Classification of head impacts based on the spectral density of measurable kinematics. Xianghao Zhan 1 , Yiheng Li 2 , Yuzhe Liu 1 , Nicholas J. Cecchi 1 , Samuel J. Raymond 1 , Zhou Zhou 1 , Hossein Vahid Alizadeh 1 , Jesse Ruan 3 , Saeed Barbat 3 , Stephen Tiernan 4 , Olivier Gevaert 2 , Michael M. Zeineh 5 , Gerald A. Grant 6 , David B. Camarillo 1 Abstract Traumatic brain injury can be caused by head impacts, but many brain injury risk estimation models are less accurate across the variety of impacts that patients may undergo. We investigated the spectral characteristics of different head impact types with kinematics classification. Data was analyzed from 3,262 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), reaching a median accuracy of 96% over 1,000 random partitions of training and test sets. To test the classifier on data from different measurement devices, another 271 lab-reconstructed impacts were obtained from 5 other instrumented mouthguards with the classifier reaching over 96% accuracy. The most important features in the classification included both low-frequency and high-frequency features, both linear acceleration features and angular velocity features. Different head impact types had different distributions of spectral densities in low-frequency and high-frequency ranges (e.g., the spectral densities of MMA impacts were higher in high-frequency range than in the low-frequency range). Finally, with the classifier, type-specific, nearest-neighbor regression models were built for 95th percentile maximum principal strain, 95th percentile maximum principal strain in corpus callosum, and cumulative strain damage (15th percentile). This showed a generally higher R 2 -value than baseline models. The classifier enables a better understanding of the impact kinematics in different sports, and it can be applied to evaluate the quality of impact-simulation systems and on-field data augmentation. Key words: traumatic brain injury, head impacts, classification, impact kinematics Introduction Traumatic brain injury (TBI) is a growing public health hazard with high mortality and morbidity, as well as a socio-economic issue causing enormous diagnosis and treatment expenses 1 . This is particularly urgent for mild TBI (mTBI), given that mTBI is notoriously underreported, difficult to diagnose, and pose a potential predisposing factor to long-term neurodegenerative processes 2 . TBI/mTBI can be caused by various types of head impacts such as accidental falls, bike accidents, car crashes, American football impacts, mixed martial arts (MMA) impacts, water polo and car crashes 3-5 . Considering the consequences and prevalence of TBI/mTBI, various biomechanical studies have focused on the estimation of brain injury risk 6-10 . Recent study 11 found that different head impact types tend to have variable biomechanical characteristics, and the impact types should not be ignored when estimating the risk of TBI/mTBI. However, the brain injury criteria (BIC) were developed based on certain types of head impacts 4,12 , and should not be used across head impact types as the different kinematic features these BIC use can weigh differently across impact types 13,14 . To better develop risk evaluation models adaptable to various head impact types for detection and monitoring of TBI/mTBI, it is worthwhile to investigate the difference in the kinematics of various types of head impacts. Sports-specific monitoring and protection strategies can be developed if we understand the difference among types of head impacts. To study the difference across head impact types, we used the kinematics of 3,262 head impacts from head model simulations, American football, MMA, automobile crashworthiness tests and car racing. We extracted the spectral density of linear acceleration and angular velocity, classified these impacts, and then analyzed the most important features for classification. Finally, we used the classification model to