HANOVER – The U.S. Office of Naval Research has awarded two Dartmouth engineering professors $600,000 to develop an artificially intelligent computer system that can model surgeons' behavior and intentions in the operating room.

Professor Eugene Santos and adjunct engineering professor Dr. Joseph Rosen, who also is a plastic surgeon at Dartmouth-Hitchcock Medical Center and a professor of surgery at Dartmouth Medical School, plan to use computer-modeling software that Santos has spent the last two decades developing to prevent operating room mishaps ranging from simple miscommunications to incorrect amputations.

"Our goal is to come up with a model of the intentions and behaviors of all the players in an environment," Santos said.

Creating the new system involves first gathering data from surgeons about a wide variety of operating room scenarios and developing a mechanism for entering patient information into the computer program. Once that information is captured, Santos and Rosen plan to outfit an actual operating room with video cameras and voice sensors to record and process operations in real time. The computer then uses the audio and visual "evidence" to model the surgeon's potential intentions at each step in the procedure, Santos said. If the surgeon deviates from his intended course, a monitor display would alert him and ask for justification, thus reducing the risk of unintended consequences.

"That whole approach has never been done in the operating room," Rosen said. "We have check lists and charts and forms "¦ but you can override that." Presently, when a surgeon does override information from charts or colleagues during surgery, whether intentionally or unintentionally, not everyone in the operating room is always informed. The new computer system would ensure that all the players have the same information and that it has been communicated correctly.

Santos said he hopes to capture as much as 90 percent of the thousands of variables that impact an operating room team during a surgical procedure, such as how many surgeons are performing the operation and whether a patient has an allergic reaction to a certain medication. As a result, one of the challenges facing Rosen and Santos as they develop the new system will be figuring out how much information is too much.

Surgeons need the freedom to make quick decisions as circumstances in the operating room change. Too many buzzers and requests for justification could distract or delay the surgical team, Santos acknowledged; the goal is finding the right balance.

"If you have too much information, it almost feels like Big Brother, but that's not our goal here," he said.

By the end of the three-year grant period, Santos and Rosen hope to have a prototype of the operating room system. "We should have a functioning prototype and we'll actually be doing some validation studies," Santos predicted.

Eventually, they envision the project becoming a marketable product and a business venture.

Santos said working with Rosen, who spends much of his time in an operating room, gives the project a more real-world feel, compared to some of the other applications for which his modeling software has been used.

For the last 20 years, Santos has been coming up with algorithms and developing software that models human decision-making. Working with the Florida-based defense contractor Securboration and other companies, he has employed the technology in various capacities, from intelligence analysis and war gaming to modeling the multiplication of breast cancer cells.

"You can take it on all different levels, from social and cultural modeling all the way down to the cellular level, where you can model how cells react and communicate and form tissues, which form organs," Santos said. "One of the things that we're already exploring is how can we model society? - How do I measure, say, political will?" Regardless of the scenario, modeling potential outcomes gives people more information and helps them make better decisions, Santos said.

"I don't want to say we have a crystal ball -- that's too much -- but we can say, These are the things that possibly could happen.'"