Using Physiological Big Data to Predict Cross Country Performance The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:38811443 Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Acknowledgements First, I would like to thank Dr. Daniel Lieberman for his support and advice throughout this study. His knowledge, patience, and guidance helped me immensely throughout the process. I also would like to thank my teammates, for this research would not have been possible without their consent for the use of their physiological data. I am also very grateful that WHOOP, Inc. provided me with the opportunity to do this research by allowing me to use their data. I would also like to thank Dr. Charles Czeisler, who helped me understand my results and helped me learn more about sleep, the circadian rhythm, and the body on such short notice. Lastly, I would like to thank Dr. Kevin Rader, Dr. Aaron Baggish, and Bo Waggoner for their help and advice along the way. Table of Contents Introduction ............................................................................................................................................................................... 1 Previous Relevant Research ........................................................................................................................................... 5 Summary ................................................................................................................................................................................. 9 Methods .................................................................................................................................................................................... 11 Subjects ................................................................................................................................................................................. 11 Study Duration and Treatments ................................................................................................................................. 11 Measurements of Performance .................................................................................................................................. 12 Physiological Data Measured ....................................................................................................................................... 14 Survey Data ......................................................................................................................................................................... 15 Statistical Methods ............................................................................................................................................................... 16 Random Effects Models ................................................................................................................................................. 16 Varying Intercept Models ........................................................................................................................................ 16 Varying Slope and Intercept Models ................................................................................................................... 18 Difference from Classical Regression ................................................................................................................. 19 Likelihood Ratio Tests .................................................................................................................................................... 19 Data Aggregation .............................................................................................................................................................. 21 Hypotheses Tested .......................................................................................................................................................... 24 Set 1: Individual Physiological Variables and Performance ...................................................................... 24 Set 2: Investigating the REM Sleep and Total Sleep Relationship ............................................................ 30 Set 3: Investigating Sleep One Night Before and Two Nights Before ..................................................... 34 Set 4: Sleep and Heart Rate Variability .............................................................................................................. 37 Standardized Coefficients ............................................................................................................................................. 40 Results ........................................................................................................................................................................................ 41 Summary Statistics: Sleep Stage Quantity ............................................................................................................. 41 Hypotheses Set 1 Important Results ....................................................................................................................... 42 Hypotheses Set 2 Important Results ....................................................................................................................... 50 Hypotheses Set 3 Important Results ....................................................................................................................... 56 Hypotheses Set 4 Important Results ....................................................................................................................... 62 Weekly Bike Efforts.......................................................................................................................................................... 67 Discussion ................................................................................................................................................................................. 69 Limitations .......................................................................................................................................................................... 76 Conclusions ........................................................................................................................................................................ 79 Appendix.................................................................................................................................................................................... 80 A .............................................................................................................................................................................................. 80 B .............................................................................................................................................................................................. 89 C ........................................................................................................................................................................................... 103 D ............................................................................................................................................................................................ 109 E ........................................................................................................................................................................................... 117 References ..............................................................................................................................................................................132 Introduction We currently live in the age of Big Data, where information is collected every second while the world attempts to derive meaning from it. Many industries are exploring how to use this endless river of information under the assumption that there is no such thing as too much data. One such example is the fitness industry. Over the last few years, wearable fitness tracking devices have become increasingly popular with exercise enthusiasts. The sensors from these devices record a variety of information, from exercise intensity to sleep quality, in an effort to inform individuals about their activity patterns and to influence them to live a healthier lifestyle. All of these data could potentially be useful for many individuals, especially endurance athletes such as cross country distance runners, whose performance is presumably affected by numerous factors. These factors include fitness, freshness on race day, running economy, ability to avoid injury, strategy, energy storage, weight, thermoregulation, and the ability to push oneself to exertion. While each of these factors likely plays a role in the performance of elite runners, it is reasonable to suggest that the most important of these to performance in endurance sports is one’s mental state, overall fitness, and how fresh one is on race day. There is no clear definition of one’s “mental state.” Mental state can be understood in many different ways, including but not limited to: how well one copes with stress, how one can deal with adversity, and