Developmental Imaging, MCRI, Melbourne, Australia.
Biomedical Engineering, University of Melbourne, Melbourne, Australia.
Sci Rep. 2024 Nov 19;14(1):28598. doi: 10.1038/s41598-024-80202-x.
Over the first few months after birth, the typical emergence of spontaneous, fidgety general movements is associated with later developmental outcomes. In contrast, the absence of fidgety movements is a core feature of several neurodevelopmental and cognitive disorders. Currently, manual assessment of early infant movement patterns is time consuming and labour intensive, limiting its wider use. Recent advances in computer vision and deep learning have led to the emergence of pose estimation techniques, computational methods designed to locate and track body points from video without specialised equipment or markers, for movement tracking. In this study, we use automated markerless tracking of infant body parts to build statistical models of early movements. Using a dataset of infant movement videos (n = 486) from 330 infants we demonstrate that infant movement can be modelled as a sequence of eight motor states using autoregressive, state-space models. Each, motor state Is characterised by specific body part movements, the expression of which varies with age and differs in infants at high-risk of poor neurodevelopmental outcome.
在出生后的头几个月，自发性、不安分的一般运动的典型出现与后期的发育结果有关。相比之下，缺乏不安分的运动是几种神经发育和认知障碍的核心特征。目前，对婴儿早期运动模式的手动评估既费时又费力，限制了它的广泛应用。计算机视觉和深度学习的最新进展带来了姿态估计技术的出现，这些计算方法旨在在没有专用设备或标记的情况下从视频中定位和跟踪身体部位，以进行运动跟踪。在这项研究中，我们使用婴儿身体部位的自动无标记跟踪来构建早期运动的统计模型。使用来自 330 名婴儿的婴儿运动视频数据集（n=486），我们证明婴儿运动可以使用自回归、状态空间模型来建模为八个运动状态的序列。每个运动状态的特点是特定的身体部位运动，其表达随年龄而变化，并且在神经发育结果不良风险高的婴儿中有所不同。
