by A team of engineering researchers has developed a machine learning platform that can efficiently analyze and predict data collected by wearable medical sensors. The team applied this platform to a new stretchy, wearable throat sensor that monitors a user’s speech and swallowing patterns by recording vibrations and electrical muscle impulses from the neck area. The research, reported in Nature Communications, aims to provide a more accurate and convenient method for continuously monitoring the muscle and swallowing movements of patients with throat conditions.
Currently, the available devices used for monitoring throat conditions provide limited information and are often bulky and uncomfortable. To address these limitations, the researchers designed a wearable patch composed of a composite hydrogel electrode interface. The hydrogel material is flexible and easy to apply and remove, making it suitable for continuous monitoring while maintaining good signal quality.
The hydrogel sensor collects vibrations and muscle movement data, which are then processed by a machine learning algorithm. The collected data is sent to a custom-built cloud interface, where healthcare providers can remotely access and analyze it. The algorithm groups the collected data at different frequencies, such as swallowing, speaking, or respiration, into a streamlined output. This streamlined data is much more useful for healthcare providers to quickly assess and make judgments.
The adaptive capabilities and memory functions of the algorithm allow it to predict patient data with over 90% accuracy after only one minute of data collection and three hours of offline training. This predictive data can be used by clinicians to inform early diagnoses and assess the effectiveness of treatments.
The research team believes that adaptive machine learning can also account for individual differences in data among a large population. This would enable researchers to make inferences on the health of a large population based on individual datasets. In summary, the development of this wearable stretchy sensor and its accompanying machine learning platform will significantly improve remote health monitoring and treatment evaluation. It provides a faster and more accurate method for analyzing and predicting health data, leading to earlier diagnoses and more effective treatments.
