Abstract
Talent identification in elite sport increasingly draws on multidimensional data, yet cognitive, physiological, and body-composition measures are rarely captured on one platform. We present a sports big-data platform that integrates Garmin wearable signals, gamified cognitive-function testing ("The Brain Gym"), and bioelectrical body-composition analysis behind athlete- and coach-facing dashboards. Comparing fourteen elite female football players with twelve female university students without regular exercise, we find significantly higher limb muscle mass and faster cognitive movement times in athletes, and a significant negative correlation between one-month median heart rate and cognitive movement and reaction times.
Problem & Motivation
Nurturing exceptional football players is a high-stakes enterprise, yet talent identification still relies heavily on subjective coach assessment. An athlete's profile combines ever-changing cognitive and psychological traits with physiological characteristics during training and recovery, and these dimensions are rarely recorded together. The open question is whether consumer wearables, mobile cognitive testing, and body-composition measurement can be integrated into a single platform that lets coaches monitor — and help anticipate — athletic performance objectively.
Method
We built a sports big-data platform (Flask, MySQL, Jinja2 / Bootstrap, Chart.js) on cloud-hosted servers, integrating three data sources: (1) a Garmin Forerunner 255 wearable streaming heart rate, HRV, SpO2, steps, stress, Body Battery, VO2max, respiration, and sleep through the Garmin Health API; (2) "The Brain Gym", an iPad cognitive-testing app developed by the Institute of Cognitive Neuroscience that measures reaction time, movement time, working-memory capacity, and executive function through game-based tasks; and (3) a Tanita MC-980 MA PLUS bioelectrical-impedance analyzer for BMI, BMR, muscle mass, and limb-specific fat and muscle. Fourteen elite female football players from National Taiwan Normal University (mean age 20.6) formed the observational group, and twelve female university students without regular exercise habits (mean age 21.6) served as controls (ethics ref. 202107EM002). The platform exposes athlete dashboards and a three-column coach interface for team-wide monitoring.
Findings
- Body composition: elite female football players had significantly higher limb muscle mass than controls, consistent with long-term training adaptation.
- Muscle mass and motor speed: muscle mass correlated negatively with cognitive movement time (r = −0.4, p = 0.041) — more muscle mass, faster movements.
- Heart rate and cognition: one-month median heart rate correlated negatively with movement time (r = −0.83, p = 0.011) and reaction time (r = −0.73, p = 0.042).
- Coaching use: daily heart-rate data may serve as a quick reference for talent identification and training-load adjustment.
Implications
The study demonstrates a feasible multimodal platform that unifies consumer wearables, cognitive testing, and body composition behind athlete- and coach-facing dashboards. The same wearable-measurement architecture underpins physiological monitoring for learning analytics in the Educational Omics program (PhysioNeuromics dimension): where a coach reads an athlete's physiological state, an instructor can read a learner's. This cross-domain transfer — from sports science to the classroom — is an early step toward physiologically-aware learning support. Heart-rate and HRV associations are reported as correlations, not as causal or affective claims.
Citation
BibTeX
@incollection{chang2024sports_wearable,
author = {Chia-Kai Chang and Yu-Lun Chen and Chi-Hung Juan},
title = {Predicting Sports Performance of Elite Female Football Players Through Smart Wearable Measurement Platform},
booktitle = {Progress in Brain Research},
publisher = {Elsevier},
year = {2024},
issn = {0079-6123},
doi = {10.1016/bs.pbr.2024.04.002},
}