The new flexible wearable device mimics the brain to analyze health data
A research team at the University of Chicago has developed a stretchy, flexible wearable device that records health data and processes it in a similar way to the functioning of the human brain. The market today offers a variety of wearable fitness bands and other health devices.
Despite this, most cannot analyze the patient's baseline measurements or spot signals of disease in a complex manner. The potential of artificial intelligence in this regard can be utilized to fill in the chasm between these two worlds. It is possible to detect patterns in complex data sets by using machine learning techniques.
The information that is sent from a device to a centralized AI location is however inefficient and energy consuming, so it is not desirable to send it there. The new study aimed to develop a chip that collects data from multiple biosensors and also uses artificial intelligence to conclude a person's health. A smartwatch always leaves a gap.
"We wanted something that could achieve very intimate contact and accommodate skin movement," explained Sihong Wang, an Assistant Professor of Molecular Engineering. The study published in Matter was also co-authored by Wang. In addition to being flexible and stretchy, polymers can also be used to make semiconductors and electrochemical transistors.
The polymers were incorporated into a device that was able to be used by artificial intelligence as a means of processing the data. Known as neuromorphic computing. The chip operates less like a computer than it does like the brain of a human. As a result of this, it can store and analyze information in an integrated manner, which makes it very effective.
As part of the research, researchers have also tested the effectiveness of the device in terms of analyzing electrocardiogram data, which is the electrical activity of the heart. A device was trained to form four categories out of the data and it was found to be able to accurately determine whether a chip was bent or not once the device was trained to classify the data.