A new tool created by researchers can diagnose a stroke based on abnormalities in the patient’s speech ability and muscular movements in the face, and with accuracy – all within minutes of interacting with a smartphone .
According to a study, researchers have developed a machine learning model to aid and possibly speed up the diagnostic process by physicians in a clinical setting.
“Currently, physicians must use their past training and experience to determine at what stage a patient should be sent for a CT scan,” said study author James Wang of Penn State University in the United States.
“We try to simulate or follow this process using our machine learning approach,” Wang added.
The team’s new approach analyzed the presence of stroke among actual emergency patients with a suspicion of stroke using computation of facial movement analysis and natural language processing to identify abnormalities in a patient’s face or voice, such as a drooping cheek or sloppy speech.
To train the computer model, the researchers built a dataset of more than 80 patients at Houston Methodist Hospital in Texas.
Each patient was asked to perform a speech test to analyze their speech and cognitive communication while it was being recorded on an Apple iPhone.
“Obtaining facial data in natural conditions makes our work robust and useful for clinical use in the real world, and ultimately empowers our method for remote diagnosis of stroke and self-assessment,” Huang said.
By testing the model on the Houston Methodist dataset, the researchers found that its performance achieved 79 percent accuracy – comparable to clinical diagnostics by emergency physicians, who use additional tests such as CT scans.
However, the model can help save valuable time diagnosing a stroke, with the ability to assess a patient within four minutes.
