A smart neckband for tracking dietary intake

This technology is particularly beneficial for managing conditions such as diabetes and obesity or optimising fitness regimes. However, distinguishing between eating and drinking motions and other similar activities like talking or walking poses a significant challenge for wearable devices. To address this, Chi Hwan Lee and his team have designed a machine-learning equipped neckband capable of differentiating between various bodily movements, speech, and the intake of food and fluids.

The sensor module, positioned on the thyrohyoid muscle, features a 45-degree pre-curved design and soft, waterproof encapsulation. Credit: Park et al

The neckband is equipped with a sensor module that includes a surface electromyography sensor, a three-axis accelerometer, and a microphone. These components work together to detect muscle activation in the thyrohyoid muscle in the neck, body movements, and sounds. In a trial involving six volunteers, the embedded machine-learning algorithm accurately identified whether movements related to eating or drinking with an impressive accuracy of about 96% for individual actions and 89% for simultaneous activities.

Constructed from a flexible, twistable, and breathable mesh textile, the neckband incorporates 47 active and passive elements and can operate for over 18 hours on a single battery charge. The developers suggest that this neckband could integrate with a continuous glucose meter and insulin pump in a closed-loop system, enabling diabetic patients to accurately calculate insulin doses based on mealtimes. Additionally, it holds potential benefits for athletes, and anyone interested in enhancing their health and wellness.