Multimodal Joint Health Sensing
We have developed a multimodal sensing systems capable of capturing signals related to joint health such as electrical bioimpedance (EBI) and joint acoustic emissions with convenient, wearable form factor to enable longitudinal and mid-activity measurements. Such a system can be deployed in clinical or at-home settings, allowing for potential tracking of rehabilitation after a musculoskeletal injury or the prediction of future joint injuries.
Joint Acoustic Emissions
The knee is one of the most commonly injured body parts and is also susceptible to numerous debilitating chronic joint diseases such as osteoarthritis and rheumatoid arthritis. To assess knee health, we have investigated knee acoustic emissions produced by knee movement, which allows us to study the joint’s structural health. With degradation or acute injury, we found these acoustic emissions to be a salient biomarker enabling the differentiation between healthy and unhealthy joints.
With the progression of wearable devices and at home data collections, the introduction of artifacts becomes more likely due to the fluctuating environment. To address this, we have developed tools for signal quality assessment for joint acoustic emissions. The flexion/extension cycles are described using a graph based spectral embedding which provides information about the underlying community structure of the signal. We found that this structure was similar across all clean cycles, regardless of health status, and largely impacted by artifacts, the removal of which improved knee health classification.
Wearable Biomechanical Sensing
To noninvasively assess biomechanical parameters such as vertical ground reaction forces during activity, our lab has developed wearable bioimpedance systems capable of capturing time-series dynamics and longitudinal changes in the electrical bioimpedance (EBI) signal related to muscle force, fatigue, and other relevant biomechanical properties.
