New computer simulation may solve questions about human brain function

Playfully leaning back in his chair while sitting in his modest, dimly lit office, molecular biology professor John Hopfield talks about his creation with all the enthusiasm of a proud parent.

Hopfield's work area does not seem like the kind of place where the finishing touches would be made to an artificial organism, but that is exactly what he and Carlos Brody, a postdoctoral fellow at the Center for Neural Science at New York University, have done.

And last week, they introduced the world to their virtual creation — the sand mouse, or mus silicium.

For this project, Hopfield and Brody created a computer simulation of 660 artificial cells that behave exactly like real brain cells. These virtual cells, which form an artificial organism named for the silicon chips that power it, possess all the properties of real nerve cells.

Hopfield said the principles that drive the sand mouse's virtual network of neurons and synapses may shed light on how the human brain works.

While molecular biology research is typically based on observing organisms or organism systems that are not fully understood, Hopfield knows all the characteristics of his sand mouse. As a result, his project encourages neurobiologists to approach research from a unique angle — transcend the data and truly search for the concepts behind it.

"[In the field of molecular biology], we often say that we need to collect more data when we don't understand something, that something we have not yet discovered is responsible for what is going on," Hopfield said. "I thought that the members of the community would appreciate the opportunity to exercise their problem-solving skills and look carefully at the data, given that everything they need to figure it out is there."

As part of the simulation, the sand mouse cells release chemicals, connect to each other through synapses and fire action potentials — a type of electric current. Through these processes, the organism can recognize the spoken word "one," even in noisy conditions.

Such behavior — easily accomplished by the human brain — is practically impossible for machines to duplicate.

Hopfield and Brody's approach to soliciting feedback for their project has been almost as creative as the experiment itself. Rather than simply publishing their discoveries in a scientific journal or magazine, they constructed a Website that presents all data necessary for a user to determine how the organism's brain operates.

Colleagues and researchers experimenting with the sand mouse are then encouraged to determine for themselves the basic principles behind the organism's function.

To encourage other scientists to inspect his work, Hopfield is running what is, in effect, an online contest. Cash prizes will be awarded to those who determine by Dec. 1 how the organism recognizes the word "one" or apply their understanding of the organism to make it behave in a different way, Hopfield said.

Get the best of ‘the Prince’ delivered straight to your inbox. Subscribe now »

Hopfield will announce the answer to the challenge Dec. 14 in his MOL 437: Computational Neurobiology and Computing Networks class.

Hopfield said his Website had received about 500 hits last Tuesday before receiving approximately 15,000 hits between last Tuesday and Thursday.

Brody noted that the decision to publicize their results by arranging an online challenge has been somewhat controversial among many scientists.

"Senior scientists have not all taken to the idea," he said, "but we wanted to raise discussion of how people think in neuroscience, and we have been successful in doing that."

But for Hopfield, the experience of developing the sand mouse has been an education, as he hopes experimenting with it will be for his colleagues.