Background: Application of machine learning for classifying human behavior is increasingly common as access to raw accelerometer data improves. The aims of this scoping review are (1) to examine if machine-learning techniques can accurately identify human activity behaviors from raw accelerometer data and (2) to summarize the practical implications of these machine-learning techniques for future work. Methods: Keyword searches were performed in Scopus, Web of Science, and EBSCO databases in 2018. Studies that applied supervised machine-learning techniques to raw accelerometer data and estimated components of physical activity were included. Information on study characteristics, machine-learning techniques, and key study findings were extracted from included studies. Results: Of the 53 studies included in the review, 75% were published in the last 5 years. Most studies predicted postures and activity type, rather than intensity, and were conducted in controlled environments using 1 or 2 devices. The most common models were support vector machine, random forest, and artificial neural network. Overall, classification accuracy ranged from 62% to 99.8%, although nearly 80% of studies achieved an overall accuracy above 85%. Conclusions: Machine-learning algorithms demonstrate good accuracy when predicting physical activity components; however, their application to free-living settings is currently uncertain.
Anantha Narayanan, Farzanah Desai, Tom Stewart, Scott Duncan and Lisa Mackay
Rachel G. Curtis, Dorothea Dumuid, Timothy Olds, Ronald Plotnikoff, Corneel Vandelanotte, Jillian Ryan, Sarah Edney and Carol Maher
Background: Substantial evidence links activity domains with health and well-being; however, research has typically examined time-use behaviors independently, rather than considering daily activity as a 24-hour time-use composition. This study used compositional data analysis to estimate the difference in physical and mental well-being associated with reallocating time between behaviors. Methods: Participants (n = 430; 74% female; 41  y) wore an accelerometer for 7 days and reported their body mass index; health-related quality of life (QoL); and symptoms of depression, anxiety, and stress. Regression models determined whether time-use composition, comprising sleep, sedentary behavior, light physical activity (LPA), and moderate to vigorous physical activity (MVPA), was associated with well-being. Compositional isotemporal substitution models estimated the difference in well-being associated with reallocating time between behaviors. Results: Time-use composition was associated with body mass index and physical health-related QoL. Reallocating time to MVPA from sleep, sedentary behavior, and LPA showed favorable associations with body mass index and physical health-related QoL, whereas reallocations from MVPA to other behaviors showed unfavorable associations. Reallocations from LPA to sedentary behavior were associated with better physical health–related QoL and vice versa. Conclusion: Results reinforce the importance of MVPA for physical health but do not suggest that replacing sedentary behavior with LPA is beneficial for health and well-being.
E. Kipling Webster, Leah E. Robinson and Danielle D. Wadsworth
Background: Activity breaks are an established way physical activity may be incorporated into the preschool day. The purpose of this study was to examine what factors influenced moderate to vigorous physical activity (MVPA) during a teacher-implemented classroom-based activity break (CBAB) in a Head Start population. Methods: Ten-minute CBAB was conducted over 2 days in a quasi-experimental design; 99 preschoolers (mean age 3.80 [0.65] y; 49.5% male) from a convenience sample participated. Accelerometers measured MVPA, fundamental motor skill competency was assessed using the Test of Gross Motor Development—second edition, and weight classification status used body mass index percentiles. Results: A significant, moderate regression was found (r = .328, P = .001) between fundamental motor skill and MVPA. There was no significant correlation between body mass index percentile and MVPA during the CBAB. In addition, the locomotor subscale was the best predictor for MVPA for children during the CBAB (r = .32, β = 0.82, P < .001). Conclusions: CBAB equally elicited MVPA for normal and overweight preschoolers. Fundamental motor skill competency was associated with MVPA during the CBAB; in particular, locomotor skills were the best predictor for physical activity. Structured activity opportunities that focus on locomotor skills may be a useful integration to prompt more MVPA in a preschool-age population.
Paul J. Collings, Diane Farrar, Joanna Gibson, Jane West, Sally E. Barber and John Wright
Background: Physical activity performed while pregnant is beneficially associated with maternal cardiovascular health. It is unknown if benefits extend to neonatal cardiovascular health. This study investigated associations of maternal physical activity with neonatal cord blood lipid and lipoprotein concentrations. Methods: Cord blood lipids were measured at birth in a pseudorandomly selected subgroup of Born in Bradford birth cohort participants (N = 1634). Pregnant women were grouped into 4 activity categories (inactive/somewhat active/moderately active/active) based on their self-reported physical activity at 26- to 28-weeks gestation. Regression was used to calculate adjusted mean differences in neonatal cord blood lipid concentrations among the 4 groups of physical activity. Results: Maternal physical activity was associated with higher neonatal cord blood high-density lipoprotein cholesterol. Cord blood high-density lipoprotein cholesterol was higher in neonates of women who were somewhat and moderately active compared with neonates of women who were inactive. There were no associations of pregnancy physical activity with triglycerides, low-density lipoprotein, very low-density lipoprotein cholesterol, or adiponectin levels. Conclusions: Maternal physical activity is favorably associated with neonatal cord blood high-density lipoprotein cholesterol levels. This novel beneficial finding highlights the potential for physical activity in pregnancy to aid the early prevention of cardiovascular disease.
Chia-Yuan Yu and Biyuan Wang
Background: This study explored the percentage change of walking to/from public transit to work from 2009 to 2017 in general and for specific sociodemographic characteristics. Furthermore, this study also examined the sociodemographic characteristics of those who walked to/from transit to work and those who walked 30 minutes or more per day to/from transit to work and compared the difference between 2009 and 2017. Methods: 2009 and 2017 National Household Travel Survey were used. This study used weighted logistic regressions to explore the sociodemographic characteristics of those who walked to/from transit to work and those who walked 30 minutes or more per day to/from transit to work in both 2009 and 2017. Results: The percentage of trips achieving the recommended level of physical activity (30 min or more per day) by walking to/from transit work solely has a slightly increase from 9 in 2009 to 9.5 in 2017. However, the weighted percentages of walking to/from transit to work decreased for low-education, low-income, and minority populations. High population density areas were related to more transit-related walking trips to work in both 2009 and 2017. Conclusions: Policymakers in terms of transit location and service should consider low-education, low-income, and minority populations to address potential equity issues.
Dimitrios Poulimeneas, Maria I. Maraki, Eleni Karfopoulou, Yannis Koutras, Stavrie Chrysostomou, Costas A. Anastasiou, Stavros A. Kavouras and Mary Yannakoulia
Background: Although plenty of evidence indicates that weight loss maintainers are highly physically active, studies focusing on the sex-specific differences in activity levels between maintainers and regainers are scarce. The authors aimed to investigate sex-specific differences in activity patterns in a cohort of Mediterranean maintainers and regainers. Methods: Sample includes 756 participants of the MedWeight registry (60.5% women), aged 18–65 years, who lost ≥10% of their initial weight, and either maintained their loss for ≥12 months or regained it. Participants completed a series of questionnaires, including demographics and weight history. Activity levels were evaluated with the International Physical Activity Questionnaire-short version. Results: Maintainers of both sexes were, in total, more active than their same-sex regainers. When specific activities were considered, women maintainers spent more time walking than regainers (P adjusted = .02), whereas men maintainers spent more time in vigorous activities (P adjusted = .001) and walking than regainers (P adjusted = .001). Modest increments in activity of sex-relevant intensity were associated with increased odds for maintenance. Conclusions: Maintainers attained a more active lifestyle than their same-sex regainers, involving more walking for both sexes and more vigorous activities for men. The detected differences, according to activity intensity, support that activity patterns associated with successful weight loss are distinguishable between sexes.
Kelsie M. Full, Eileen Johnson, Michelle Takemoto, Sheri J. Hartman, Jacqueline Kerr, Loki Natarajan, Ruth E. Patterson and Dorothy D. Sears
Background: For breast cancer survivors, moderate to vigorous physical activity (MVPA) is associated with improved survival. Less is known about the interrelationships of daytime activities (sedentary behavior [SB], light-intensity physical activity, and MVPA) and associations with survivors’ health outcomes. This study will use isotemporal substitution to explore reallocations of time spent in daytime activities and associations with cancer recurrence biomarkers. Methods: Breast cancer survivors (N = 333; mean age 63 y) wore accelerometers and provided fasting blood samples. Linear regression models estimated the associations between daytime activities and cancer recurrence biomarkers. Isotemporal substitution models estimated cross-sectional associations with biomarkers when time was reallocated from of one activity to another. Models were adjusted for wear time, demographics, lifestyle factors, and medical conditions. Results: MVPA was significantly associated with lower insulin, C-reactive protein, homeostatic model assessment of insulin resistance, and glucose, and higher sex hormone-binding globulin (all P < .05). Light-intensity physical activity and SB were associated with insulin and homeostatic model assessment of insulin resistance (both P < .05). Reallocating 18 minutes of SB to MVPA resulted in significant beneficial associations with insulin (−9.3%), homeostatic model assessment of insulin resistance (−10.8%), glucose (−1.7%), and sex hormone-binding globulin (7.7%). There were no significant associations when 79 minutes of SB were shifted to light-intensity physical activity. Conclusions: Results illuminate the possible benefits for breast cancer survivors of replacing time spent in SB with MVPA.
Tobias Lundgren, Gustaf Reinebo, Markus Näslund and Thomas Parling
Despite the growing popularity of mindfulness and acceptance-based performance enhancement methods in applied sport psychology, evidence for their efficacy is scarce. The purpose of the current study is to test the feasibility and effect of a psychological training program based on Acceptance and Commitment Training (ACT) developed for ice hockey players. A controlled group feasibility designed study was conducted and included 21 elite male ice hockey players. The ACT program consisted of four, once a week, sessions with homework assignments between sessions. The results showed significant increase in psychological flexibility for the players in the training group. The outcome was positive for all feasibility measures. Participants found the psychological training program important to them as ice hockey players and helpful in their ice hockey development. Desirably, future studies should include objective performance data as outcome measure to foster more valid evidence for performance enhancement methods in applied sport psychology.
Matthieu Dagenais, Nancy M. Salbach, Dina Brooks and Kelly K. O’Brien
Purpose: To assess the criterion and construct validity of the Fitbit Zip® to measure physical activity among adults living with HIV. Methods: Participants were video recorded completing 2 walk tests while wearing the Fitbit Zip® and completed 3 self-reported physical activity questionnaires 1 week later. The authors calculated intraclass correlation coefficients (ICCs) to determine agreement between the number of steps taken and distance walked (Fitbit Zip®) with the visual count of number of steps taken and actual distance walked (walk tests). The authors tested 15 a priori hypotheses about predicted associations between questionnaire scores and physical activity measured by the Fitbit Zip®. Results: Among the 34 participants, there was “excellent” agreement between the number of steps taken measured by the Fitbit Zip® and visually counted number of steps taken (ICC = .99) and number of steps taken at slow (ICC = .75), moderate (ICC = .85), and fast (ICC = .78) walking speeds. There was “poor” agreement between the Fitbit Zip® recorded distance and actual determined distance walked (ICC = .20). Three (20%) construct validity hypotheses were confirmed. Conclusions: The Fitbit Zip® demonstrated criterion validity for its ability to measure number of steps taken but not distance walked, and did not demonstrate construct validity for measuring physical activity among adults with HIV.
Melanna F. Cox, Greg J. Petrucci Jr., Robert T. Marcotte, Brittany R. Masteller, John Staudenmayer, Patty S. Freedson and John R. Sirard
Purpose: Develop a direct observation (DO) system to serve as a criterion measure for the calibration of models applied to free-living (FL) accelerometer data. Methods: Ten participants (19.4 ± 0.8 years) were video-recorded during four, one-hour FL sessions in different settings: 1) school, 2) home, 3) community, and 4) physical activity. For each setting, 10-minute clips from three randomly selected sessions were extracted and coded by one expert coder and up to 20 trained coders using the Observer XT software (Noldus, Wageningen, the Netherlands). The coder defines each whole-body movement which was further described with three modifiers: 1) locomotion, 2) activity type, and 3) MET value (used to categorize intensity level). Percent agreement was calculated for intra- and inter-rater reliability. For intra-rater reliability, the criterion coder coded all 12 clips twice, separated by at least one week between coding sessions. For inter-rater reliability, coded clips by trained coders were compared to the expert coder. Intraclass correlations (ICCs) were calculated to assess the agreement of intensity category for intra- and inter-rater comparisons described above. Results: For intra-rater reliability, mean percent agreement ranged from 91.9 ± 3.9% to 100.0 ± 0.0% across all variables in all settings. For inter-rater reliability, mean percent agreement ranged from 88.2 ± 3.5% to 100.0 ± 0.0% across all variables in all settings. ICCs for intensity category ranged from 0.74–1.00 and 0.81–1.00 for intra- and inter-rater comparisons, respectively. Conclusion: The DO system is reliable and feasible to serve as a criterion measure of FL physical activity in young adults to calibrate accelerometers, subsequently improving interpretation of surveillance and intervention research.