A time to exhaustion model during prolonged running based on wearable accelerometers

Area of Science:

• Biomechanics
• Sports Science
• Human Movement Analysis

Background:

• Optimizing training requires understanding the relationship between running mechanics and fatigue.
• Accelerometry offers a non-invasive method for collecting kinematic data during locomotion.

Purpose of the Study:

• To develop a biomechanical model predicting time to exhaustion using accelerometry data.
• To evaluate the influence of sensor placement on fatigue prediction accuracy.

Main Methods:

• Ten volunteers performed a treadmill running test until exhaustion at a constant speed (13.5 km/h).
• Three accelerometers were placed on the foot, tibia, and lumbar spine (L4-L5).
• Multiple linear regression analyses were used to model time to exhaustion based on derived indicators.

Main Results:

• Accurate prediction of time to exhaustion was achieved using simultaneous measurements (R² and 21 indicators).
• Individual models showed predictive capabilities for lumbar (R² and 11 indicators), tibia (R² and 11 indicators), and foot (R² and 12 indicators).
• The location of accelerometry measurement significantly influenced the information and predictive accuracy.

Conclusions:

• Accelerometry-based biomechanical modeling can accurately predict time to exhaustion in running.
• Different body locations provide unique insights into running fatigue mechanisms.
• Future research should investigate homogeneous populations to refine predictive models and minimize errors.