Comparison of methods of derivation of the yank-time signal from the vertical ground reaction force–time signal for identification of movement-related events

• Published in: Journal of biomechanics 2021-01, Vol.115, p.110048-110048, Article 110048
• Main Authors: Sahrom, Sofyan, Wilkie, Jodie Cochrane, Nosaka, Kazunori, Blazevich, Anthony J.
• Format: Article
• Language: English
• Subjects: Butterworth; Butterworth filters; Countermovement jump; Data smoothing; Derivation; Derivative; Differentiation; Filter; Force platform; Inflection points; Kinematics; Muscle contraction; Muscles; Savitsky-Golay; Signal processing; Signal to noise ratio; Standard deviation; Three dimensional motion; Vertical forces
• ISSN: 0021-9290; 1873-2380
• DOI: 10.1016/j.jbiomech.2020.110048

Temporal changes in ground reaction force magnitudes reflect movement strategy, and thus underlying muscle activation patterns, during movement tasks. Speculatively, these changes may be observed more readily when the force–time signal is differentiated, yielding the yank-time signal. However, the differentiation process, including the signal filtering used before or after differentiation, can significantly affect the signal-to-noise ratio (SNR) and likelihood of meaningful inference. The aim of the present study was to compare three methods of deriving the yank-time signal: Method 1 derived the yank-time signal using 2nd-order central differentiation subsequent to application of a 4th-order Butterworth filter; Method 2 included the same process as Method 1 but additionally filtered the yank-time data with a Savitzky-Golay smoothing filter; and Method 3 directly and simultaneously derived and smoothed the yank-time signal using a Savitzky-Golay digital differentiation filter. The current analyses revealed Method 2 had the best SNR, followed by Methods 3 and 1, but caused a small loss of signal amplitude. With regards to timing of inflection points in the yank-time data, no significant difference was observed. Therefore, Method 3 led to the best derivation of the yank-time signal due to its efficiency and preservation of signal characteristics and good SNR. Also, a strong association between the first maximum point of the yank-time signal and the start of the downward movement of the body’s centre of mass during a countermovement jump, as identified by 3-D motion analysis, was observed. Thus, subtle events (e.g. start of downward movement) can be easily observed in the yank-time signal.