Hearables: The new health centers on the ear

Over the next five years, researchers from the Universities of Oldenburg and Bremen want to investigate what additional tasks hearing aids can take on as part of the HEARAZ (Hearable-centered Assistance: From Sensor to Participation) research training group.

Prof. Andreas Hein, professor of automation and measurement technology at the University of Oldenburg and spokesperson for the new research training group, talks to MEDICA-tradefair.com about the new functions that could be integrated.

Prof. Hein, what could the devices be used for in the medical field in the future?

Prof. Andreas Hein: The vision of hearables is that both hearing aids and miniaturized headphones will record health data using additional sensors. This would allow us to reach younger target groups and monitor developments over a longer period of time in order to detect deficits or the development of diseases at an earlier stage.

With additional sensors in the wearables on the ear, for example, oxygen saturation, pulse and heart rate variability can be measured, as well as brain activity with limited EEG. The data is then transferred to the smartphone, where we carry out an AI-based analysis.

We want to recognize patterns and symptoms, such as difficulty swallowing. They can be measured as acceleration at the ear and as activation of the chewing muscles. This allows us to detect abnormal behavior, but we will not perform any diagnostics.

What developments are necessary for this?

Hein: We collect and validate diagnostic information for early detection. In the field of geriatrics, for example, we are looking at mobility and the risk of falls. In neurology, we are looking at behavioral changes that occur over the course of Parkinson's, such as mobility disorders or swallowing disorders.

For these clinical pictures, we can derive information from sensor technology close to the ear. The results must be translated for the treating physicians and processed by AI. As we observe hearing aid wearers throughout the day in their normal lives, we expect a significantly higher variance and need to understand the context. We may collect sensitive data, but for privacy reasons we limit ourselves to certain characteristics of speech. For example, we look at fluency, inflections and possible changes in the voice due to emotional or physical changes. The content of conversations is not recorded.

Which projects will be carried out at the start?

Hein: In the funding, we have planned ten doctoral theses starting in parallel that cover the spectrum of data acquisition described. This involves adapting the hardware, processing the sensor data, linking it to clinical processes, as well as questions of acceptance and privacy and integrating future users.

We will select the topics for the second cohort of doctoral students in around three years' time based on the approaches that have proven successful. We will then focus on results that can be quickly transferred into practice.

Are there any issues that are considered particularly urgent?

Hein: Sensor technology is already integrated into many hearing aids, i.e. acceleration sensors and the audio channel, which can be used to analyze mobility and determine the risk of falling. As we can build on the hardware of existing products, the focus here is on the development of software and its clinical application.

It makes sense to examine the interaction between mobility and speech production in particular. If I can no longer walk and speak at the same time, this is an indication of impaired cognitive abilities. This leads to an increased risk of falling or can be an early indicator of cognitive changes, dementia or Parkinson's disease. I see potential here for early implementation in products and services in the healthcare sector. In the future, the collected and analyzed data should be integrated into the telematics infrastructure so that the doctors treating patients can access it directly.