Chi-Ming Huang

Chi-Ming Huang
Associate Professor
Biological and Chemical Sciences

Contact Info
529B SH - Office 506 SH- Lab
MEMS sensors fortified with artificial intelligence and machine learning


The outcome of head-impact depends upon not only the impact force but also the biomechanical properties of his or her head-and-neck. MicroElectroMechanical Systems (MEMS) head-impact sensors can only measure the physical parameters of external impact forces. Our hypothesis is that MEMS sensors fortified with artificial intelligence and machine learning can incorporate individualized human factors into consideration and help to define concussive threshold that is personalized as in precision medicine. In laboratory tests, the prototype smart sensor we built can machine-learn the biomechanical properties of the head-and-neck of the user without being programmed with that information. It then measures the magnitude of the impact against a personalized threshold of the user, in real time. These capabilities allow our smart sensor to accurately determine the potential for concussive injuries for a given individual. Our lab continues to work to improve the technology for concussion diagnosis.

At present, there are few effective clinical treatments for acute concussions outside of rest, vestibular therapy and gradual exercise. Concussions can progress into chronic traumatic encephalopathy(CTE). Concussion prevention is therefore particularly important. Our laboratory has developed a concussion avoidance training (CAT) to significantly reduce concussion risks. The principles of CAT are grounded in the neuroscience of motor learning and a Pavlov-like conditioned response (CR). The acquisition of the CR is facilitated by virtual reality technology. The effectiveness of this CR in reducing concussions is based on biomechanics of the head-and-neck and the role of neck stiffness in response to external impact forces. Simply put, a dynamic increase in neck stiffness helps to resist impact force and reduce head angular accelerations, thereby decreasing concussion risk. Our hypothesis is that the dynamic increase in neck stiffness should reduce concussive risks significantly. Our smart sensor is fully integrated into the training in order to quantitatively monitor the increase in neck stiffness. We can therefore validate the effectiveness of the acquired CR in reducing concussion risks. Our lab continues to work to improve the CAT for concussion prevention.