a computer, the mechanics of the body simplify the type of control signals and networks needed.”
Using MRI recordings of the slug eating and a newly developed biomechanical model, the researchers discovered that two feeding behaviors that appear very similar to the naked eye actually rely on entirely different mechanical strategies.
When the sea slug rejects food, it elongates part of its feeding structure, stretching a muscle so it can generate greater force to push food away. During biting, however, that same strategy wouldn’t work. Instead, surrounding muscles bend and wrap around the feeding organ, redirecting forces so the animal can hold on to the food.
To the human eye, the motions look nearly identical, but how the sea slug rearranges its muscles to interact entirely changes the way its body accomplishes the different tasks.
“One of the most exciting aspects of this project is how it helped us identify a set of biomechanical tools that aren’t just used by the slug, but that might be used by a whole subclass of muscular hydrostats,” said Michael Bennington, first author of the research paper and Ph.D. candidate in Webster-Wood’s lab. “So while the study was initiated to answer a specific question about Aplysia behavior, it’s given us really interesting insights into hydrostats in general and the mechanical control strategies that they might be using.”
Essentially, instead of having to control every minute detail of motion with a computer, the mechanics of the body simplify the type of control signals and networks needed.
Vickie Webster-Wood, Associate Professor, Mechanical Engineering
This finding not only reveals a new principle for how soft biological systems work, it can also guide the design of more capable soft robots. Instead of trying to control every movement with complex software, future soft robotics may be able to let their physical structure do more of the work.
“Biology provides proof of existence that autonomous, adaptable, multifunctional behavior can be achieved in robust, energy-efficient, soft-bodied systems,” said Webster-Wood. “Biological brains are far more efficient than state-of-the-art algorithms. By studying how biological organisms’ brains and bodies work together to create this behavior, we can learn new principles to bring to brain-inspired computing and robotic design to help make future soft robots and controllers more capable and efficient.”