Peer-reviewed article
Artificial intelligence algorithms effectively classify 38 movements in infants born full-term and preterm recorded in the laboratory and at home.
Yohanes Purwanto, Erick Chandra, Po-Nien Tsao, Ting-An Yen, Wei-Chih Liao et al. ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This study developed and tested artificial intelligence algorithms to automatically recognize and classify 38 different infant movements in babies from birth to 18 months old. The AI was trained using videos of full-term and preterm infants recorded both in a clinical laboratory and at home by parents using a mobile app. The AI achieved high accuracy in classifying movements in laboratory videos and reasonable accuracy in home videos, and the results correlated well with assessments by physiotherapists.
Key points
- AI model achieved 91% accuracy classifying 38 movements from laboratory videos and 84% accuracy from home videos
- The system worked for both full-term and preterm infants across ages 4 to 18 months
- Age-based movement sets showed 99% agreement with standard physiotherapist assessment (AIMS)
- The approach combined controlled laboratory recording with parent-uploaded home videos via the Baby Go mobile application
- Home video classification was less accurate than laboratory video, likely due to variation in lighting, angles and recording quality
Limitations
The study does not report whether the AI system actually improved early detection of developmental delays or changed clinical outcomes for infants. It also does not discuss the usability, acceptability or feasibility of the mobile app for parents in routine clinical practice.
Abstract
BACKGROUND: Although recent advancements in artificial intelligence (AI) provide an alternative for infant motor assessment, such applications focus mainly on infants' movements within a narrow age range in a laboratory setting. This study aimed to develop and validate AI algorithms for movement recognition in full-term and preterm children in the laboratory and at home throughout infancy. METHODS: This prospective cohort study included 85 full-term infants and 84 preterm infants who were video recorded during the Alberta Infant Motor Assessment (AIMS) administered by physiotherapists in a laboratory, while parents uploaded movement videos at home using the Baby Go mobile application (app) from 4 to 18 months of age. An AI model was developed to classify movements using physiotherapists' labeling results as the ground truth. The AI-classified movements were organized into age-based sets to test concurrent validity against the AIMS examiners' results. RESULTS: Validation of the AI model for classifying 38 movements in full-term and preterm infants revealed an accuracy of 0.91, precision of 0.92, recall of 0.90, and F1 score of 0.91 with the laboratory videos and an accuracy of 0.84, precision of 0.84, recall of 0.77, and F1 score of 0.78 with the home videos. These movements were dispatched into age-based sets with two to five movements per age that showed high concurrent validity with the AIMS results (agreement = 0.99). CONCLUSION: The AI model accurately classified 38 movements in full-term and preterm infants performed in the laboratory and at home. The age-based sets also highly correlated with the physiotherapists' assessment results.
Findings the Hub has extracted
Each claim below is labelled by its type of design and links back to this report.
- Intervention
- AI algorithm trained to automatically classify 38 infant movements from video recordings, administered in laboratory and home settings via mobile app
- Population
- 85 full-term infants and 84 preterm infants, ages 4 to 18 months
Key finding
Laboratory video accuracy: 91% (precision 0.92, recall 0.90, F1 score 0.91)