Related Experiment Video
Updated: Aug 16, 2026
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Automated detection of abnormal general movements from pressure and positional information in hospitalized infants
Nathalie L Maitre1, Caitlin P Kjeldsen2, Andrea F Duncan3
1Department of Pediatrics, Emory University School of Medicine and Children's Healthcare of Atlanta, Atlanta, GA, USA. nmaitre@emory.edu.
Insights
Automated detection of abnormal (cramped synchronized, or CS) general movements assessment (GMA) in infants shows promise for early identification of neuromotor disorders. This machine learning approach may increase access to GMA screening for at-risk newborns.
Area of Science:
- Neonatal neurology
- Developmental pediatrics
- Computational neuroscience
Background:
- Prechtl's general movements assessment (GMA) identifies abnormal (cramped synchronized, or CS) movement patterns with high sensitivity for predicting neuromotor disorders.
- Clinical adoption of GMA is limited by training needs and subjective interpretation.
Purpose of the Study:
- To develop and validate a preliminary, automated approach for detecting CS movements in infants using machine learning.
- To create a clinically applicable software interface for automated GMA analysis.
Main Methods:
- A three-phased approach combining unsupervised and supervised machine learning.
- Dual data collection using video and a pressure-sensor mat from 335 hospitalized infants.
- Feature extraction from clinician- and mat-derived data, followed by classification modeling.
Main Results:
- A classification model integrating normalization, clustering, and decision tree methods achieved 100% sensitivity for CS movements.
- Automated results were available within 20 minutes of data recording via a software interface.
Conclusions:
- A feasible preliminary method for automated detection of abnormal GMA in neonates was developed by combining clinical research, machine learning, and sensor mat technology.
- Further software and algorithm refinement is necessary for widespread clinical use.
Background:
Prechtl's general movements assessment (GMA) allows visual recognition of movement patterns that, when abnormal (cramped synchronized, or CS), have very high sensitivity in predicting later neuromotor disorders; however, training requirements and subjective perceptions from some clinicians may hinder universal adoption of the GMA in the newborn period.
Methods:
To address this, we used a three-phased approach to design a preliminary and clinically-oriented approach to automated CS GMA detection. 335 hospitalized infants were dually recorded on video and a pressure-sensor mat that collected time, spatial, and pressure data. Video recordings were scored by advanced GMA readers. We then conducted a series of unsupervised machine learning and supervised classification modeling with features extracted from clinician- and mat-driven datasets. Finally, the resulting algorithm was converted to a software interface.
Results:
A classification model combining normalization, clustering, and decision tree modeling resulted in the highest sensitivity for CS movements (100%). Results were delivered via the software interface within 20 min of data recording.
Conclusion:
The combination of clinical research, machine learning, and repurposing of existing sensor mat technology produced a feasible preliminary approach to automatically detect abnormal GMA in infants while still in the NICU. Further refinements of software and algorithms are needed.
Impact Statement:
Machine learning can differentiate cramped synchronized general movement patterns in the neonatal intensive care unit with good sensitivity and specificity. Increasing access to the GMA through automated detection methods may allow for earlier identification of a greater number of children at high risk for movement delay. Large studies leveraging new artificial intelligence approaches could increase the impact of such detection.
