Optimal Head Impact Signal Processing and the Description and Perception of Head Impact Exposure in Female Adolescent Ice Hockey Players

TR Number

Date

2024-08-29

Journal Title

Journal ISSN

Volume Title

Publisher

Virginia Tech

Abstract

Concussion and repetitive head impact exposure are significant media, clinical, and research topics. Long-term symptomatic outcomes of sub-injurious head impact exposure have become a topic of concern for professional and youth contact sport athletes alike. Vulnerable populations such as females and pediatric athletes deserve special attention but are understudied. It is known that females tend to be diagnosed with concussions more frequently and incur them at lower peak head kinematic values than their male counterparts. Sex-specific symptom presentation has been demonstrated, but little is known about the effect of athlete perception of head impact exposure or skill level on symptom reporting.

Injury biomechanics research has begun to converge on using instrumented mouthguards to monitor head impact exposure in various sports. These devices offer six-degree-of-freedom head kinematic measurements and direct coupling to the measurement point of interest, the skull, by connecting to the upper dentition. However, published post-processing recommendations for these devices differ from one another and manufacturer practices. This lack of commonization makes cross-study comparison difficult. Additionally, the devices are plagued by relative motion artifacts that can leak into reported kinematic signals.

The research presented in this dissertation aimed to first address a lack of common post-processing methods available for instrumented mouthguard measurements made with three linear accelerometers and three gyroscopes. We developed an optimal combination of cutoff frequencies for filters applied to these instruments by quantifying the minimal error from a transformation function. We then applied those same methods to instrumented mouthguards, minimizing error based on sport-specific impact duration. Next, mouthguard decoupling artifacts were described in a laboratory study. Decoupling increased kinematic error relative to ground truth measurements from an instrumented headform. We used these data to develop a classification algorithm that found signal features related to mouthguard decoupling while recording an acceleration event in an instrumented mouthguard. We proceeded to salvage impacts with decoupling artifacts. When decoupling was identified, the primary head acceleration signal could be salvaged by wavelet deconstruction. We removed high-frequency content that was representative of decoupling artifacts.

We applied these optimized post-processing techniques to instrumented mouthguard data from a group of adolescent female ice hockey players. Their documented head impact exposure was correlated to symptom outcomes and ocular motor evaluation scores. Their sleep and menstrual cycle patterns were included as potential confounding factors. In this sample of athletes, sleep was more strongly associated with symptom presentation than head impact exposure. Ocular motor results showed a possible association with head acceleration exposure severity and menstrual cycle phase, but further study is warranted. Finally, a self-reported association between symptoms and head impact exposure appears to be individual- and skill-level specific, as we saw many variations between individuals of the same sex in what they called a "memorable" head impact.

Description

Keywords

Concussion, Instrumented Mouthguard, Post-Processing, Decoupling, Perception

Citation