Frequency content and filtering of head sensor kinematics: A method to enable field-based inter-study comparisons

Gregory Tierney,Steve Rowson,Ryan Gellner,Sadaf Iqbal,Pardis Biglarbeigi,James Tooby,James Woodward,Amir Farokh Payam
DOI: https://doi.org/10.1136/bmjsem-2023-001758
2024-09-23
Abstract:Wearable head sensor systems use different kinematic signal processing approaches which limits field-based inter-study comparisons, especially when artefacts are present in the signal. The aim of this study is to assess the frequency content and characteristics of head kinematic signals from head impact reconstruction laboratory and field-based environments to develop an artefact attenuation filtering method (artefact attenuation method). Laboratory impacts (n=72) on a test-dummy headform ranging from 25-150 g were conducted and 126 elite-level rugby union players were equipped with instrumented mouthguards (iMG) for up to four matches. Power spectral density (PSD) characteristics of the laboratory impacts and on-field HAE (n=5694) such as the 95th percentile cumulative sum PSD frequency were utilised to develop the artefact attenuation method. The artefact attenuation method was compared to two other common filtering approaches (Fourth order (2x2 pole), zero-lag Butterworth filter with 200 Hz (-6 dB) cut-off frequency (Butterworth-200Hz) and CFC180 filter) through signal-to-noise ratio (SNR) and mixed linear effects models for laboratory and on-field events, respectively. The artefact attenuation method produced an overall higher SNR than the Butterworth-200Hz and CFC180 filter and on-field peak linear acceleration (PLA) and peak angular acceleration (PAA) values within the magnitude range tested in the laboratory. Median PLA and PAA were higher for the CFC180 filter than the Butterworth-200Hz (p<0.01) and artefact attenuation method (p<0.01), reporting values as high as 294 g and 31.2 krad/s2. The artefact attenuation method can be applied to all commercially available iMG kinematic signals with adequate sample rates to enable field-based inter-study comparisons.
Medical Physics
What problem does this paper attempt to address?