An electrocardiogram, also known as an ECG, is a medical test that monitors the state of the heart by recording electrical signals produced by the heart. Many diseases throughout the body display cardiac symptoms that can be picked up by an ECG. These ailments include maladies such as traumatic brain injuries, COVID-19, and more.

Diagnosing a patient for such maladies is difficult to do from an ECG because the visual indicators are small and can be hard to notice, even to a trained eye. Deep learning, however, has proven to be a powerful tool in solving such detailed problems. Nathan Chapman, Jack Flitcroft and Kyle Rodriguez are a team of seniors majoring in Computer Science at Milwaukee School of Engineering (MSOE) who are working with the Medical College of Wisconsin to build Deep Learning based models that predict whether a patient has a high likelihood of having Covid-19 or a traumatic brain injury.

They are training deep learning models using Rosie, MSOE’s supercomputer, based on the data collected from ECGs.

“ECG’s are really inexpensive and quick medical tests, and provide a lot of information about the body. The issue is that to the human eye, it’s difficult to say the least, to gather from an ECG whether someone has some illness. This is where AI comes in,” said Flitcroft. “AI is really good at spotting small differences within a healthy vs unhealthy ECG, and with this we can predict whether someone has a number of different health conditions from this simple test.”

Flitcroft and his teammates are taking this technology one step further and putting it directly in the hands of anyone who owns a smart watch. “Even smart watches have ECG’s in them now,” he said. “Imagine if you hit your head hard and want to know if you have a concussion. We are building an app to take your ECG on the smart watch, and let you know if the model thinks you have a concussion.”

From athletes to the average person, everyone is at risk for concussion after a big bump or fall. This new technology is another tool to encourage individuals to seek medical attention.