A team from the University of Chicago have produced a flexible computing chip able to process huge amounts of health information using brain-mimicking artificial intelligence. The device can analyze health data in real time, a breakthrough that could improve wearable health technologies and how health data is processed in the future.
As part of the transition from hospital or clinic visits to continuous health tracking by wearable electronics, wearable biosensors are being developed that are unobtrusive and can track indicators of health such as levels of oxygen, sugar, metabolites and immune molecules in people’s blood.
As these biosensors start to collect more and more data, its analysis is becoming more difficult since all information has to be seen in the broader perspective of a patient’s history and other health parameters. While smartphones can’t achieve the analysis required, innovative artificial intelligence platforms that can integrate machine learning to find patterns in large datasets can. However, transmitting this information from a device to a centralized AI location is slow, energy inefficient and invasive in terms of privacy.
The researchers therefore looked to design a chip that could collect data from multiple biosensors and draw conclusions about a person’s health using machine learning. It was also key that it was wearable and integrated seamlessly with skin, something smart watches, for instance, cannot manage.
As reported in the journal Matter [Dai et al. Matter (2022) DOI: 10.1016/j.matt.2022.07.016], polymers, which are used in semiconductors and electrochemical transistors but can stretch and bend, were assembled into a device that allowed the AI-based analysis of health data. Rather than work like a typical computer, the chip – called a neuromorphic computing chip – functions similar to a human brain, and can store and analyze data in an integrated way.
To test the device, electrocardiogram (ECG) data representing the electrical activity of the human heart was assessed by training it to classify ECGs. It was shown to accurately classify heartbeats regardless of whether the chip was stretched or bent, and based on real-time information the device was shown to make intelligent decisions about adjusting medication levels. As researcher Sihong Wang said: “With this work we’ve bridged wearable technology with artificial intelligence and machine learning to create a powerful device which can analyze health data right on our own bodies”.
Although it remains early research for integration of AI with wearable electronics, the team hope to test the device’s power in assessing patterns of health and disease, sending patient or clinician alerts, and automatically tweaking medications, as well as trying new iterations to increase the type of devices with which it can integrate and the types of machine learning algorithms it can use.
“With this work we’ve bridged wearable technology with artificial intelligence and machine learning to create a powerful device which can analyze health data right on our own bodies”Sihong Wang
The wearable neuromorphic chip, made of stretchy semiconductors, can implement artificial intelligence (AI) to process massive amounts of health information in real time. Above, Asst. Prof. Sihong Wang shows a single neuromorphic device with three electrodes. Photo by John Zick