AI-powered insoles could spot early Parkinson’s symptoms

By Stephen Beech

A new "smart" insole that monitors how people walk, run and stand can improve posture and even provide early health warnings.

The wearable device can detect early signs of conditions, including Parkinson’s disease, say American scientists.

Constructed using 22 small pressure sensors and powered by small solar panels on the tops of shoes, the new system offers "real-time health tracking" based on how a person walks.

Detailed personal health data is transmitted via Bluetooth to a smartphone for quick and detailed analysis, according to a study published in the journal Science Advances.

Study co-author Dr. Jinghua Li, Assistant Professor of materials science and engineering at The Ohio State University, said: “Our bodies carry lots of useful information that we’re not even aware of.

“These statuses also change over time, so it’s our goal to use electronics to extract and decode those signals to encourage better self-health care checks.”

It is estimated that around one in 14 people suffer from "ambulatory difficulties" - activities that include walking, running or climbing stairs.

While efforts to manufacture a wearable insole-based pressure system have risen in popularity in recent years, many previous prototypes were met with low energy limitations and unstable performance.

To overcome those challenges, Dr. Li and study lead author Qi Wang, a PhD student in materials science and engineering at Ohio State, sought to ensure that their wearable was durable, had a high degree of precision when collecting and analyzing data, and provided consistent and reliable power.

Dr. Li said: “Our device is innovative in terms of high resolution, spatial sensing, self-powering capability, and its ability to combine with machine learning algorithms.

“So we feel like this research can go further based on the pioneering successes of this field.”

She said the system is also made unique through its use of AI.

Using an advanced machine learning model, the wearable can recognize eight different motion states, including static ones such as sitting and standing to more dynamic movements like running and squatting.

Dr. Li says that since the materials the insoles are made of are flexible and safe, the device - much like a smartwatch - is "low-risk and safe" for continuous use.

After the solar cells convert sunlight to energy, that power is stored in tiny lithium batteries that don’t harm the user or affect daily activities.

The researchers could see how the pressure on parts of the foot is different in activities, such as walking compared to running, because of the distribution of sensors from toe to heel.

Dr. Li explained that during walking, pressure is applied sequentially from the heel to the toes, whereas during running, almost all sensors are subjected to pressure simultaneously.

She said, in health care, the smart insoles could support gait analysis to detect early abnormalities associated with foot pressure-related conditions - such as diabetic foot ulcers, musculoskeletal disorders - such as plantar fasciitis, and neurological conditions, including Parkinson’s disease.

The new system also used machine learning to learn and classify different types of motion.

The researchers said that offers opportunities for personalized health management, including real-time posture correction, injury prevention and rehabilitation monitoring.

They say customized fitness training may also be a future use.

The smart insoles showed no notable deterioration in performance after 180,000 cycles of compression and decompression, according to the study.

Dr. Li said: “The interface is flexible and quite thin, so even during repetitive deformation, it can remain functional.

“The combination of the software and hardware means it isn’t as limited.”

The researchers expect the technology will likely be available commercially within the next three to five years.

Next steps to advance the work will be aimed at improving the system’s gesture recognition abilities, which, according to Dr. Li, will likely be helped with further testing.

She added: “We have so many variations among individuals, so demonstrating and training these fantastic capabilities on different populations is something we need to give further attention to.”