Stroke victims soon might breathe easier as a result of a new method for diagnosing stroke patients developed by an Iowa State professor.

Julie Dickerson, assistant professor of electrical and computer engineering and co-author of a book on a data analysis technique called fuzzy logic, said she received a phone call from neurologist Cathy Helgason from the University of Illinois-Chicago two and a half years ago.

Dickerson said Helgason had read a chapter of her book and wanted her to analyze data she’d recorded from stroke victims.

“We started an initial study looking for multiple interacting causes of stroke,” she said. “The doctors could see from the patients that they did not just have one thing wrong with them; they had many things wrong.”

Dickerson said they found an important link between vitamin B12 and homocystiene while studying the relationships between the levels of seven vitamins, metabolic acids and amino acids in stroke patients. The seven variables can interact in different ways to cause a stroke.

Dealing with those seven different measurements to determine a patient’s cause of stroke has been difficult in the past using traditional statistical methods, Dickerson said.

“They were using several poor assumptions in the way they treat it,” she said.

The reliance on simple vitamin B12 levels is not sufficient in stroke diagnosis, Dickerson said.

“Today, there is a crisp threshold. If you’re below the threshold in levels of vitamin B12, you’re in danger of stroke. If you’re above the threshold, you’re fine. It doesn’t make a lot of sense,” she said.

Dickerson’s application of fuzzy logic in dealing with the seven variables could make diagnosis easier.

“Fuzzy logic does not call for a ‘yes’ or ‘no’ answer. It allows for the shades of gray that most people see every day, especially physicians. They’ll see a number of interacting variables that are present to different degrees,” she said.

And although the new diagnostic method still is being tested, Dickerson said it showed a lot of promise as being an effective tool for physicians.

“Fuzzy logic is related to traditional statistics, and it is a more convenient way to express conditional expectations in terms that human beings are comfortable with,” she said.