Rat-brain controlled robot aids memory research

University of Reading researchers in the U.K. study memory formation and destruction from a robot controlling its own movements with rodent brain matter.

A robot using biological brain matter from rodents to control its movements is helping researchers learn more about human neurological diseases such as Alzheimer's and Parkinson's, according to University of Reading researchers in the U.K.

The robot represents a multidisciplinary effort within the University of Reading, whose team includes Kevin Warwick, head of Cybernetics in the School of Systems Engineering, and Ben Whalley, pharmacist and professor in the School of Pharmacy.

University of Reading robot and the multi electrode array containing the rodent brain matter that controls it. University of Reading

The neuron culture being used to control the robot operates from what is essentially a sophisticated, temperature-controlled Petri dish with electrodes called a multi electrode array (MEA).

The MEA, in which the delicate brain matter is kept in an incubated environment and fed, consists of 60 electrodes. After about one to two weeks of growth, the brain matter is mature enough to start learning and use the MEA to communicate with the robot. The MEA both stimulates the brain matter by sending electrical signals to it and responds to the electrical impulses sent out by the cultured neurons.

"We feed them every couple of days, a pink liquid with nutrients not too dissimilar to what the Olympians might drink for energy," Warwick said in a phone interview. "It keeps the neurons alive and allows them to grow. Within 24 hours they make connections. Within a week there is a brain-like activity. If you stimulate one electrode you get spontaneous firing. We then use that basic operation by linking it up to the robot body."

The electrical output from the brain received by the MEA is transferred to the robot via ultrasonic sonar. The sonar signal causes the robot's wheels to move forward, left, or right, depending on the signal. In return, the brain is sent signals to stimulate it when it nears an object in the hopes that it will respond to avoid colliding with it.

The process has been working so well, the researchers are now considering adding infrared sensors and possibly even an audio component to enhance the robot's functionality, according to Warwick.

"One of the things we could look to do is to put sound in and get sound out which essentially means communicating," he said. "There's no reason why we can't do that and I see that as something. What it will mean to the (neuron) culture by putting audio in and getting different audio out, is that it would allow us to whistle to it and teach it in some way to respond back."

Using the current system, the researchers have been able to teach the robot how to navigate around obstacles and to respond to specific signals with specific movements and positions. This has allowed them to monitor how the brain learns and forms memories as the robot learns how to recognize where it is and navigate itself accordingly.

"Simply from the habit of doing things repetitively the neuron connections strengthen and allow the robot to be able to do it more accurately," Warwick said. "In terms of teaching the cells to do more, that is part of the research. Can we get it to improve by applying different voltage and chemicals and things?"

The researchers have worked on this project with several different batches of brain matter used for months at a time. The brain matter currently being used has grown over three months from a tiny group of neurons to about 100,000 neurons--roughly the size of a brain in a bee or wasp.

After the researchers track how the rodent brain matter learns and creates memories on a neurological level, they move to phase two.

The researchers purposefully damage the brain matter in an attempt to recreate what happens to a brain with trauma, or diseases like Alzheimer's and Parkinson's. From there, they are studying the brain's reactions to damage on a neurological level in conjunction with how it operates with memory and movement on the robotic level.

"We are doing that now, applying chemicals to part of the brain structure so that it stops growing and the neurons will die off and then see if the actions the robot has been doing are taken over. So it's like looking at what happens in a stroke," Warwick said.

"Can the brain learn to do things it used to do? How can it learn to do the things it used to do? The key thing is--as we understand things a bit more--is to put it into different positions or a particular situation and teach it so that we can see how the memory of what to do in that situation is constructed in the brain by the connection," he said.

Rodents have been thought to be a promising specimen source for learning more about memory formation.

Last year, researchers in Europe discovered that a cellular molecule in the brain's hippocampus is responsible for mice's ability to learn, build memories, and recall them.

Other multidisciplinary efforts to incorporate biological brain matter and robotics are also underway at other leading universities. In 2003, a group from the Georgia Institute of Technology created a robot with biological brain matter that it dubbed a "hybrot."

About the author

In a software-driven world, it's easy to forget about the nuts and bolts. Whether it's cars, robots, personal gadgetry or industrial machines, Candace Lombardi examines the moving parts that keep our world rotating. A journalist who divides her time between the United States and the United Kingdom, Lombardi has written about technology for the sites of The New York Times, CNET, USA Today, MSN, ZDNet, Silicon.com, and GameSpot. She is a member of the CNET Blog Network and is not a current employee of CNET.

 

Join the discussion

Conversation powered by Livefyre

Show Comments Hide Comments