Scientists have built the first silicon chip equipped with living nerve cells. The "neurochip," a silicon rectangle about 4 centimeters wide immersed in a petri dish, may be the forerunner of bionic eyes or other medical devices engineered from combinations of silicon and living neurons. For now, neurobiologists hope to use the device, described this week at the meeting of the Society for Neuroscience in New Orleans, for a humbler purpose: understanding how nerve cells grow and communicate with each other.
Biologists have studied individual neurons for years, and they can listen in on the chatter of neurons in nerves and brains. But trying to learn how neurons communicate from such measurements was like trying to learn basic electronics by studying the workings of a computer. Any electronic engineer learns by building small circuits first, so neuroscientists have long wanted to build simple circuits of neurons. But when they tried to link living nerve cells, they usually ended up injuring or killing them.
Jerome Pine, a neurophysicist at the California Institute of Technology in Pasadena, along with a team of electrical engineers and biologists, created a microscopic silicon landscape that confined individual neurons, while allowing them to establish connections: a set of 16 tiny wells, each about 1/40 of a millimeter in diameter, with short tunnels leading to the surface. The researchers placed an embryonic rat brain cell in each well; as the cells grew, they sent out long dendrite arms through the tunnels toward neighboring wells. Wires in the underlying silicon monitored the electrical behavior of the neurons.
Eventually the dendrites made contact with those of neighboring cells and established normal electrical activity. "Now," Pine says, "the biggest challenge is maintaining a healthy network." So far the researchers have only been able to keep the cells alive for two weeks at a time. Once they can keep the circuit going for long enough, perhaps a month, they hope to study, for example, how small sets of neurons "learn" after being stimulated repeatedly.
"The work will certainly be important for future medical applications," says Peter Fromherz, a neurophysicist at the Max Planck Institute for Biochemistry in Munich, Germany, who is also trying to marry neurons and silicon. "The retina, with its planar structure, may be the best system" to try to imitate with flat neurochips, he says. But such applications are a ways off. Says Pine, "You shouldn't expect anything in our lifetime."