Towards novel classification of infants' movement patterns supported by computerized video analysis

Area of Science:

• Pediatric neurology
• Developmental pediatrics
• Movement analysis

Background:

• Infant positional preferences and movement asymmetry can signal abnormal clinical conditions.
• Lack of standardized nomenclature complicates assessment and documentation of infant motor patterns.
• Video analysis offers a safe, reproducible method for assessing infant movements, crucial for physiotherapy and goal planning.

Purpose of the Study:

• To develop an objective classification system for infant movement patterns.
• To emphasize the distribution of muscle tension in infant movements.
• To enhance accuracy and reproducibility in infant assessments using computer analysis of video recordings.

Main Methods:

• Recorded spontaneous supine movements of 51 healthy infants (6-15 weeks old).
• Experts identified postural-motor patterns, including symmetry, typical asymmetry (linked to asymmetrical tonic neck reflex), and deviations.
• Developed a computer-based inference system for automatic classification of movement patterns.

Main Results:

• Classified motor patterns into normal (typical and atypical), positional preference, and abnormal.
• The automatic classification system achieved 84% accuracy in mapping expert decisions.
• High reproducibility was confirmed for atypical patterns; typical patterns showed lower reproducibility (≤70%).

Conclusions:

• Infant spontaneous movements can be categorized into typical and atypical patterns.
• Computer-based analysis of infant movement patterns objectifies and reliably reproduces diagnostic decisions.