Spontaneous Leg Movements Measured by Wearable Sensors in Infancy Differentiate Later Risk for Cerebral Palsy

BACKGROUND: To determine if early spontaneous leg movements of infants differentiate risk status for cerebral palsy (CP) in infants who had a perinatal brain injury. METHODS: Infants born at risk for CP (due to a perinatal brain injury) were classified as high risk for CP (n = 33) or not at high risk (n = 21), at 4 months of age, using General Movements Assessment and Test of Infant Motor Performance. Full-day leg movement data were recorded using wearable inertial sensors on each ankle (2 days per month, from age 1 to 4 months). Kinematic parameters were derived from accelerometer and gyroscope data as follows: (1) number of leg movements per hour awake; (2) median leg movement acceleration; (3) median leg movement duration; and (4) fuzzy entropy of peak leg movement acceleration. The leg kinematic parameters measured at month 1 and month 4 were analyzed using mixed-effects models. RESULTS: Infants classified as high risk for CP had statistically significantly fewer number of leg movements per hour awake and shorter median leg movement duration than infants not at high risk (P= 0.0322 and P= 0.0023, respectively). CONCLUSIONS: The spontaneous movements of infants can be characterized in their natural environments using wearable sensors and have potential to inform recommendations for early intervention and follow-up with experts in CP diagnosis.