Development and Validation of a "Gold Standard" in vivo Head Impact Sensor

The original project, referred to as Project 1, seeks to determine how the approach to affixing an in vivo impact sensor to the head influences its measurement accuracy.  It has been shown that existing in vivo head impact sensors cannot reliably distinguish concussion-causing head impacts from non-concussive impacts based on their head kinematics (acceleration) measurements. We hypothesized that one reason for this is that the non-rigid fixation of existing in vivo sensors to the head leads to measurement error.  Therefore, a novel sensor design that is rigidly fixed to the upper teeth via an orthodontic appliance should provide more accurate head kinematics measurements that may be more predictive of concussion. In this pilot study, we used a free-drop testing paradigm in helmeted and un-helmeted cadaveric head-neck specimens to compare head kinematics (acceleration) measurements obtained from a sensor rigidly fixed to the upper teeth using our novel orthodontic approach to measurements obtained from an intracranial reference sensor and multiple commercially-available sensors mounted in/on a football helmet, skin-patch, headband, or protective athletic mouth guard. Project 1 is dually-supported by this UM IC Pilot Grant and an M-Cubed award to Drs. Eckner, Ashton-Miller, and Conley (Co-PI’s). This project represents a significant step toward determining the influence of sensor fixation method on the measurement accuracy of wearable head impact sensors. Our observation that a novel approach involving rigid sensor fixation to an orthodontic appliance out performed existing commercially-available sensors raises concern about the value of data obtained from existing wearable head impact sensors affixed to the head by integration into a piece of protective athletic equipment or otherwise non-rigidly affixed to the head. This project moves the field one step closer to the ability to accurately identify concussion-causing head impacts vs. non-concussive impacts. 

The second project, referred to as Project 2, seeks to increase our understanding of concussion’s effect on the autonomic nervous system (ANS) and sleep. There is increasing recognition that brain injury has measurable effects on autonomic functions such as beat-to-beat heart rate variability (HRV), the galvanic skin (sweat) response, as well as sleep quality and duration. This remains an emerging area, with more known about the autonomic effects of moderate-severe brain injury relative to mild traumatic brain injury (mTBI)/concussion, and little known about ANS function in the post-concussion syndrome (PCS). In this pilot study, we used the wrist-worn Empatica E4 device to measure HRV, galvanic skin response, and sleep/activity measures in patients with acute concussion, post-concussion syndrome, and matched controls. Project 2 is dually-supported by this UM IC Pilot Grant and a MICHR Seed Pilot Grant to Dr. Stork (PI). This pilot project demonstrates the feasibility of our group using the wrist-worn Empatica E4 device in mTBI and PCS populations to study the effect of concussion on autonomic function. Upon completion of data analysis, we will also begin characterizing the effects of acute concussion and PCS on measures of autonomic function. This pilot project has been used to support a larger internal grant to move the field one step closer to identifying an objective concussion biomarker capable of diagnosing concusion and predicting those at high risk for developing PCS.

Principal Investigator: 
James T. Eckner
Assistant Professor
Physical Medicine and Rehabilitation, University of Michigan
2015 - 2016
Abstract Text: 
Project 1: AIM 1 was to develop a working prototype head impact sensor that can be rigidly affixed to the upper teeth. This was accomplished. The prototype sensor/fixation system was used during our lab testing, but will require additional device modification for field use in human subjects. AIM 2 was to perform initial validation of the prototype head impact sensor’s measurements against a reference sensor affixed intracranially at the center-of-mass of a human cadaver head during laboratory-based free drop tests. This has been accomplished. Our analyses to-date demonstrate that the novel test impact sensor’s peak linear and angular acceleration measurements were more strongly correlated with the intracranial reference sensor then the same measurements taken by any of the commercial impact sensors. Our results suggest that measurement accuracy differs between head impact sensors utilizing different methods of sensor-to-head fixation, with greater accuracy achieved by the novel rigid orthodontic fixation than any of the existing commercial sensors tested (helmet-based, adhesive patches, protective mouth guard, headband). This finding supports the need for additional work to develop an orthodontically-fixed impact sensor suitable for use in living human subjects outside of the lab environment that can be used in future field-based research. Improving measurement accuracy represents a critical step toward the ability to accurately distinguish concussion-causing head impacts from non-concussive head impacts in athletes in real time on the field of play. Project 2: AIM 1 was to examine the feasibility and acceptability of the E4 Wristband in those with concussion, PCS, and matched healthy controls. No significant feasibility or acceptability issues with the E4 devices were raised by our subjects. AIM 2 is to perform exploratory comparisons of E4 Wristband generated physiological data between study groups: concussion, PCS, and matched controls. Data collection is finished. We anticipate proceeding with analysis of the physiological data this summer in collaboration with collaborators in Computer Science Engineering. If our results demonstrate measurable autonomic differences between patients with acute concussion vs. PCS vs. healthy matched controls, this finding will support the need for additional work to determine which acute ANS findings predict development of future PCS. A reliable biomarker capable of both distinguishing those with acute concussion from non-concussed individuals, as well as predicting future development of PCS would have tremendous clinical value.
PI Title: 
Assistant Professor, Physical Medicine and Rehabilitation, University of Michigan