Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics
Background: Traumatic brain injury can be caused by head impacts, but many brain injury risk estimation models are not equally accurate across the variety of impacts that patients may undergo, and the characteristics of different types of impacts are not well studied. We investigated the spectral ch...
Main Authors: | , , , , , , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2023-09-01
|
Series: | Journal of Sport and Health Science |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S2095254623000236 |
_version_ | 1797738985152315392 |
---|---|
author | Xianghao Zhan Yiheng Li Yuzhe Liu Nicholas J. Cecchi Samuel J. Raymond Zhou Zhou Hossein Vahid Alizadeh Jesse Ruan Saeed Barbat Stephen Tiernan Olivier Gevaert Michael M. Zeineh Gerald A. Grant David B. Camarillo |
author_facet | Xianghao Zhan Yiheng Li Yuzhe Liu Nicholas J. Cecchi Samuel J. Raymond Zhou Zhou Hossein Vahid Alizadeh Jesse Ruan Saeed Barbat Stephen Tiernan Olivier Gevaert Michael M. Zeineh Gerald A. Grant David B. Camarillo |
author_sort | Xianghao Zhan |
collection | DOAJ |
description | Background: Traumatic brain injury can be caused by head impacts, but many brain injury risk estimation models are not equally accurate across the variety of impacts that patients may undergo, and the characteristics of different types of impacts are not well studied. We investigated the spectral characteristics of different head impact types with kinematics classification. Methods: Data were analyzed from 3262 head impacts from lab reconstruction, American football, mixed martial arts, and publicly available car crash data. A random forest classifier with spectral densities of linear acceleration and angular velocity was built to classify head impact types (e.g., football, car crash, mixed martial arts). To test the classifier robustness, another 271 lab-reconstructed impacts were obtained from 5 other instrumented mouthguards. Finally, with the classifier, type-specific, nearest-neighbor regression models were built for brain strain. Results: The classifier reached a median accuracy of 96% over 1000 random partitions of training and test sets. The most important features in the classification included both low- and high-frequency features, both linear acceleration features and angular velocity features. Different head impact types had different distributions of spectral densities in low- and high-frequency ranges (e.g., the spectral densities of mixed martial arts impacts were higher in the high-frequency range than in the low-frequency range). The type-specific regression showed a generally higher R2 value than baseline models without classification. Conclusion: The machine-learning-based classifier enables a better understanding of the impact kinematics spectral density in different sports, and it can be applied to evaluate the quality of impact-simulation systems and on-field data augmentation. |
first_indexed | 2024-03-12T13:50:47Z |
format | Article |
id | doaj.art-334dac0a60b244f79dfcd59cee5d0eca |
institution | Directory Open Access Journal |
issn | 2095-2546 |
language | English |
last_indexed | 2024-03-12T13:50:47Z |
publishDate | 2023-09-01 |
publisher | Elsevier |
record_format | Article |
series | Journal of Sport and Health Science |
spelling | doaj.art-334dac0a60b244f79dfcd59cee5d0eca2023-08-23T04:32:58ZengElsevierJournal of Sport and Health Science2095-25462023-09-01125619629Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematicsXianghao Zhan0Yiheng Li1Yuzhe Liu2Nicholas J. Cecchi3Samuel J. Raymond4Zhou Zhou5Hossein Vahid Alizadeh6Jesse Ruan7Saeed Barbat8Stephen Tiernan9Olivier Gevaert10Michael M. Zeineh11Gerald A. Grant12David B. Camarillo13Department of Bioengineering, Stanford University, Stanford, CA 94305, USADepartment of Biomedical Data Science, Stanford University, Stanford, CA 94305, USADepartment of Bioengineering, Stanford University, Stanford, CA 94305, USA; Corresponding author.Department of Bioengineering, Stanford University, Stanford, CA 94305, USADepartment of Bioengineering, Stanford University, Stanford, CA 94305, USADepartment of Bioengineering, Stanford University, Stanford, CA 94305, USADepartment of Bioengineering, Stanford University, Stanford, CA 94305, USAFord Motor Company, 3001 Miller Rd, Dearborn, MI 48120, USAFord Motor Company, 3001 Miller Rd, Dearborn, MI 48120, USATechnological University Dublin, Dublin, D07 EWV4, IrelandDepartment of Biomedical Data Science, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Neurosurgery, Stanford University, Stanford, CA 94305, USADepartment of Bioengineering, Stanford University, Stanford, CA 94305, USABackground: Traumatic brain injury can be caused by head impacts, but many brain injury risk estimation models are not equally accurate across the variety of impacts that patients may undergo, and the characteristics of different types of impacts are not well studied. We investigated the spectral characteristics of different head impact types with kinematics classification. Methods: Data were analyzed from 3262 head impacts from lab reconstruction, American football, mixed martial arts, and publicly available car crash data. A random forest classifier with spectral densities of linear acceleration and angular velocity was built to classify head impact types (e.g., football, car crash, mixed martial arts). To test the classifier robustness, another 271 lab-reconstructed impacts were obtained from 5 other instrumented mouthguards. Finally, with the classifier, type-specific, nearest-neighbor regression models were built for brain strain. Results: The classifier reached a median accuracy of 96% over 1000 random partitions of training and test sets. The most important features in the classification included both low- and high-frequency features, both linear acceleration features and angular velocity features. Different head impact types had different distributions of spectral densities in low- and high-frequency ranges (e.g., the spectral densities of mixed martial arts impacts were higher in the high-frequency range than in the low-frequency range). The type-specific regression showed a generally higher R2 value than baseline models without classification. Conclusion: The machine-learning-based classifier enables a better understanding of the impact kinematics spectral density in different sports, and it can be applied to evaluate the quality of impact-simulation systems and on-field data augmentation.http://www.sciencedirect.com/science/article/pii/S2095254623000236ClassificationContact sportsHead impactsImpact kinematicsTraumatic brain injury |
spellingShingle | Xianghao Zhan Yiheng Li Yuzhe Liu Nicholas J. Cecchi Samuel J. Raymond Zhou Zhou Hossein Vahid Alizadeh Jesse Ruan Saeed Barbat Stephen Tiernan Olivier Gevaert Michael M. Zeineh Gerald A. Grant David B. Camarillo Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics Journal of Sport and Health Science Classification Contact sports Head impacts Impact kinematics Traumatic brain injury |
title | Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics |
title_full | Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics |
title_fullStr | Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics |
title_full_unstemmed | Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics |
title_short | Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics |
title_sort | machine learning based head impact subtyping based on the spectral densities of the measurable head kinematics |
topic | Classification Contact sports Head impacts Impact kinematics Traumatic brain injury |
url | http://www.sciencedirect.com/science/article/pii/S2095254623000236 |
work_keys_str_mv | AT xianghaozhan machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT yihengli machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT yuzheliu machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT nicholasjcecchi machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT samueljraymond machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT zhouzhou machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT hosseinvahidalizadeh machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT jesseruan machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT saeedbarbat machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT stephentiernan machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT oliviergevaert machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT michaelmzeineh machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT geraldagrant machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics AT davidbcamarillo machinelearningbasedheadimpactsubtypingbasedonthespectraldensitiesofthemeasurableheadkinematics |