Individual versus Group Calibration of Machine Learning Models for Physical Activity Assessment Using Body-Worn Accelerometers

Accelerometer techniques for capturing human movement validated against direct observation: a scoping review

Objective.Accelerometers are devices commonly used to measure human physical activity and sedentary time. Accelerometer capabilities and analytical techniques have evolved rapidly, making it difficult for researchers to keep track of advances and best practices for data processing and analysis. The objective of this scoping review is to determine the existing methods for analyzing accelerometer data for capturing human movement which have been validated against the criterion measure of direct observation.Approach.This scoping review searched 14 academic and 5 grey databases. Two independent raters screened by title and abstract, then full text. Data were extracted using Microsoft Excel and checked by an independent reviewer.Mainresults.The search yielded 1039 papers and the final analysis included 115 papers. A total of 71 unique accelerometer models were used across a total of 4217 participants. While all studies underwent validation from direct observation, most direct observation occurred live (55%) or using recordings (42%). Analysis techniques included machine learning (ML) approaches (22%), the use of existing cut-points (18%), receiver operating characteristic curves to determine cut-points (14%), and other strategies including regressions and non-ML algorithms (8%).Significance.ML techniques are becoming more prevalent and are often used for activity identification. Cut-point methods are still frequently used. Activity intensity is the most assessed activity outcome; however, both the analyses and outcomes assessed vary by wear location. This scoping review provides a comprehensive overview of accelerometer analysis and validation techniques using direct observation and is a useful tool for researchers using accelerometers.

Floating Epoch Length Improves the Accuracy of Accelerometry-Based Estimation of Coincident Oxygen Consumption

Estimation of oxygen consumption (VO2) from accelerometer data is typically based on prediction equations developed in laboratory settings using steadily paced and controlled test activities. These equations may not capture the temporary changes in VO2 occurring in sporadic real-life physical activity. In this study, we introduced a novel floating epoch for accelerometer data analysis and hypothesized that an adaptive epoch length provides a more consistent estimation of VO2 in irregular activity conditions than a 6 s constant epoch. Two different activity tests were conducted: a progressive constant-speed test (CS) performed on a track and a 6 min back-and-forth walk test including accelerations and decelerations (AC/DC) performed as fast as possible. Twenty-nine adults performed the CS test, and sixty-one performed the AC/DC test. The data were collected using hip-worn accelerometers and a portable metabolic gas analyzer. General linear models were employed to create the prediction models for VO2 that were cross-validated using both data sets and epoch types as training and validation sets. The prediction equations based on the CS test or AC/DC test and 6 s epoch had excellent performance (R2 = 89%) for the CS test but poor performance for the AC/DC test (31%). Only the VO2 prediction equation based on the AC/DC test and the floating epoch had good performance (78%) for both tests. The overall accuracy of VO2 prediction is compromised with the constant length epoch, whereas the prediction model based on irregular acceleration data analyzed with a floating epoch provided consistent performance for both activities.