Spontaneous Infant Movement Analysis: A Scoping Review of Non-linear Approaches

Spontaneous movements serve as a critical window into early neuromotor development, offering valuable insights into the maturation of the central nervous system (CNS) and the emergence of motor control strategies. From birth, infants generate intrinsically driven, non-goal-directed movements, which gradually evolve into more structured and intentional motor behaviors as the nervous system develops. The observation and analysis of these movements have been extensively used for the early detection of neurodevelopmental disorders, particularly in preterm infants or those at risk of motor impairments such as cerebral palsy (Tacchino et al., 2020), (Miyagishima et al., 2018), (Marcroft et al., 2015). Traditionally, studies investigating infant spontaneous movements have relied on linear approaches, such as the quantification of movement amplitude, velocity, or frequency. These methods provide valuable information about motor development trajectories, but they are often limited in capturing the complex, adaptive, and dynamic nature of motor control (Igarashi et al., 2017), (Olsen et al., 2015). Human movement is inherently nonlinear, meaning that its variability and complexity cannot be fully described using purely linear models. Nonlinear methods, widely used in fields such as dynamical systems theory and chaos analysis, have the potential to offer deeper insights into movement organization, stability, adaptability, and variability (Deng et al., 2023), (Smith et al., 2017), (Hesse et al., 2018). Over the past two decades, an increasing number of studies have explored nonlinear measures in human movement analysis, particularly in populations with neurological disorders such as stroke or Parkinson’s disease (Igarashi et al., 2016). These measures—such as Lyapunov exponent, entropy-based metrics, recurrence quantification analysis (RQA), fractal dimension, and complexity index—have been successfully applied to gait analysis, postural control, and upper limb movement variability (Groos et al., 2020), (Passmore et al., 2024). Despite this progress, their application in infant spontaneous movement analysis remains scarce and fragmented (Tsuji et al., 2020), (Kieszczynska et al., 2024). Some of the most promising entropy-based measures, such as approximate entropy and sample entropy, have been used to assess movement complexity and variability. These studies indicate that nonlinear approaches can differentiate between healthy and at-risk infants, particularly in the context of preterm birth and early motor dysfunctions (Smith et al., 2017). Several key aspects remain underexplored. First, existing studies often use a single nonlinear measure without systematically comparing the strengths and limitations of different nonlinear approaches in capturing movement dynamics (Deng et al., 2023). Second, there is a lack of studies that combine nonlinear measures with other advanced techniques, such as machine learning or sensor-based motion capture, to enhance movement classification (Cenci et al., 2017). Third, most studies provide cross-sectional data, but there is limited evidence on how nonlinear measures evolve over time in early development (Passmore et al., 2024). Finally, while nonlinear measures have shown promise, their clinical applicability for early diagnosis and intervention remains uncertain. Further research is needed to establish their predictive power for identifying motor impairments before clinical symptoms emerge (Kieszczynska et al., 2024), (Marcroft et al., 2015). Given these gaps, this scoping review aims to answer the following question: What nonlinear measures have been used to analyze spontaneous infant movements in scientific literature? This review seeks to map and synthesize research on nonlinear measures in spontaneous infant movements. Specifically, it aims to identify which nonlinear measures have been applied, examine the methodological approaches used, highlight current limitations and research gaps, and provide recommendations for future research directions to enhance the use of nonlinear methods in early motor assessment.