Objective profiling of varied human motion based on normative assessment of magnetometer time series data

Abstract

By using multi-dimensional analysis of similarity measures between participants rather than direct parameterisation of the physiological data, complex and varied patterns of physical motion can be quantified, allowing objective and robust profiling of relative function across participant groups.

A Narrative Review of Motor Competence in Children and Adolescents: What We Know and What We Need to Find Out

“…Pervasive technologies, such as accelerometers, inertial measurement units and magnetometers, have been used with reasonable success. In the study of Barnes et al [94], a magnetometer (which measures the direction, strength, or relative change of a magnetic field at a given location) was worn during a MC assessment. Using robust machine learning, data was subsequently processed, yielding visualizations of the relative performance in three-dimensions and a relative distance between children within the multidimensional scaling that could be used to create an automated sensor-based rank scoring, resulting in good agreement between observer and sensor scores.…”

New Perspectives through Emerging Technologies

“…In particular, Table 1 outlines solutions proposed for quantitative activity monitoring, considering both product and process-oriented approaches, while Table 2 refers to quantitative motion analysis, which is a detailed process-oriented description of specific motor tasks. Magnetometer (Barnes et al, 2018); Force sensitive resistor (Xiao and Menon, 2014) Dynamic time warping (Barnes et al, 2018); contour mapping (Barnes et al, 2018); Extreme learning machine classifier (Xiao and Menon, 2014) Daily living Visual; time parameters; space parameters Electro-oculography (Bulling et al, 2011); gyroscope (Leutheuser et al, 2013); Radio frequency identification (RFID) (Spinney et al, 2015) Support vector machine (Bulling et al, 2011;Leutheuser et al, 2013); k-Nearest Neighbor classifier (Leutheuser et al, 2013); classification and regression tree (Leutheuser et al, 2013); linear correction (Spinney et al, 2015) PA intensities Visual; tangible; frequency parameters…”

Physical activity, motor competence and movement and gait quality: A principal component analysis

“…Whilst Barnes et al (Barnes et al, 2018) show that comparison of an automated, sensor-based method to the standard approach has a strong correlation to subjective human-assessed scores.…”

“…Recently there have been developments in technological and analytical capability, permitting the quantification of complex human movement behaviours (Clark, Barnes, Stratton, et al, 2016) which have as yet untapped potential to be applied to the assessment of motor competence. Pervasive technologies, such as accelerometers, inertial measurement units and magnetometers have been used, albeit in only a small number of studies, with reasonable success to automatically assess and score motor competence (Barnes, Clark, Rees, Stratton, & Summers, 2018;Bisi, Panebianco, Polman, & Stagni, 2017). For example, Barnes et al (2018) demonstrated good agreement between observer and magnetometry derived motor competency scores, where raw tri-axial magnetometer traces underwent pattern recognition and were systematically compared against human-assessed scores, with correlation coefficients of the overall score in the range of 0.62-0.71 for different cohorts.…”

“…Pervasive technologies, such as accelerometers, inertial measurement units and magnetometers have been used, albeit in only a small number of studies, with reasonable success to automatically assess and score motor competence (Barnes, Clark, Rees, Stratton, & Summers, 2018;Bisi, Panebianco, Polman, & Stagni, 2017). For example, Barnes et al (2018) demonstrated good agreement between observer and magnetometry derived motor competency scores, where raw tri-axial magnetometer traces underwent pattern recognition and were systematically compared against human-assessed scores, with correlation coefficients of the overall score in the range of 0.62-0.71 for different cohorts. Whilst Bisi et al (2017), with the application of inertial measurement units, which consisted of an in-built, tri-axial, magnetometer, gyroscope and accelerometer, showed that automatic assessment, compared to observer assessment, yielded an agreement of 87% on average across an entire cohort for each skill.…”