Related Experiment Video
Updated: Jul 2, 2025
Quantified Assessment of Infant's Gross Motor Abilities Using a Multisensor Wearable
Published on: May 17, 2024
Automatic assessment of infant carrying and holding using at-home wearable recordings
Manu Airaksinen1,2, Einari Vaaras3, Leena Haataja4,5
1BABA Center, Pediatric Research Center, Department of Clinical Neurophysiology, New Children's Hospital and HUS Imaging, Helsinki University Hospital, Helsinki, Finland. manu.airaksinen@hus.fi.
Insights
Automated detection of infant carrying and holding (C/H) is now feasible using wearable sensors. This technology offers new insights into infant development and parent-child interactions during at-home studies.
Area of Science:
- Developmental Psychology
- Wearable Technology
- Machine Learning in Healthcare
Background:
- Assessing infant carrying and holding (C/H), a key physical infant-caregiver interaction, is crucial for developmental research.
- Automated C/H detection is needed for long-term, at-home studies using wearable devices to monitor infant neurobehavior.
Purpose of the Study:
- To develop and evaluate deep learning classifiers for automatic detection of infant C/H behaviors from multi-sensor wearable recordings.
- To compare automated C/H detection performance with human expert agreement and existing methods like actigraphy.
Main Methods:
- Developed a phenomenological categorization for five C/H behavior definitions.
- Trained and assessed deep learning classifiers using multi-sensor wearable data from infant motor development assessments.
- Benchmarked automated detection against video-based human expert agreement and actigraphy.
Main Results:
- Automated C/H detection achieved few-second temporal accuracy, with the best definition yielding 96% accuracy and 0.56 kappa.
- Performance varied with C/H definition, influenced by infant movement presence.
- Actigraphy-based methods were found to ignore typical C/H behaviors.
Conclusions:
- Deep learning-based automated detection of infant C/H from wearable sensors is feasible and accurate.
- This method offers a valuable tool for studying C/H behavior across infant development, providing novel insights when matched with motor abilities.
- Automated C/H detection surpasses traditional actigraphy in capturing these essential interactions.
Abstract:
Assessing infant carrying and holding (C/H), or physical infant-caregiver interaction, is important for a wide range of contexts in development research. An automated detection and quantification of infant C/H is particularly needed in long term at-home studies where development of infants' neurobehavior is measured using wearable devices. Here, we first developed a phenomenological categorization for physical infant-caregiver interactions to support five different definitions of C/H behaviors. Then, we trained and assessed deep learning-based classifiers for their automatic detection from multi-sensor wearable recordings that were originally used for mobile assessment of infants' motor development. Our results show that an automated C/H detection is feasible at few-second temporal accuracy. With the best C/H definition, the automated detector shows 96% accuracy and 0.56 kappa, which is slightly less than the video-based inter-rater agreement between trained human experts (98% accuracy, 0.77 kappa). The classifier performance varies with C/H definition reflecting the extent to which infants' movements are present in each C/H variant. A systematic benchmarking experiment shows that the widely used actigraphy-based method ignores the normally occurring C/H behaviors. Finally, we show proof-of-concept for the utility of the novel classifier in studying C/H behavior across infant development. Particularly, we show that matching the C/H detections to individuals' gross motor ability discloses novel insights to infant-parent interaction.
