AI Approaches towards Prechtl’s Assessment of General Movements: A Systematic Literature Review

Abstract

General movements (GMs) are spontaneous movements of infants up to five months post-term involving the whole body varying in sequence, speed, and amplitude. The assessment of GMs has shown its importance for identifying infants at risk for neuromotor deficits, especially for the detection of cerebral palsy. As the assessment is based on videos of the infant that are rated by trained professionals, the method is time-consuming and expensive. Therefore, approaches based on Artificial Intelligence have gained significantly increased attention in the last years. In this article, we systematically analyze and discuss the main design features of all existing technological approaches seeking to transfer the Prechtl’s assessment of general movements from an individual visual perception to computer-based analysis. After identifying their shared shortcomings, we explain the methodological reasons for their limited practical performance and classification rates. As a conclusion of our literature study, we conceptually propose a methodological solution to the defined problem based on the groundbreaking innovation in the area of Deep Learning.

The future of General Movement Assessment: The role of computer vision and machine learning – A scoping review

“…We should not forget that GMs have a large, complex, and variable repertoire, bringing difficulties for machine learning approaches to acquire a representative dataset. For example, when relying on retrospective videos from infants with atypical development, due to the uncertain representativeness of the training datasets, it might be challenging to achieve high external validity when testing the created model on novel samples ( Irshad et al, 2020 ).…”

“…To maintain the non-intrusive character of the GMA, vision-based markerless approaches appear more favourable than the ones using wearable sensors, or attaching markers to the infant’s body. Although marker- and wearable sensor-based approaches have technical merits (see Irshad et al, 2020 ), it is yet to be examined whether these markers or sensors may interfere with the infants’ spontaneous movements, or whether the device-attaching procedures, usually time-consuming and during which the infant has to be touched or manipulated, could affect the infant’s consequential behavioural state (e.g., becoming fussy and distracted).…”

“…Indeed, in the past two decades, a boom of technological approaches aiming at automated or technology-assisted GMA have surfaced. These efforts range from mobile-app-based recording tools, e.g., the Baby Moves ( Spittle et al, 2016 ) and the GMApp (Marschik, Pokorny, et al, 2017), to automated pose estimation through sensor-based or markerless approaches (e.g., Irshad, Nisar, Gouverneur, Rapp, & Grzegorzek, 2020 ; Marcroft, Khan, Embleton, Trenell, & Plötz, 2015 ; Marschik et al, 2017 ).…”

“…We focus on video-based approaches only, since GMA is in origin a visual-based method. Advanced computer vision technology remains authentic to the non-intrusive character of the classic GMA, allowing automated analyses of the infant’s spontaneous movements, which is not influenced by the use of wearable sensors and other devices (for recent overviews on diverse sensors targeting GMs, please see Hyde et al, 2019 ; Irshad et al, 2020 ; Marcroft et al, 2015 ). Different from previous reviews, we examine in particular if the existent technological attempts targeting automated GMA are in accordance with the fundamental concepts underlying this unique clinical tool.…”

RGB-D Videos-Based Early Prediction of Infant Cerebral Palsy via General Movements Complexity

“…Although GMA has proven to be a non-invasive, and accurate diagnostic tool, it requires significant time and resource investment to train evaluators. Therefore, several studies have been conducted in an attempt to evaluate the feasibility of automated GMA [12]. In 2001, Prechtl proposed GMA [2], which analyzed the spontaneous movement of infants and can effectively evaluate the function of infants' nervous system.…”