-State- -1st Letter of City- Submit Adverti Italian soccer powerhouse looks to artificial intelligence for edge on the field Pila dr Try More Arsenal's Robert Pires receives treatment on his injured right knee during FA Cup soccer action in March. The injury forced Pires out of the game. By Linda Carroll MSNBC CONTRIBUTOR May 29 The ink was barely dry on Fernando Redondo s multi-million dollar contract with the soccer powerhouse AC Milan when the player blew out his knee just three minutes into a routine workout on a treadmill. His pain was felt acutely by team owners, who were left to wonder how this might have been foreseen and prevented. Take our interactive quizzes Sign up for our free health e-newsletter 1 of 6 6/3/2002 8:15 AM
Yellow Pages Auctions at ubid Personals Channel Professional Tips Newsletters Shopping Right now we ve got pretty rudimentary tools to predict injuries with. It would be nice to have tools that would allow us to see more subtle abnormalities. DR. ARTHUR BARTOLOZZI chief of sports medicine at the Pennsylvania Hospital in Philadelphia FOR YEARS team doctors and coaches have looked for crystal balls that would show ACL injuries in the making, soothsayers who could hold forth on hamstrings that might blow, genies that could warn of a rotor cuff about to explode in the new hot prospect s shoulder. AC Milan may have found such an oracle: a computer smart enough to recognize the signs of an athlete coming apart. The renowned Italian soccer club which has four players competing in this year s World Cup has teamed up with Computer Associates International to test the feasibility of using neural networks, a form of artificial intelligence, to predict injuries and optimize conditioning for each athlete, perhaps even to help select which players to sign. NEURAL NETWORKS Neural networks are different from traditional computers in that they learn to recognize patterns instead of being given a set of rules in the form of a computer program. So, in the case of sports injuries, the neural network is shown many cases of athletes who have been injured. The computer is fed data on each player s performance and habits in the weeks and months prior to the injury. 2 of 6 6/3/2002 8:15 AM
2002 World Cup Pool play: May 31-June 14 Mexico beats Croatia 1-0 Batistuta's big goal lifts Argentina over Nigeria Sweden, England tie Spain ends opening-round jinx South Africa, Paraguay tie Bondy: For U.S., it will be three... and out Interactive: U.S. World Cup stars to watch Newsweek preview: Italy over Argentina for title All-world World Cup team Top Cup contenders Top Cup playmakers Team-by-team capsules Daily scoreboard Day-by-day schedule Schedules by group Group standings Vote: Who will win? More on soccer Do you think sports teams should be allowed to use neural networks to help select which players to sign? Yes No In theory, after the computer has been exposed to many cases, it will see what led up to the injury and be able to alert coaching staff the next time it recognizes a similar pattern in another player. Computer Associates isn t keen on giving out the details of the system designed to give AC Milan the edge. Matteo Aliberti, project manager for Computer Associates, said that the players are tested every 15 days. But he declined to say exactly how the athletes were being monitored. We re looking at the fitness of the athlete, Aliberti said. We re looking at how they react to external stimuli; how they move; how they train; what they eat. Jean Pierre Meersseman, the head of AC Milan s medical team, provided a bit more insight into the process. Players are asked to wear sensors about 18 to 24 of them on their bodies during workouts. The sensors transmit information back to the neural network via radiowaves. This way the computer gets feedback while the athlete plays soccer, runs a 40-meter dash or works out on the machines in the AC Milan gym. The computer also is fed psychological data, as well as information on what each player eats. 3 of 6 6/3/2002 8:15 AM
PROMISING RESULTS An 18-month test of the system gave promising results, according to Aliberti. We were able to run the test because [AC Milan] had a lot of data on its athletes already, he Vote to see results said. The pilot program showed that injury prediction was a possibility, Meersseman said. We had information from more than 5,000 tests on our players done over the last four years, he explained. We put this into the network to see if certain parameters changed before a player was injured. Ultimately the neural network correctly predicted injuries 84 percent of the time, Meersseman said. The mathematicians think they can get this number up to 96 percent, he added. The new system could be a great help to coaches trying to predict which players are at risk, said Dr. Arthur Bartolozzi, chief of sports medicine at the Pennsylvania Hospital in Philadelphia. A lot of factors can contribute to injury: conditioning, fatigue, equipment, weather conditions, surface conditions, Bartolozzi said. Advertisement If a system could be developed to take all of these things into account, it might allow team doctors to pull players before they get injured, he added. Right now we ve got pretty rudimentary tools to predict injuries with, he said. It would be nice to have tools that would allow us to see more subtle abnormalities. The program might also have saved AC Milan a lot of money, had it been around when the team was bidding for Fernando Redondo. We spent an enormous amount of money buying a player who got hurt after three minutes on the treadmill, Meersseman said. Maybe we would have thought twice about buying him. Maybe the price would have been different. And if the team had gone through with the deal anyway, the neural network might have told them how to avoid the big injury. Once we buy a player we should be able to monitor him to see if his condition is declining, Meersseman said. Then we could adapt the training schedule. If the new technology turns out to be everything promised, it could make things more accurate and thus fairer for athletes, Bartolozzi said. Right now, before signing a new player, doctors and coaches do their best to predict whether a player will stay healthy. But the assessment is in many ways unscientific, he said. The new system might allow teams to make a more accurate determination of a player s future fitness, he added. Beyond that, Bartolozzi said, It s very intriguing to be 4 of 6 6/3/2002 8:15 AM
able to study players in real time. Bartolozzi sees other possible uses of the new technology. We might be able to use it to figure out the differences in injury rates between male and female athletes, he said. Now that more girls are playing soccer and other sports, we re seeing more of them tearing their ACLs. This might help us explain the sex difference. Penalty-kick tiebreakers Bad passes back to the goalkeeper Yellow cards The hooligans Offside traps The goalkeeper and his teammates Weird formations Different nations display very different styles. England and Ireland prefer the long balls, and high crosses. Italy utilizes short passes and, like Argentina, tries to beat you 1-0. France and Portugal have creative midfielders who dominate with playmaking. The Germans and Poles are extremely physical. The African nations, and Cameroon in particular, are known for a more open, daring style. Brazil changes pace on a dime, and still tries to win with individual technical flair. Printable version When it comes to predicting injuries, ultimately, the neural networks may simply give validity to what team doctors often know intuitively. If I think I see something wrong with an athlete and can t say exactly what it is, it s hard to tell the coach to not use a player, Meersseman said. He wants the player to play if there s nothing really wrong. I hope now I ll have the numbers to say how much possibility there is that this player will be injured. It will give a little more power to what I say. 5 of 6 6/3/2002 8:15 AM
WashPost: Hospital embraces remote ICU Toxic tubing? TV linked to obesity in very young Toxins pose risks for many in U.S. One couple's battle to start a family MSNBC Cover Page MSNBC's Health Library Return to Health front page MSN Health MSNBC VIEWER'S TOP 10 Would you recommend this story to other viewers? not at all 1-2 - 3-4 - 5-6 - 7 highly MSNBC is optimized for Microsoft Internet Explorer Windows Media Player MSNBC Terms, Conditions and Privacy 2002 Cover News Business Sports Local News Health Technology & Science Living & Travel TV News Opinions Weather Comics Information Center Help News Tools Jobs Write Us Terms & Conditions Privacy 6 of 6 6/3/2002 8:15 AM