Systematic analysis of different low-pass filter cut-off frequencies on lumbar spine kinematics data and the impact on the agreement between accelerometers and an optoelectronic system

Measurement of spinal sagittal motion remains of critical importance in the assessment and treatment of low back pain (Arshad et al., 2019). Currently, numerous methods exist to achieve this aim (Reddy et al., 2021, Serafino et al., 2021, Williams et al., 2013, Williams et al., 2014, Van Herp et al., 2000). However, biomechanics has seen a shift from traditional laboratory-based optoelectronic camera systems which are constrained usually to a specific environment, costly, and require line of sight, to body worn sensor systems such as accelerometers and/or inertial measurement units (Frey et al., 2020, Williams et al., 2014). Such devices are portable, cost effective, and can be worn under clothing making them attractive for clinical applications. Accelerometers in isolation can measure sagittal angle through the relationship to the vertical (i.e., gravity vector) (Frey et al., 2019, Alqhtani et al., 2016). However, raw data from accelerometers are subject to high frequency noise and filtering is a recommended step for such kinematic data (Derrick et al., 2020). A necessary step in the validation of accelerometers for the measurement of angles is to determine the levels of agreement with current gold standard methods. However, it is possible that this agreement is a function of the filtering parameters used on the data.

The decision pertaining to filtering parameters, especially cut-off frequency, often causes great discussion within and across research teams. Challenges include deciding the appropriate cut-off frequency, method to determine the appropriate cut-off frequency, and then ultimately whether to apply the same filter to a group of individuals or to take a more individualised approach to filtering. Such concern over ‘getting this right’ can be witnessed through criticism by journal reviewers or by conference delegates, where chosen filtering parameters are often challenged resulting in different parameters being used throughout the literature. Similar discussions over appropriate cut off frequency were held within our research group measuring sagittal range of motion of the spine and prompted this technical note. Many types of filtering are possible, but it is common in the literature to utilise a low pass Butterworth filter to remove the high frequency noise component of the signal, often associated with collecting data of human movement (Kristianslund et al., 2012). A recent review demonstrated cut off frequencies range between 1 and 15 Hz (Papi et al., 2017) with 4–6 Hz being commonplace in lumbar kinematics research (Bauer et al., 2015, Senington et al., 2020, Tulipani et al., 2018, Wong and Wong, 2008, Alqhtani et al., 2015).

The aim of this technical note is to (1) systematically determine the effect of applying different filter frequency cut-offs on the peak range of motion (ROM) during forward bending as measured by accelerometers and an optoelectronic (OE) system respectively, (2) explore the influence of filtering on the agreement between accelerometers and an OE system, and (3) determine the difference in peak ROM measurement between accelerometers and an OE system.

留言 (0)

沒有登入
gif