Stroke patients watch their virtual image on screen and respond as virtual tennis balls are thown at them from all directions.Israeli computer scientists have developed a new virtual reality program for stroke victims, that promises not only to be able to diagnose patients, but also, eventually, calculate the probability of recovery and recommend the most effective treatment option for the patient.

The program, which is already in use in Israeli hospitals, was created by researchers Dr. Larry Manevitz of the University of Haifa, Dr. Uri Feintuch, a neuroscientist from Hebrew University, and Eugene Mednikov, a computer science graduate student. It uses a variety of video sessions to differentiate between different types of brain injury, cerebrovascular accident (CVA), and traumatic brain injury (TBI).

In a typical session from this program, stroke patients watch their virtual image on a screen and respond as tennis balls are thrown at them virtually from all directions. The patient’s actual hand motions are recorded on the screen.

These video sessions are then fed into the newly developed computer program, which “learns” to differentiate between the different types of brain injuries. In further testing, it was found that 90-98 percent of the time, the computer program was able to accurately diagnose whether the patient was healthy, or had suffered a stroke or a traumatic brain injury.

While diagnosis is the most basic part of treatment, says Manevitz, from the university’s Department of Computer Science, any doctor or healthcare worker can correctly diagnose severe brain injuries. The significant advance here comes in the next phase of development for this program.

“As soon as the computer identified the injury, we have a model that we can use for further testing and analysis – something that cannot be done on live patients,” explained Manevitz. “Using a computer model, we can experiment with different treatment options and decide which will be the most effective. The computer can also define how much the patient will be able to rehabilitate. These are things that would take a long time for medicine to accomplish, and some of them cannot be done at all.”

For example, the computer can simulate how the patient will respond if the virtual reality therapy throws more balls to the patient’s left side than to the right, or if any other change would be beneficial for rehabilitation.

The computer can quickly examine tens of different possibilities in a very short time. Using the computer will help physicians and patients alike avoid spending time on treatments that will not benefit the patient, or worse, cause harm.

“Our next step is to find similarities in the behavior of people in sub-groups of brain injuries. The human eye may not be able to see such similarities, but a computer would easily be able to pick them up. As soon as we are able to identify similarities in different sub-groups, new avenues of effective treatment will open up for doctors,” said Manevitz.


Slashdot It!