Emory Report

September 27, 1999

 Volume 52, No. 6

Emory-Georgia Tech make world's first 'leech-ulator'
It sounds like the stuff of science fiction--combining real biological nerve cells and electronics into a system that can perform computations--but that's the goal of collaborative research being conducted by Emory neuroscientist Ron Calabrese, two Georgia Tech biomedical engineering professors and Emory MD/PhD student Paul García, who serves as a the lead experimenter and "cross-pollinates" between the three labs.

The unusual biological calculator-which the group jokingly refers to as a "leech-ulator"-- actually uses living neurons from leeches. But as Calabrese emphasized: "The goal is not to make a calculator out of leeches. This particular research is specific to neurons, rather than leeches."

Leech neurons were a logical choice for this project since Calabrese has used leeches in his research for about 25 years. "I know the cells like I know my own children," he explained. Another big advantage is that this research is unlikely to incite any animal rights activists.

According to García, the first step was to take living neurons and do simple addition using those neurons and electronics; two separate nerve cells in nerve tissue (ganglia) were placed separately in a solution to keep them alive. Then electrical recordings were made from them, and those recordings were interfaced with a computer to get addition.

"We encode numbers into the firing of these two neurons--one number into each neuron," García said. "Then we connect the neurons with an artificial, one-way synapse, or connection, that simulates a realistic biological connection. The answer can then be interpreted by examining the firing of the receiving cell when it has both the influence of the number originally encoded into it plus the influence of the other cell with an encoded number."

The ultimate goal is to actually have the neurons in cell culture, rather than in pieces of neural tissue, Calabrese said. "The neurons will sit in cell culture and actually live on top of electronic chips which will then directly interface the cells with the digital computer." Right now the computer and neurons interface through electrodes stuck in the cells.

The research has both biological and engineering implications. Since neurons in the brain are involved in performing logic, math and other complicated functions, the researchers are interested in figuring out how neurons actually achieve these complex tasks. Another goal is to see if it is possible to use the properties of neurons that allow them to perform computations, like addition, to build hybrid systems, comprising both biological and electronic components, that may improve computer technology.

García stressed that the six-month-old project is still in its infancy. "We're really just at the beginning," he said. Noting that the brain is made of billions of neurons with trillions of connections, he said, "We believe the neurons and their dynamics are responsible for our computation. We're trying to use them in different ways to see if we can get anything similar."

Ultimately they hope "to harness the complicated dynamics that neurons are thought to have and use them to try to build some kind of computer--not based on digital computing, but on chaotic computing," Calabrese said. "Obviously we're not going to replace the traditional computer for word processing; it does a perfectly good job. But there may be applications for which different kinds of computers would be more appropriate."

Calabrese said he and his colleagues are currently generating feasibility data so they can write funding proposals. "We're trying to do something a little innovative that has potential technological implications," he said. "We're hoping we can increase our efforts in this project by looking for external funding."

--Linda Klein



Return to September 27, 1999, contents page