Related Experiment Video
Updated: Apr 5, 2026
Quantified Assessment of Infant's Gross Motor Abilities Using a Multisensor Wearable
Published on: May 17, 2024
Development of a machine learning model to detect toddlers' physical activity and sedentary time using
Elyse Letts1, Sara King-Dowling2, Natascja Di Cristofaro1
1Child Health & Exercise Medicine Program, Department of Pediatrics, McMaster University, Canada.
Insights
A new machine learning model accurately detects toddlers' physical activity (PA) and sedentary time (SED), outperforming existing methods. This advancement improves measurement of children's movement patterns.
Area of Science:
- Pediatric exercise science
- Biomedical engineering
- Machine learning applications in health
Background:
- Accelerometer-based physical activity (PA) and sedentary time (SED) detection methods face challenges with toddler movement patterns.
- Accurate measurement of PA and SED in toddlers is crucial for understanding health outcomes and developing interventions.
Purpose of the Study:
- To develop a novel machine learning (ML) model for detecting physical activity (PA) and sedentary time (SED) in toddlers.
- To compare the performance of the developed ML model against existing cut-point methods for analyzing toddler PA.
Main Methods:
- 111 toddlers (21±7 months) wore waist-worn accelerometers during two 1-hour visits.
- Video recordings provided ground truth for activity classification (SED, total PA (TPA), light PA (LPA), moderate-to-vigorous PA (MVPA), non-volitional movement (NVM)).
- Four gradient boosted tree ML models were trained using extracted accelerometer features and validated against existing methods.
Main Results:
- ML models achieved 82% (NVM/SED/TPA) and 74% (NVM/SED/LPA/MVPA) accuracy, with low mean absolute differences (3.0-3.2 min/h).
- Existing methods showed significantly lower accuracies (33-74%) and higher mean absolute differences (7.6-25.8 min/h) for TPA and MVPA estimation.
- The developed ML models demonstrated superior performance in classifying toddler movement patterns compared to traditional methods.
Conclusions:
- The novel ML models (NVM/SED/TPA and NVM/SED/LPA/MVPA) offer a significant improvement for measuring toddlers' physical activity and sedentary time.
- These models align with PA guidelines and the health links of MVPA, providing a more accurate assessment tool.
- An open-access, user-friendly interface for these ML models is provided, facilitating wider adoption and advancing research in pediatric PA measurement.
Objectives:
Toddler movement patterns challenge current accelerometer-based detection of physical activity (PA) and sedentary time (SED). The objectives of this study were to: (1) develop a novel machine learning (ML) model to detect toddlers' PA and SED; and (2) compare this ML model to existing cut-point methods to analyze toddlers' PA (independent sample cross-validation of existing methods).
Design And Methods:
We recruited 111 toddlers (21 ± 7 months; 51% female) to two 1-hour semi-structured visits wearing a waist-worn ActiGraph wGT3X-BT accelerometer. Video recordings were manually annotated using a modified Children's Activity Rating Scale to determine a ground truth. We extracted 40 time and frequency domain features from raw accelerations and trained 4 gradient boosted tree ML models (distinguishing SED, total PA (TPA), light PA (LPA), moderate-to-vigorous PA (MVPA), and non-volitional movement (NVM)). Models were assessed using accuracy, F1 scores, and confusion matrices. For the validation of 11 existing methods, we calculated accuracy, F1, and mean absolute differences in TPA and MVPA estimation.
Results:
ML models classifying NVM/SED/TPA and NVM/SED/LPA/MVPA reached 82% and 74% accuracy with mean absolute differences of 3.0 and 3.2 min/h, respectively. Independent sample cross-validation found accuracies from 33 to 74% and mean absolute differences from 7.6 to 18.6 min/h in TPA and 10.7 to 25.8 min/h in MVPA.
Conclusions:
We recommend the NVM/SED/TPA or NVM/SED/LPA/MVPA models' given alignment with toddler TPA guidelines and MVPA link to health outcomes, respectively. We additionally present an open-access interface for using these ML models that does not require coding knowledge. This presents a substantial step forward in the measurement of toddlers' physical activity.
