• No results found

In this study we have shown that VO2max can be predicted from submaximal running using data fusion of heart rate and accelerometer measurements. We extracted various features from these measurements and used a data-driven approach based on greedy forward selection to select a small subset of these features. The best model found in this paper has an explained variance of 0.784 and uses four features: two descriptive features (gender and body weight), one heart rate feature (HR-10) and one accelerometer feature (SD-1t,total,0). This model can predict an individual’s maximal oxygen uptake from

objective variables calculated from running on a treadmill at only 8 or 9 km · hr−1for four minutes.

While the method presented in this study is accessible for recreational runners, it would be more convenient if VO2max can be predicted during outdoor running instead.

Future work can explore whether the feature extraction and selection methods used in this paper can be applied for outdoor running as well. It will be important then to find features that are independent of running speed.

References

1. Coyle EF. Integration of the physiological factors determining endurance performance ability. Exercise and sport sciences reviews. 1995;23(1):25–64.

2. Joyner MJ. Modeling: optimal marathon performance on the basis of physiological factors. Journal of Applied Physiology. 1991;70(2):683–687.

3. Abut F, Akay MF, George J. Developing new VO2 max prediction models from maximal, submaximal and questionnaire variables using support vector machines combined with feature selection. Computers in Biology and Medicine.

2016;79:182–192.

4. Artificial neural network models for predicting maximum oxygen uptake from submaxiaml exercise involving walking, jogging or running, author=Acikkar, M and Akay, MF and George, JD and Delil, M and Akt¨urk, E,

booktitle=International Symposium on Electrical and Electronics Engineering and Computer Systems. Lefka, North Cyprus, pages=17–21, year=2012;.

5. Cao ZB, Miyatake N, Aoyama T, Higuchi M, Tabata I. Prediction of maximal oxygen uptake from a 3-minute walk based on gender, age, and body composition.

Journal of Physical Activity and Health. 2013;10(2):280–287.

6. Abut F, Akay MF, Yildiz I, George JD. Performance Comparison of Different Machine Learning Methods for Prediction of Maximal Oxygen Uptake from Submaximal Data; 2015. p. 367–370.

7. Akay MF, Zayid EIM, Akt¨urk E, George JD. Artificial neural network-based model for predicting VO 2 max from a submaximal exercise test. Expert Systems with Applications. 2011;38(3):2007–2010.

8. George JD, Stone WJ, Burkett LN. Non-exercise VO2max estimation for physically active college students. Medicine and science in sports and exercise.

1997;29(3):415–423.

9. Bradshaw DI, George JD, Hyde A, LaMonte MJ, Vehrs PR, Hager RL, et al. An accurate VO2max nonexercise regression model for 18–65-year-old adults.

Research quarterly for exercise and sport. 2005;76(4):426–432.

10. Akay MF, Inan C, Bradshaw DI, George JD. Support vector regression and multilayer feed forward neural networks for non-exercise prediction of VO 2 max.

Expert Systems with Applications. 2009;36(6):10112–10119.

11. Shenoy S, Tyagi B, Sandhu J, Sengupta D. Development of non-exercise based VO2max prediction equation in college-aged participants in India. The Journal of sports medicine and physical fitness. 2012;52(5):465–473.

12. Weyand PG, Kelly M, Blackadar T, Darley JC, Oliver SR, Ohlenbusch NE, et al.

Ambulatory estimates of maximal aerobic power from foot-ground contact times and heart rates in running humans. Journal of Applied Physiology.

2001;91(1):451–458.

13. T¨onis T, Gorter K, Vollenbroek-Hutten M, Hermens H. Comparing VO2max determined by using the relation between heart rate and accelerometry with submaximal estimated VO2max. The Journal of sports medicine and physical fitness. 2012;52(4):337–343.

15. Borg GA. Psychophysical bases of perceived exertion. Med sci sports exerc.

1982;14(5):377–381.

16. Moe-Nilssen R. A new method for evaluating motor control in gait under real-life environmental conditions. Part 1: The instrument. Clinical Biomechanics.

1998;13(4-5):320–327.

17. Buskirk E, Hodgson J. Age and aerobic power: the rate of change in men and women. In: Federation proceedings. vol. 46; 1987. p. 1824–1829.

18. Rowell LB. Human circulation: regulation during physical stress. Oxford University Press, USA; 1986.

19. McGregor SJ, Busa MA, Yagie JA, Bollt EM. High resolution MEMS

accelerometers to estimate VO2 and compare running mechanics between highly trained inter-collegiate and untrained runners. PLOS ONE. 2009;4(10).

20. Sch¨utte KH, Maas EA, Exadaktylos V, Berckmans D, Venter RE, Vanwanseele B.

Wireless tri-axial trunk accelerometry detects deviations in dynamic center of mass motion due to running-induced fatigue. PLOS ONE. 2015;10(10):1–12.

21. Hoerl AE, Kennard RW. Ridge regression: Biased estimation for nonorthogonal problems. Technometrics. 1970;12(1):55–67.

22. Tibshirani R. Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society Series B (Methodological). 1996; p. 267–288.

23. Gaudino P, Gaudino C, Alberti G, Minetti AE. Biomechanics and predicted energetics of sprinting on sand: hints for soccer training. Journal of Science and Medicine in Sport. 2013;16(3):271–275.

Bijlage B

Poster

maximal rate at which an individual can consume oxygen (ml/kg/min) VO2max

Depending on age, gender, genetic markers, training status, …

Upper limit of endurance performance

Measured using a

maximal incremental exercise test Practical limitations

Expensive lab set-up required Physical limitations

Risks for heart disease patients, physically demanding, hard to incorporate in strict training regimen

Predict from

submaximal exercise Based on data fusion of physiological and

• 1 min. rest between stages

• Increasing running speed

• Until volitional exhaustion Data collected during the test:

• Heart rate

• Acceleration from tibia, lower back and upper back

• Oxygen consumption (VO2)

25 subjects, of which From warm-up and first 3

stages (submaximal effort)

Highly intercorrelated features

Not all features useful for predicting VO2max

• Descriptive features (D): gender and body weight

• Heart rate (HR): (reciprocal of) mean heart rate for each stage [8 features]

• Biomechanical features (B): (reciprocal of) mean, stddev, RMS, energy, phase plot height, width and ratio (=width/height) for each direction (x, y, z, total) at each

accelerometer, for each stage [672 features] Best result for S4 R2= 0.784 S1(D)

S2(D+HR) S3(D+B) S4(D+HR+B)

S1= {gender, body weight}

S3= S1 {phase plot ratio of total lower back acc. in stage 2}

S2= S1 {reciprocal of mean heart rate in warm-up stage}

S4= S1 S2 {reciprocal of stddev of total tibia acc. in warm-up stage}

Best result:

• Least squares linear regression + greedy forward selection

+ physiological and biomechanical features

• Selected features are from the warm-up stage only:

minimal physical effort required

Prof. Jesse Davis, Wannes Meert, Prof. Benedicte Vanwanseele

• Improved results of model by Weyand et al.

• No questionnaires or self-chosen running speed Select best feature

Add f* toS if

Select features Susing greedy forward selection:

a

Evaluated by R2with leave-one-subject-out cross-validation

S4(D+HR+B) S2(D+HR) S1(D)