Proof-of-concept of a video analysis for the detection of infant movement gestalts
Introduction
The assessment of general movements in young infants can be used as a prognostic tool to identify infants that will develop cerebral palsy, especially during the fidgety movement’s period (9-18 weeks post term age) [1]. The general movement assessment is a qualitative method that assesses gestalts in the infant movements [1]. Previous studies have shown that fidgety movements (FMs) can be detected and cerebral palsy can be predicted by use of computer-based video analysis software [2]. However, video-based movement analysis has only assessed infant movements as a center of motion and not assessed the gestalts as coordinative movements between the infant’s body segments.
Section snippets
Research Question
The aim of this study was to introduce a proof-of-concept of the validity of computer-based movement analysis capturing gestalts of FMs.
Methods
A typically developing term born infant with a normal motor repertoire was included. Video recording at 12 weeks post term age were observed and analyzed according to Prechtl’s General Movement Assessment by an experienced and certified observer. Three epochs with FMs (duration) and three epochs with non-FMs with similar duration was selected from the video. A video analysis based on fixed-reference multilevel optical flow estimation and principle component analysis was developed according to
Results
93% of the movement variation was explained by 10 PCs. Fig. 1C indicate that PC number 2, 4, and especially 7 was the most sensitive PCs to difference between FMs and non-FMs. The video vector field indicates that composite movements in both arms were most related to FMs which was confirmed by certified observer (see PC #7 in Fig. 1B and D).
Discussion
The PC number 7 had the potential to define the gestalt for the FMs periods that was most present in the upper extremities for this infant. However, more studies are needed(?) to validate the present proof-of-concept on a larger cohort of infants and identify specific measures of the PCs valid for the FM pattern and which could improve predict models of cerebral palsy and other movement disabilities.
References (3)
- et al.
Using computer-based video analysis in the study of fidgety movements
Early Hum Dev
(2009)