Movidea: An Innovative Semi-Automatic Software to Detect Early Motor Signatures in Autism Spectrum Disorder

Poster Presentation
Thursday, May 2, 2019: 5:30 PM-7:00 PM
Room: 710 (Palais des congres de Montreal)
F. Fulceri1, L. Gila1, A. Caruso1, T. Salvitti1, M. Bulgheroni2, W. Baccinelli2 and M. L. Scattoni1, (1)Research Coordination and Support Service, Istituto Superiore di Sanità, Rome, Italy, (2)Ab. Acus srl, Milan, Italy
Background: Motor abnormalities have been reported as one of the earliest markers of autism spectrum disorder (ASD). The Italian network for early detection of ASD (NIDA) has been established in 2012 with the aim to define a surveillance protocol of High-Risk (HR, siblings of children with a diagnosis of ASD) and Low-Risk infants (LR, siblings of typically developing children) from birth to 36 months of age. The NIDA clinical/biological protocol includes also video recordings of spontaneous motor behaviors in the first 6 months of life. Based on the Pretchl’s method, the NIDA network has recently detected abnormalities in general movements of HR infants during the first two weeks of age.

Objectives: To identify early motor signatures of ASD through an innovative, semi-automatic and advanced tool.

Methods: MOVIDEA is a software aimed to analyze bi-dimensional video-recordings of infant’s movements by implementing a semi-automatic tracking of limbs. MOVIDEA has been developed by Ab.Acus, an Italian company with high expertise in the field of biomedical science and in developing advanced technologies for studying human kinematic. The measures extracted by the software have been validated by a child psychiatrist and two researchers of the Istituto Superiore di Sanità (ISS) with high expertise in infant’s motor assessment and analysis. The NIDA network enrolled 113 LR and 276 HR infants. For each infant, five structured video-recordings have been collected [at 10 days, 6-12-18-24 weeks]consisting of approximately 1900 video recordings. The median length of each scored video is 3 minutes and the median length of time to analyze it is 30 minutes. Infants were subdivided in three groups based on their clinical outcome: typical development (TD), neurodevelopmental disorders (NDDs) and autism spectrum disorder (ASD). Statistical analysis was performed with STATA 13.1.

Results: Between-group differences emerged in the standard deviation of y-directions of the centroid of motion (spatial center of the positive pixels in the motion image) at 10 days [NDD>TD], in the Cerebral Palsy Predictor [NDD>TD at 10 days and 6 weeks], in the cross-correlation indexes between right and left foot [NDD>TD at 24 weeks], in the area differing from moving average of foots [NDD>TD at 10 days] and of right foot both in x- and y-direction [NDD>TD at 10 days ]. No differences emerged in the quantity of motion and in other values such as the centroid of motion, the area from moving average and cross-correlation indexes.

Conclusions: The NIDA Network is a longitudinal prospective study aimed to identify early signs of ASD. Preliminary data suggest that infants later diagnosed with NDD presented early motor abnormalities. MOVIDEA is a useful innovative tool to describe early motor development, but further research should be performed to increase sample’s size.