Development of a wearable multimodal sensing ring for quantifying skin scratch intensity

Chronic itch can be caused by a variety of conditions, including eczema and psoriasis. Itch can lead to scratching behaviors, which can affect people's sleep, work productivity, mood, and overall health. Therefore, objective measurement of chronic itch is important to improve patient care. Currently, wearable devices have shown promise in scratching behavior, but most existing wearable devices cannot assess scratch intensity, which makes it difficult to fully understand the impact of itch on an individual.

According to MEMS Consulting, researchers from the Robocs Institu, Carnegie Mellon University and other institutions have recently developed a multimodal sensing ring that can be worn on the finger to measure scratch intensity and can distinguish different types and intensities of scratching behavior. Further development of the ring will help to measure scratching behavior more consistently and comprehensively, helping doctors and patients better understand and treat chronic itching. The relevant research results were published in the journal Communications Medicine under the title "A multimodal sensing ring for quantifation of scratch intensity".

In this work, the researchers proposed a system architecture for quantifying scratch intensity, including a multimodal wearable ring, a method to capture real data of scratch intensity from a pressure-sensitive tablet, and a supervised method for regressing scratch intensity in the power range of 0 ~ 600 mW. The multimodal ring consists of a three-axis accelerometer and a vibrating contact microphone, which can provide complementary acoustic and mechanical features for scratching behavior. These can be used to capture low-frequency finger and arm movements, as well as high-frequency vibrations propagating from the nail to the ring. For the processing of raw sensor data, the raw data is analyzed by extracting frequency features and using two machine learning algorithms to evaluate scratching behavior and scratch intensity in real time. To train the scratch intensity algorithm, the researchers used a method to automatically generate intensity from a touch-sensitive surface (i.e., a pressure-sensitive tablet) embedded with an array of pressure sensors.

Multimodal wearable ring and its raw data captured from scratching behavior

Real-time assessment of scratching behavior and scratch intensity using machine learning algorithms

The researchers evaluated the analytical capabilities of the multimodal algorithm by leaving one subject out (LOSO) cross validation (CV) on data from 20 healthy subjects, demonstrating that the proposed multimodal ring can accurately assess scratch intensity in the power range of 0 ~ 600 mW with a mean absolute error (MAE) of 49.71 mW. In addition, this multimodal sensing strategy has significant advantages in scratch detection, with the model using a combination of accelerometers and contact microphones achieving an accuracy of 89.98%, while the models using only accelerometers and only contact microphones have an accuracy of 86.24% and 79.98%, respectively.

Furthermore, compared to existing patient-reported scales that only capture itch symptoms at specific time points, the technology proposed in this study is expected to provide continuous monitoring of scratching behavior, thereby reducing the workload for doctors and patients.

Evaluating the analytical ability of multimodal algorithms on scratch data by leave-one-subject cross-validation

The researchers said that although the current research focuses on distinguishing scratching behavior from other daily behaviors, future research work can also include daily behaviors similar to scratching, such as kneading, picking and rubbing.

In summary, this research work proposes a wearable multimodal sensor ring that can not only be used for scratching detection, but also quantify scratching intensity, both of which are critical for tracking itch symptoms. In clinical trials, this wearable ring and multimodal algorithm system can provide researchers and clinicians with more in-depth treatment references by directly reflecting scratching behavior in real time. By providing direct and objective quantification of scratches, this multimodal ring will help improve patient treatment outcomes and aid in the discovery and validation of treatment methods.

Review editor: Liu Qing