The General Movements Motor Optimality Score in High-Risk Infants: A Systematic Scoping Review

Abstract

Purpose: 

The aim of this systematic scoping review was to explore the use of the motor optimality score in the fidgety movement period in clinical practice, and to investigate evidence for the motor optimality score in predicting neurodevelopmental outcomes.

Summary of Key Points: 

Thirty-seven studies, with 3662 infants, were included. Studies were conceptualized and charted into 4 categories based on the motor optimality score: prediction, outcome measure, descriptive, or psychometric properties. The most represented populations were preterm or low-birth-weight infants (16 studies), infants with cerebral palsy or neurological concerns (5 studies), and healthy or term-born infants (4 studies).

Conclusion: 

The motor optimality score has the potential to add value to existing tools used to predict risk of adverse neurodevelopmental outcomes. Further research is needed regarding the reliability and validity of the motor optimality score to support increased use of this tool in clinical practice. What this adds to the evidence: The motor optimality score has potential to improve the prediction of adverse neurodevelopmental outcomes. Further research on validity and reliability of the motor optimality score is needed; however, a revised version, the motor optimality score-R (with accompanying manual) will likely contribute to more consistency in the reporting of the motor optimality score in future.

INTRODUCTION

There is evidence that early, targeted intervention can improve functional outcomes for infants at risk of neurodevelopmental impairment such as cerebral palsy (CP).^1–4 However, referral for early therapy programs and specialist input relies on accurate identification of infants at risk. The field has made significant progress in the early identification of adverse neurodevelopmental outcomes, although many conditions are still only diagnosed in later infancy, and tools to predict severity of outcomes at an early age are still lacking. The general movements assessment (GMA) motor optimality score (MOS)⁵ may be an accessible tool that can contribute to our ability to identify infants at risk of adverse outcomes at an early age, with the potential to add to the ability to predict severity of outcomes.

The GMA is recognized as an important tool in the early and accurate diagnosis of CP, with the international guidelines for early detection of CP recommending use of the GMA at 3 to 5 months of age, in combination with the Hammersmith Infant Neurological Examination and brain imaging.² There is evidence in the literature supporting the use of the GMA for prediction of CP.^6–9 The association between the GMA and other neurodevelopmental disorders, such as cognitive or behavioral disorders, developmental motor delay, autism spectrum disorder (ASD), and intellectual disability, has been investigated^10–14 although the accuracy of the GMA to predict outcomes other than CP is significantly lower.

In addition to identifying the presence or absence of fidgety movements at 3 to 5 months of age using the GMA, the infant's movement and posture can be analyzed using the MOS.⁵ The MOS involves using a standardized scoring form to assess the quality of movements observed during the GMA video. The following domains are scored: (i) quantity and normality of movement patterns (eg, kicking and hand-to-hand contact); (ii) age-related movement repertoire; (iii) observed postural patterns (eg, head and trunk symmetry); and (iv) movement character (eg, smooth and fluent, monotonous). The original MOS was published in 2004 in the Prechtl Manual.⁵ In 2009, clarifications were made regarding the number of normal patterns of movement required for age adequacy and these scoring criteria have been applied in many publications since.¹⁵ The most recent consensus revised version (MOS-R) was developed by Einspieler and colleagues,¹⁶ with the addition of definitions to assist with scoring. The highest possible score on the MOS is 28 points; however, a score between 25 and 28 is regarded as optimal.^16,17 For infants with absent fidgety general movements (GMs), the highest score possible is 17. Clinicians able to use this assessment have received training at an advanced level from the General Movements Trust and are certified assessors.

There have been no systematic reviews reporting on the use of the MOS in clinical practice or the reliability of the MOS in predicting neurodevelopmental outcomes. There are, however, indications that components of the MOS may be associated with adverse neurodevelopmental outcomes.¹⁰

The aim of this study was to explore the current evidence regarding the use of the MOS, including when it is used in clinical practice, and the evidence to support use of the MOS in predicting neurodevelopmental outcomes. A systematic scoping review was chosen due to the heterogeneity of available studies. The goal of a scoping review is to determine the range of research findings, identify gaps in research, and clarify conceptual boundaries of a topic.¹⁸

The review questions included: (1) how is the MOS used in clinical practice for infants 3 to 5 months of age (fidgety movement period)? (2) Can the MOS predict neurodevelopmental outcomes? And if so, (3) what are these outcomes and are there specific components of the MOS that may indicate that children are at risk of specific adverse outcomes? (4) Is there evidence to support the validity and inter-/intrarater reliability of the MOS?

METHODS

This scoping review is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR).¹⁸ We followed the guidelines by the Joanna Briggs Institute (JBI).¹⁹ A scoping review was used to determine the range of information available for use of the MOS in clinical practice. We were interested in the original and revised versions of the MOS.

Data Sources

A systematic literature search was performed up to April 2021 using the following databases: Medline, CINAHL, Embase, and Google Scholar. A search based on the following key terms was used: general movements, Prechtl, optimality list, optimality score, motor repertoire, and MeSH (Medical Subject Headings) term “cerebral palsy.” A comprehensive search of the General Movements Trust website, an author search and reference searches of eligible studies were subsequently conducted. No studies prior to 2008 met criteria. EndNote 20²⁰ was used for citation collation and duplicates were removed manually. Ethical approval was not required, as this was a scoping review and did not contain information directly identifying people.

Study Selection

The participant, concept, context (PCC) framework for scoping reviews was used to define the review focus.¹⁹ Our population of interest was any infant who had a GMA within the fidgety age range (9-20 weeks), which included use of the MOS. We included studies that used the original or revised versions of the MOS applied to the scoring of their fidgety movements.^5,16 We included a range of study designs (controlled trials, cohort studies, and case studies) from all countries.

Studies were excluded if (a) they only reported on the global scoring categories for fidgety general movements; (b) optimality scoring was only used to assess writhing movements at preterm/term age using the General Movements Optimality Scoring (GMOS); or (c) study was only presented as an abstract/poster. Supplemental Digital Content 1 (available at: https://links.lww.com/PPT/A417) lists excluded studies.

Data Extraction

Two independent reviewers reviewed the titles and abstracts of the studies identified. Studies that could not be excluded based on the title/abstract were reviewed in full by 2 independent reviewers according to the inclusion criteria. Differences of opinion regarding eligibility of a study were discussed until agreement was reached. A study flow diagram, as per the PRISMA statement,²¹ illustrates the process and results of the search (Figure 1).

Fig. 1.: PRISMA flow diagram.

Data were independently extracted for included studies by 2 authors into a predetermined data extraction table. Data were extracted on participant characteristics, location, study design, level of evidence, aims, and main findings of each study. Key concepts on how the MOS was used in each study were mapped and charted according to the scoping review methodology outlined by the JBI.¹⁹ One author completed this process to group studies into themes based on the primary use of the MOS, and then findings were discussed among all authors until agreement was reached.

Data were also specifically extracted for overall MOS as well as scores on each individual domain of the MOS. This information was visually mapped to identify themes and trends based on infant population groups, and these findings were discussed among authors until agreement was reached. Consideration was given to the number of participants and quality of the study when analyzing results.

Risk of Bias and Quality Assessment

Although formal assessment of methodological quality of included studies is not generally performed in a systematic scoping review, we assessed a subset of studies for risk of bias to address 1 of our specific review questions. These studies specifically reported the MOS as a predictive tool (n = 12) in association with outcomes (eg, mobility, cognition, language, or gross motor severity of CP) using the Gross Motor Function Classification System (GMFCS) level. We assessed risk of bias using the Quality in Prognosis Studies (QUIPS) (see Supplemental Digital Content 2, available at: https://links.lww.com/PPT/A418).²²

RESULTS

The search yielded 83 unique articles of which 37 articles were included (comprising a total of 3662 infants) from 20 countries in 6 continents (Figure 1).

Studies were conceptually grouped according to use of the MOS or MOS-R in each study; for prediction, as an outcome measure, as a descriptive tool, or exploring the psychometric properties of the MOS (Figure 2). In this article, the term “MOS” reflects all versions used, unless versions are specifically identified. Table 1 provides an overview of included studies.

Fig. 2.: Key concepts.

TABLE 1 - Characteristics of Included Studies (Grouped According to Use of MOS)

Author, Country	Study Design	Level of Evidence	Target Population	Participants (n, Gestational Age, wk)	MOS Version and Outcome Assessments	Primary Aim	Main Findings
PREDICTIVE
Peyton et al,14 USA	Prospective cohort	Level 3	Infants <31 wk and <1500 g who required oxygen at birth	n = 123 mean 27.1 (SD 2.0)	MOS Outcome assessments: Bayley III	To investigate relationship between resting state functional brain connectivity, general movements, and developmental outcomes	MOS (separate from fidgety movements alone) correlates with cognitive, language, and motor outcomes There was a relationship between MOS and functional brain connectivity
Einspieler et al,16 global	Retrospective cohort	Level 3	Cerebral palsy	n = 468 Median 34 (23-42)	MOS-R Outcome assessments: GMFCS	To identify early specific markers for gross motor function, topography, and type of cerebral palsy	GMFCS strongly associated with MOS-R MOS-R of >14 strongly associated with GMFCS I or II; GMFCS IV or V hardly ever associated with MOS >8 Asymmetrical segmental movements strongly associated with unilateral CP Circular arm movements strongly associated with dyskinetic CP
Einspieler et al,30 Brazil	Prospective case-control	Level 3	Infants vertically exposed to Zika and control infants	Total N = 444 n = 111 exposed to Zika (35 with microcephaly) n = 333 controls	MOS Outcome assessments: Bayley III	To describe MOS in Zika-exposed cohort Associations between MOS, clinical risk factors, microcephaly, and outcome at 12 mo	All infants with microcephaly had absent fidgety GMs and low MOS 70% of infants of Zika-positive mothers but no microcephaly had suboptimal MOS MOS differentiated outcomes
Salavatir et al,17 Austria	Prospective cohort	Level 4	Singleton; term (38-41 wk); residing close to facility; German-/Austrian-speaking family	n =22 Mean 39.9 (SD 0.8)	MOS Outcome assessments: At 4 y: AWST-R, NVT, PPVT (3rd ed) At 10 y: HAWIK, WISC, TROG	To determine the association between the early motor repertoire and language development	Smooth and fluent movement character at 3 mo was associated with better language performance at 4 and 10 y of age
Adde et al,26 India	Prospective longitudinal cohort	Level 3	LBW infants ≤1500 g	n =243 Median 31 (26-39)	MOS Outcome assessments: PDMS-II	Associations between general movements, MOS, and developmental outcome	Abnormal quality of the concurrent motor repertoire, and reduced or absent age-adequacy of the concurrent motor repertoire were significantly associated with lower gross and total motor scores at 12 mo post term
Zang et al,27 China	Prospective cohort	Level 3	Infants <1500 g Excluded brain malformations, chromosomal defects, syndromes	n =74 Mean 29 (SD 2)	MOS Outcome measure: PDMS-II	To explore how motor repertoire relates to neonatal complications, and motor skills at 12 mo of age	Motor optimality was lower in infants with pneumonia and/or bronchopulmonary dysplasia Lower MOS associated with poor fine and gross motor performance at 12 mo
Hitzert et al,29 the Netherlands	Retrospective analysis of a cohort of infants from a larger randomized controlled trial	Level 3	Healthy Dutch term-born infants	n = 74 Mean 40.1 (range 38.0-42.6)	MOSa Outcome assessments: WPPSI-R-NL, TEACh-NL, AVLT, NEPSY-II, M-ABC, DCD-Q, CBCL, ADHD questionnaire	To determine whether MOS is associated with cognitive, motor, and behavioral outcomes at early school age	MOS at 3 mo of age is associated with cognition and behavior at school age, but not motor
Yang et al,24 China	Retrospective cohort	Level 3	Infants with CP at 2 who had motor performance videoed at 4 mo	n = 79	MOS Outcome assessments: GMFCS	Does MOS correlate with GMFCS and type CP and is this different in term vs preterm?	Moderate correlation between MOS and better mobility Individual domains of MOS had correlations with GMFCS in term babies but not preterm
Yuge et al,28 Japan	Prospective cohort	Level 4b	Neurological concerns	n = 41 Median 37 (23-42)	MOS Outcome assessments: Kyoto scale of psychological development, Enjoji's analytical developmental test	To examine to what extent MOS can predict adverse neurological outcomes at 5 y	MOS predicted CP (100% accuracy) but not associated with DCD, developmental delays, PDD, or ADHD
Brugginki et al,23 the Netherlands and Italy	Longitudinal cohort	Level 4b	Cerebral palsy, born preterm (<34 wk)	n = 37 Mean 29.1 (SD 1.9)	MOS Outcome assessments: GMFCS	To determine predictive value of early motor repertoire for level of self-mobility in children with CP at school age	Higher MOS score associated with better gross motor function
Bruggink et al,15 the Netherlands	Prospective Cohort	Level 3	Preterm <34 wk	n = 82	MOS Outcome assessments: Touwen	To determine whether quantitative aspects of general movements are predictive of minor neurological dysfunction (MND) at school age	Infants later classified as complex MND showed fewer movement patterns at 11-16 wk than infants later classified as normal. Presence of an obligatory ATN posture in infants with normal FMs increased risk of complex MND to 75%.
Butcher et al,13 the Netherlands	Retrospective cohort	Level 3	Infants <33 wk	n = 65 Mean 29.9 (SD 1.83), range 25.7-33.7	MOS Outcome assessments: Touwens Test; WISC III; Child Behaviour Checklist	Whether MOS are predictive of intellectual and behavioral problems at 7-11 y	Positive association between normal postural patterns and intelligence (WISC) trend toward lower overall MOS being associated with poor attention
MOS AS OUTCOME MEASURE
Fjortoft et al,33 Norway	Prospective controlled cohort study	Level 3	Infants exposed to alcohol or drugs prenatally + controls	Total N = 214 Rx group n = 108 Mean 38.8 (SD 2.0) Control n = 106 Mean 39.7 (SD 0.9)	MOS	To determine whether infants suffering from prenatal exposure to addictive drugs and alcohol develop abnormal motor behavior at 3-4 mo (using GMs, MOS, and AIMS)	Abnormal movement character of infants exposed to alcohol and/or addictive drugs in pregnancy 68 (63%) infants in the study group displayed an abnormal movement character, compared with 23 (22%) controls. On AIMS—46 (44%) study group scored <10th percentile; 2 (3%) controls (P = .001)
Katušići et al,39 Croatia	Longitudinal cohort	Level 4b	Preterm <32 wk	n = 30 Median 28 (IQR 26 + 4 to 30 + 4)	MOS-R Outcome assessment: infant motor profile	Primary—to explore relationship between MRI as a predictor of GMs (including MOS) and motor development at 12 mo of age	The more visible the sagittal strata, the higher the MOS
Mebius et al,38 the Netherlands	Prospective observational	Level 4b	Infants with congenital heart disease	n = 36 Median 39.1 (range 36.1-40.3)	MOSa	To assess associations between prenatal Dopplers, postnatal oxygen saturations, and GMs/MOS	MOS moderately associated with blood flow in the MCA prenatally and negatively correlated with umbilical artery flow
Kepenek-Varol et al,40 Turkey	Prospective cohort	Level 4b	Preterm <37 wk	n = 32 Mean 32.3 (SD 3.1)	MOS	Effect of 1 session of physiotherapy on MOS	No difference in MOS after physiotherapy session
Fjortoft et al,41 Norway	Randomized controlled trial	Level 2	Infants ≥32 wk	n = 130 Mean 29.7 (SD 2.2)	MOS Outcome assessment: AIMS	To investigate the effectiveness of a parent intervention	No significant between group differences in GMs, movement character, or MOS
Tanis et al,37 the Netherlands	Prospective cohort	Level 4b	Growth restriction	n = 48 (43 assessed at 3 mo) Median 35 wk (range 26-40)	MOSa	To investigate whether Doppler pulsatility in infants with growth restriction are associated with GMs	No association between fetal Doppler parameters in infants who were growth restricted in utero and GMs at 3 mo
Hitzert et al,32 the Netherlands	Prospective observational	Level 4b	Preterm <32 wk	n = 17 Median GA 26.7 (range 25.0-29.7)	MOSa	To assess the effect of low-dose dexamethasone (DXM) treatment on the quality of GMs and fidgety GMs	Low-dose DXM infants had higher MOS at 3 mo than high-dose DXM infants. MOS at 3 mo did not differ between low-dose DXM infants and hydrocortisone infants
Luxwolda et al,36 Tanzania and the Netherlands	Prospective observational	Level 3	Breastfed infants	n = 112 Tanzania, n = 97 The Netherlands, n = 15 GA not reported	MOSa	Associations between MOS scores, docosahexaenoic acid (DHA) levels, and ethnicity	DHA status related positively to the observed movement patterns (OMP) score. Higher OMP score may discriminate between sufficient and optimal motor development.
Berghuis et al,35 the Netherlands	Prospective observational	Level 3	Healthy term infants	n = 97 Mean 40 (SD 1)	MOSa	To determine whether prenatal background exposure to PCBs (polychlorinated biphenyls) and OH-PCBs (hydroxylated metabolites of PCBs) was associated with the motor development of 3-mo-old infants	Prenatal exposure to high levels of OH-PCB-107 associated with lower MOS (MOS < 26) (P = .04)
De Vries et al,34 the Netherlands	Prospective observational	Level 3	Normal infants; mothers with or without SSRI treatment for depression	SSRI group, n = 63 GA 39.1 (36.7-42.7) Control group, n = 44, GA 40 (37.3-41.7)	MOS	To determine the effects of prenatal SSRI exposure on early neurological functioning of infants	Lower MOS and abnormal GMs more likely in SSRI-exposed group. Taking SSRIs in pregnancy had an adverse effect on early neurological functioning.
Hitzert et al,31 the Netherlands	Longitudinal, observational case-control	Level 3	<32 wk, in NICU and needing corticosteroid therapy due to BPD	Total N = 56 HC group, n = 17 Mean 27.1 (SD 24.9-31) DXM group, n = 17 Mean 27.9 (SD 26-30.3) Controls, n = 22 Mean GA 27.2 (SD 26-29.6)	MOSa	To determine the effects of hydrocortisone (HC) and dexamethasone (DXM) on movement quality	Neurological functioning at 3 mo is better in infants treated with HC than in infants treated with DXM. At 3 mo, HC infants had a higher median motor optimality score (MOS) than DXM infants (P = .015).
DESCRIPTIVE
Alkan et al,25 Turkey	Retrospective cohort	Level 3	HIE and healthy controls	n = 76 HIE group Mean 38.68 (SD 1.98) Control group Mean 37.02 (SD 3.34)	MOS	To describe and compare motor repertoire of infants with HIE of varying severity compared with healthy infants	Infants with HIE had poorer MOS compared with healthy infants. There were significant differences in age-adequate motor repertoire and movement patterns. Severity of HIE (grade I, II or III) was correlated with overall MOS.
Lokmanoğlu et al,54 Turkey	Case study	Level 5	Infant with cri du chat	n = 1 40 wk	MOS-R	To describe spontaneous movements in cri du chat syndrome, and the effect early intervention has on an infant with cri du chat	This case study of infant with cri du chat showed very poor MOS score. This infant also had very poor developmental outcomes at 12 mo of age.
Örtqvist et al,46 Austria	Retrospective observational	Level 4	Extremely preterm (<27 wk) compared with term infants	Total N = 106 Preterm group, n = 53 Mean 25 wk Term group = 53 Mean 40 wk	MOS-R	To compare movements and postures of extremely preterm infants to term-born controls	Significant difference in MOS between preterm (median MOS 18), compared with term (median MOS 26)
Salavati et al,45 the Netherlands	Longitudinal observational	Level 3	Preterm infants >30-wk gestation, compared with term infants 37- to 42-wk gestation	n = 360 Mean preterm cohort 27.9 (SD 1.5) Mean term cohort 40.1 (SD 1.2)	MOS-R	To gather reference data on the MOS-R in preterm and term populations	Median motor optimality scores in preterm infants lower (24) than term infants (26). Compared with term infants, preterm infants had poorer scores on the subscales age-adequate movement repertoire, observed postural patterns and movement character. There were no differences in observed movement pattern scores.
Kahraman et al,50 Turkey	Prospective for obstetric brachial plexus lesion (OBPL) and retrospective for controls	Level 4	Infants with OBPL diagnosis and no other major medical problems, as well as the control group	Total N = 40 OBPL, n= 20 Median 39.5 (36-40) Controls, n= 20 Median 38.5 (37-40)	MOS	To define the movement characteristics of infants with OBPL and compare to infants with normal neurological outcome	OBPL did not affect the overall quality of fidgety movements. Infants with OBPL had slightly lower scores for posture and movement quality. Reliability: Excellent interscorer agreement on MOS. Overall intraclass correlation coefficient 0.948 (CI 0.912-0.0971), P < .01
Rodijk et al,49 the Netherlands	Prospective cohort study	Level 4	Infants with biliary atresia	n = 35 Median 40 (36-42)	MOS-R	To assess whether neurological impairments in infants with biliary atresia are already present in early infancy before surgery	Early motor repertoire in biliary atresia is already affected at time of diagnosis. Significantly more infants with BA scored atypical on the GMOS (56% vs 10%, P < .001) and MOS (<25) (37% vs 18%, P = .04) compared with the appropriate reference groups in literature.
Sharp et al,42 Australia	Observational cohort	Level 4	Infants ≤25 wk admitted to NICU	n = 40 Median 24 wk, 3 d (23-25 wk, 6 d)	MOS	To describe the GMs and MOS in extremely preterm infants	MOS is lower in preterm infants than for term infants, even with normal fidgety
Herrero et al,47 Brazil, Turkey, China, the Czech Republic, and Austria	Exploratory cohort	Level 4b	Down syndrome	n = 47 Median 37 (29-41)	MOS	To describe MOS in infants with Down syndrome and assess relationship between pre- and perinatal risk factors and motor performance	A wide variability of MOS in Down syndrome cohort (median MOS in was between CP and normal). Neither preterm birth or any perinatal risk factor was related to fidgety movements or reduced MOS.
Fjortoft et al,43 Norway	Prospective, cohort with comparison group	Level 3	Infants <28 wk and/or <1000 g; healthy controls	Total N = 169 EPT and/or ELBW, n = 82 Mean 26.6 (1.8) Control, n = 87 Mean 39.6 (1.0)	MOSa	To compare MOS between EPT/ELBW and healthy controls	MOS lower in the EPT/ELBW group. Preterm infants more likely to have abnormal movements (not correlated with degree of prematurity)
Einspieler et al,53 Japan	Case study	Level 5	Smith-Magenis syndrome	n = 1 40 + 3 wk	MOS	To describe the motor behavior of a 4-mo old with Smith-Magenis syndrome	Fidgety movements absent, MOS was 10
De Vries and Bos,44 the Netherlands	Prospective cohort	Level 4b	ELBW <1000 g	n = 13 Mean 27.9 (SD 2.9)	MOSa	To assess motor repertoire at term age and at 3 mo in relation to neurological outcome at 2 y	At 3 mo, general movements are mostly normal in extremely low-birth-weight infants, but concurrent movements are not. MOS range is 11-28. None of the infants were diagnosed with CP.
Phagava et al,48 Italy	Retrospective cohort with comparison group	Level 4	Infants with ASD; matched controls with videos	Total N = 40 20 ASD 20 control	MOS	To detect whether abnormalities in movement can be observed in the first months of life in children with ASD	MOS significantly lower in the ASD group (P < .001). Values not provided
PSYCHOMETRIC PROPERTIES
Fjortoft et al,51 Norway	Cross-sectional	N/A	Born at Trondheim University Hospital between 1999 and 2005	n = 24 24-28 wk in 13 infants; 29-33 wk in 5 infants; 6 born at term	MOS	To examine the interrater reliability of GMs	High between all 4 observers. ICC was 0.87, and ICCs(3,1) for the pairwise analyses ranged between 0.80 and 0.94. High agreement for fidgety movements, but other categories ranged from moderate to high.
Kuo-Kuang et al,52 Taiwan	Cross-sectional (longitudinal follow-up)	N/A	PMA 28-60 wk	n = 37 Mean 32.2 (6.5)	MOSa	To examine the intraobserver reliability for global GMA, the GMOS, and the MOS	Moderate to good, ICC 0.718-0.828 for 5 items, and total score was good (ICC3,1 = 0.900). Videos of infants in the PMA 49- to 52-wk group had less reliability than those from other age groups (PMA 52-56 wk and PMA 56-60 wk)

Abbreviations: ADHD, attention deficit hyperactivity disorder; AIMS, Alberta Infant Motor Scale; ASD, autism spectrum disorder; ATN, asymmetric tonic neck; AVLT, Rey's Auditory Verbal Learning Test; AWST-R, Aktiver Wortschatztest–Revision; Bayley III, Bayley Scales of Infant and Toddler Development–Version III; BPD, bronchopulmonary dysplasia; CBCL, Child Behaviour Checklist; CP, cerebral palsy; DCD, Developmental Coordination Disorder; DCD-Q, Developmental Coordination Disorder Questionnaire; ELBW, extremely low birth weight; EPT, extremely preterm; FM, fidgety movement; GA, gestational age; GM, general movement; GMFCS, Gross Motor Functional Classification System; GMOS, General Movements Motor Optimality Score Writhing Age; HAWIK, Hamburg-Eschsler-Intelligence Test, 4th Edition; HIE, hypoxic-ischemic encephalopathy; ICC, intraclass correlation coefficient; IQR, interquartile range; LBW, low birth weight; M-ABC, Movement Assessment Battery for Children; MND, minor neurological dysfunction; MOS, General Movements Motor Optimality Score (fidgety age); MOS-R, General Movements Motor Optimality Score (fidgety age)–Revised Version; N/A, not applicable; NEPSY-II, Developmental Neuropsychological Assessment, 2nd Edition; NICU, neonatal intensive care unit; NVT, Noun-Verb Test; PDD, Pervasive Developmental Disorders; PDMS-II, Peabody Developmental Motor Scales–Version II; PMA, postmenstrual age; PPVT, Peabody Picture Vocabulary Test, Third Edition; SSRI, selective serotonin reuptake inhibitor; TEACh-NL, Test of Every Attention for Children; TROG, Test for Reception Grammar–Deutsch; WISC, Wechsler Intelligence Scale for Children; WPPSI-R-NL, shortened form of the Wechsler Preschool and Primary Scale of Intelligence, revised (Dutch version).

^aStudy applies scoring criteria for age adequacy as per Bruggink et al.¹⁵

^bGraded down due to imprecision (small sample size).

MOS Used for Prediction

There were 12 articles that used the MOS as a predictor of neurodevelopmental outcome including 3 which described use of the MOS with a cohort of infants later diagnosed with CP.^16,23,24 Across these studies the total MOS was consistently negatively correlated with the GMFCS level, indicating that a lower MOS is predictive of lower gross motor function. The median MOS for infants later diagnosed with CP, GMFCS I and II, was 12 (range 9-22) as compared with the MOS of 9 (range 7-9) for infants with GMFCS III to V.²³ Another study reported the median MOS in the CP population as 6 (interquartile range 6-9).²⁴ A study of infants with hypoxic-ischemic encephalopathy (HIE)²⁵ reported a significant correlation between the MOS and severity of HIE (Sarnat 1 vs Sarnat 3) and that across the entire HIE cohort the median MOS was 22, which was significantly lower than the median MOS of 25.5 in the neurotypical comparison group.

The other prediction studies reported on the association of the MOS with general developmental outcomes (cognitive, motor, language, and behavior) or minor neurological dysfunction^{13–15,17,26–28} Although the MOS predicted CP with 100% accuracy, it was not diagnostically associated with developmental delay, developmental coordination disorder, pervasive developmental disability, or attention deficit hyperactivity disorder (ADHD). However, the subsample for conditions such as ADHD was very small; thus, it is difficult to draw conclusions about prediction of these disorders.²⁸

In healthy, low-risk, term-born infants, the MOS was associated with cognition and behavior, but not motor performance at early school age.²⁹ The MOS differentiated developmental outcomes in infants exposed to the Zika virus with some postural patterns with statistical significance (eg, finger spreading and asymmetrical body posture).³⁰

In infants born preterm and/or low birth weight, there was a positive association between early quality of movement and later intelligence, especially normal postural patterns¹³ and a lower MOS was associated with poor fine and gross motor performance at 12 months of age.²⁷

The quality of studies reporting on the MOS as a predictive tool ranged from low to moderate quality as rated on the QUIPS (see Supplementary Digital Content 2, available at: https://links.lww.com/PPT/A418).²² Studies were rated lower for reasons such as the sampling method used (eg, convenience sampling), or if they had not considered potential confounders (ie, early intervention) in analysis.

MOS Used as an Outcome Measure

There were 11 articles with the MOS used as an outcome measure or to describe/compare the results of a treatment or intervention in the antenatal or neonatal period. Two of these studies reported on the effects of corticosteroids on movement quality in preterm infants with bronchopulmonary dysplasia.^31,32 Further studies reported the MOS in association with prenatal events. These included maternal exposure to alcohol/drugs,^33,34 environmental pollutants,³⁵ or maternal nutrition intake/diet,³⁶ and investigations into whether prenatal Doppler measurements^37,38 or MRI³⁹ in at risk infants was associated with the MOS at 3 months of age. The final 2 studies examined the MOS following interventions targeting infants born prematurely such as physical therapy and parent interventions in the neonatal intensive care unit.^40,41

MOS Used Descriptively With a Specified Cohort

In 12 studies, the MOS was used to describe the movement quality in a specific diagnostic group, often in comparison to a control group. Five of these studies described the MOS in extremely preterm infants or infants born with extremely low birth weight.^42–46 These studies reported a lower MOS in this population in comparison to term-born infants; however, median and range of scores were inconsistent with variation between studies.

The MOS in infants with trisomy 21 was significantly lower (median 13, range 10-28) than in infants with normal neurological outcome (median 26, range 10-28), although higher than the MOS in infants with CP (median 6, range 5-20).⁴⁷ Similarly, the MOS in infants with ASD was significantly lower than in controls.⁴⁸

The MOS was described as atypical in 37% of infants with biliary atresia (MOS <25) compared with 18% of peers,⁴⁹ and infants with obstetric brachial plexus lesion had the lower mean MOS (24.5, SD 2.21) compared with peers (25.5, SD 2.59).⁵⁰

Psychometric Properties

There were 2 studies that examined the reliability of the MOS as the main aim of the article.^51,52 The study by Fjortoft and colleagues⁵¹ reported satisfactory interrater reliability in cases of the high and low total MOS with some inconsistency in the middle range of the scale. Another reliability study reported moderate to good intrarater reliability of the MOS, with better reliability in videos of older infants (12-16 weeks), compared with younger infants (9-12 weeks).⁵²

Although not the focus of the study, other articles presented data on interrater reliability. Sharp and colleagues⁴² reported agreement between 2 scorers in 34 out of 38 cases (90%), with the additional 4 cases (10%) differing by a score of 1. Specifically looking at the age adequacy of the concurrent movement repertoire, high interrater reliability was reported in a study of 82 infants who were preterm with raters agreeing measured by Cohen's κ of 0.89 in 145 randomly selected recordings.²³ For the presence of normal/atypical postural patterns, observers disagreed in 10% of cases. Another study of infants prenatally exposed to the Zika virus reported high interrater agreement ranging from 0.88 to 0.96 using Cohen's or Fleiss' κ, and intraclass correlation coefficients exceeding 0.90.³⁰ In a single case study of an infant with a genetic syndrome, high interrater agreement for assessment of posture was reported as Fleiss' κ of 0.82.⁵³

When combining results for both the writhing and fidgety general movements, interrater agreement when reporting on optimality was reported as: Cohen's κ = 0.66 (moderate) when allowing for a 1-point MOS difference; Cohen's κ = 0.84 (high) when allowing for a 2-point difference (Table 2).³²

TABLE 2 - MOS Domains and Scores: Results

Study	MOS Score	Observed Movement Patterns	Age-Adequate Repertoire	Observed Postural Patterns	Movement Character
Adde et al26 Participants: Infants with extremely low birth weight	NR	No significant associations with outcomes at 12 mo	Reduced or absent age-adequate repertoire significantly associated with lower gross and total motor scores at 12 mo on the PDMS (P = .025)	No significant associations with outcome at 12 mo	Normal quality of concurrent motor repertoire associated with higher fine motor scores on PDMS (P = .02)
Alkan et al25 Participants: HIE and healthy controls	MOS scores significantly different between groups (P < .01) Median MOS in HIE cohort = 22; healthy cohort 25.5 HIE grade I median MOS = 24 HIE grade II median MOS = 20 HIE grade III median MOS = 8	Significant difference between groups in quality of movement patterns (P < .006) HIE group: N > A: 32 (84.2%) N = A: 3 (7.9%) N < A: 3 (7.9%) HIE grade I median = 4 HIE grade II median =4 HIE grade III median = 2 Healthy infants: N > A: 25 (65.8%) N = A: 13 (34.2%) N < A: 0	Significant difference in age-adequate repertoire between groups (P < .001) HIE group: Adequate: 28.9% Reduced: 34.2% Absent: 36.8% Healthy infants: Adequate: 36 (94.7%) Reduced: 1 (2.6%) Absent: 1 (2.6%)	No significant differences between groups Postures in the HIE group: N > A: 14 (36.8%) N = A: 9 (23.7%) N < A: 15 (39.5%) HIE grade I median = 4 HIE grade II median =1.5 HIE grade III median = 1 Healthy infants: N > A: 23 (60.5%) N = A: 7 (18.4%) N < A: 8 (21.1%)	No significant differences between groups Quality HIE group: Smooth/Fluent: 9 (23.7%) Abnormal: 29 (76.3%) CS: 0 HIE grade I median = 2 HIE grade II median = 2 HIE grade III median = 2 Healthy infants Smooth/fluent: 12 (31.6%) Abnormal: 26 (68.4%) CS: 0
Berghuis et al35 Participants: Healthy term	MOS range 18-28 MOS < 26, n = 17 MOS ≥ 26, n = 80 Prenatal exposure to OH-PCB-107 associated with lower overall MOS (<26) (P = .04)	Absent antigravity movements more common in those with prenatal exposure to PCB-118, OR 1.84 (1.02-3.32) and absent manipulation more common in those with prenatal exposure to OH-PCB-172, OR 0.19 (0.05-0.79) Normal n = 95 Reduced n = 2	No significant associations Normal n = 77 Reduced n = 20	No associations reported Normal n = 96 Reduced n = 1	High levels of PCB-118 associated with CS character (P = .05) Smooth/fluent n = 50 Abnormal n = 47
Bruggink et al23 Participants: Cerebral palsy	GMFCS III-V: Median MOS = 9 (range 7-15) GMFCS I-II: Median MOS = 12 (range 9-22) MOS ≤9 for GMFCS level III or IV Sensitivity 86 (72-100) Specificity 47 (22-72) PPV 70 (53-87) NPV 70 (42-98)	No association between number of normal/atypical movement patterns and GMFCS Kicking pattern (monotonous, repetitive) associated with GMFCS level (more severe functional limitations)	Absence of age-adequate repertoire associated with GMFCS level	Non-flat posture associated with GMFCS level Sensitivity 68 (49-87) Specificity 73 (51-95) PPV 79 (61-97) NPV 61 (38-84)	Cramped character associated with GMFCS level Sensitivity 64 (44-84) Specificity 80 (60-100) PPV 82 (64-100) NPV 60 (39-81)
Bruggink et al15 Participants: Preterm <34 wk	11-16 wk most reliable period for MOS as predictor of MND vs normal outcome MOS score differentiated between CP and normal outcome at all fidgety periods	Complex MND: Fewer movement patterns at 11-16 wk than infants developing typically	Reduced or absent age-adequate repertoire in 11 of the 13 infants who developed CP Reduced age adequacy strongly associated with neurological outcome at school age 70% infants with reduced age adequacy had abnormal outcome (35% CP; 35% complex MND)	Presence of ATN at 11-16 wk predictive of abnormal neurological outcome (35% of MND infants; 7% of normal) Normal infants: More normal postural patterns at all age periods, and fewer abnormal postural patterns at 11-16 and 17-24 wk than infants who developed CP Number of normal and atypical postural patterns differentiated normal from complex MND (χ2 test for trend = 4.5, P < .05)	A predictor in LR models of infants with normal FMs
Butcher et al13 Participants: Preterm infants <33 wk	Strong relationship between MOS score and intelligence in an LR model mean MOS 15.7 (SD 3.6) Range 9-20 (note that 20 highest possible score in this MOS scoring version)	No significant associations	No significant associations	Association between number of normal postural patterns and intelligence at 7-11 y: TIQ (b = 0.35, P = .11), VIQ (b = 0.34, P = .015)	No significant associations
De Vries et al44 Participants: Infants with extremely low birth weight	MOS range 11-28 at 3 mo. Three infants (23%) had abnormal fidgety.	NR	NR	Atypical postures n = 7 (54%) infants—mostly because of flat body and limb position (38%) and/or presence of ATN reflex (23%)	Monotonous and jerky movements in n = 9 (69%) infants
De Vries et al34 Participants: Normal infants (mothers with depression)	Median MOS was lower in SSRI-exposed infants 26 (range 7-28) vs 28 (range 21-28)	NR	NR	NR	Exposed group had more monotonous movements (48% vs 20%); seen twice as often (OR 3.5, 95% CI, 1.5-8.6, P = .005)
Einspieler et al30 Participants: Infants with Zika virus and the control group	Between-group differences statistically significant (P < .01) Median MOS = 23 (IQR 21-26) in children with normal development, 12 (IQR 8-19) in children with adverse outcomes, and 5 (IQR 5-6) in children with microcephaly	Between-group differences were statistically significant (P > .001) Zika without microcephaly Normal = median 3 Atypical = median 1 Zika infants with microcephaly Normal = median 0 Atypical = median 4 Control group Normal = median 5 Atypical = median 0	Between-group differences were statistically significant (P > .001) Zika without microcephaly Age adequate = 46.1% Zika with microcephaly Age adequate = 2.9% Control group Age adequate = 73.9%	Between-group differences were statistically significant (P > .001) Zika with microcephaly Normal = median 3 Atypical = median 2 Zika with microcephaly Normal = median 1 Atypical = 6 Control group Normal = median 3 Atypical = median 1 Postural patterns that were statistically significant: Finger spreading, external rotation of hips, and asymmetry body	Between-group differences were statistically significant (P > .001) Zika without microcephaly Smooth/fluent = 15.8% Abnormal = 78.9% Cramped-sync = 0% Zika with microcephaly Smooth/fluent = 0% Abnormal = 100% Cramped-sync = 54.4% Control group Smooth/fluent = 44.7% Abnormal = 55.3% Cramped-sync = 0%
Einspieler et al16 Participants: Cerebral palsy	MOS in cohort of CP had a median of 7 (range 5-24) MOS and GMFCS had a strong negative correlation (ρ = −0.66) and stronger for term than preterm infants	Strong negative association between quantity of N movements and GMFCS (ρ = −0.57) Most commonly associated with more severe CP: Atypical mouth movements, atypical foot-to-foot contact, atypical arching, and atypical visual exploration Most commonly associated with unilateral CP: Normal hand regard, asymmetrical segmental movements Circular arm movements associated with more severe CP, bilateral CP, and dyskinetic CP	Only 5% had age-adequate repertoire, 9% had reduced, and 86% had absent repertoire	Negative association between quantity of normal postural patterns and GMFCS severity (ρ = −0.38) and positive for atypical and GMFCS (ρ = 0.41) Infants with unilateral CP had more N patterns than bilateral CP Atypical head centering and atypical variability of finger postures associated with more severe CP	Smooth and fluent very rare Cramped synchronized movements associated with severe CP
Einspieler et al53 Single-case study: Smith-Magenis syndrome	Smith-Magenis infant MOS 10	No hand-to-hand or fiddling. Abnormal foot-to-foot contact.	NR	Limited finger postures; head posture not in midline	Jerky and monotonous movements
Fjortoft et al43 Participants: Extremely preterm (EPT)/extremely low-birth-weight (ELBW) and term controls	Between-group differences statistically significant Median MOS 26 (IQR 23-28) in the ELBW/EPT group and 28 (IQR 28-28) in the control group (P = .001)	Between-group differences not significant Control group N > A 100% ELBW/EPT group N > A 96% N = A 1% N < A 2%	Age-adequate repertoire occurred significantly more often in the control group compared with the ELBW/EPT group (P = .006) Hand-to-hand contact seen twice as frequently in the control group Foot-to-foot contact also seen significantly more frequently in the control group	Control group had N > A postural patterns significantly more than 94% of control group had N > A postural patterns compared with 83% ELBW/EPT group (P = .039)	ELBW/EPT group more likely to have abnormal movement quality; OR 4.1 (95% CI, 2.0-8.7) 36 (44%) smooth and fluent movements and 41 (50%) monotonous Control infants: 70 (81%) smooth and fluent and 13 (15%) monotonous
Fjortoft et al 202033 Participants: Prenatal drug exposure and controls	Between-group differences statistically significant Prenatal drug exposure: Median MOS 26 (IQR 26-28) Control group: Median MOS 28 (IQR 28-28) (P = .001)	Hand-to-mouth contact more frequent in the control group than in the study group (81 [79%] vs 67[62%]; P < .008), and foot-to-foot manipulation more common in the control group (58 [57%] vs 44 [41%]; P < .019)	No significant difference between groups	Significantly more normal variability of finger postures in the control group (P < .001)	Smooth and fluent movement character was more than twice as high in the control group (81 [78%] versus 40 [37%]; P < .001) n = 68 (63%) infants in the study group had an abnormal movement character, compared with 23 (22%) controls (OR 6.0; 95% CI, 3.3-11.0)
Fjortoft et al41 Participants: Infants ≤32 wk	No significant between-group differences in GMs or MOS (P = .460)	NR	NR	NR	No significant between-group differences in movement character. Approximately half in each group had abnormal movement character.
Herrero et al47 Participants: Infants with Down syndrome	Median MOS 13 (range 10-28) 13 infants had absent fidgety movements, 20 abnormal, and 14 normal	N > A = 39 infants (83%) N = A = 5 infants (10.5%) N < A = 3 infants (6.5%) Average number normal movement patterns = 3 (range: 0-8); abnormal movement pattern = 1 (range: 0-3)	Age-adequate infants n = 12 (25.5%) Reduced infants n = 20 (42.5%), mainly due to a lack of movements to the midline Age-inadequate infants n=15 (32%)	N > A = 22 infants (47%) N = A = 10 infants (21%) N < A = 15 infants (32%) Lack of variability of finger postures (51%)	Smooth and fluent n = 3 infants (6.5%) Monotonous, stiff, jerky and/or tremulous n = 44 infants (93.5%)
		Most frequent normal movement patterns: Visual scanning (32/47; 68%), side-to-side movements of head (22/47; 47%), foot-to-foot contact (14/47; 30%), hand-to-mouth contact (12/47; 25.5%), and kicking (10/47; 21%) Patterns observed in <10 infants: Smiling (9/47; 19%), fiddling (9/47; 19%), hand regard (8/47; 17%), swipes (7/47; 15%), hand-to-hand contact (6/47; 13%), arching (6/47; 13%), and leg lifting (5/47; 11%)
Hitzert et al31 Participants: Preterm infants with bronchopulmonary dysplasia	Hydrocortisone infants had higher MOS (median 25) than dexamethasone infants (median 21) (P = .015).	NR	NR	NR	NR
Hitzert et al32 Participants: Preterm <32 wk	Low-dose dexamethasone infants had higher MOS at 3 mo than high-dose dexamethasone infants	NR	NR	NR	NR
Hitzert et al29 Participants: Healthy, term infants	Median MOS 26 (range 13-28) MOS associated with cognition and behavior, but not motor outcome	Presence of antigravity, midline leg, manipulation movements, was related to poorer cognition N > A = 73 N = A = 1 N < A = 0	Age-adequate repertoire related to poorer cognition, although within normal range Age adequate = 72% Reduced = 22% Severely reduced = 6%	N > A = 71 (96%) N = A =2 N < A = 1 Variable finger postures related to better cognition Absence of variable finger postures associated with borderline and abnormal visual-spatial perception (OR 20, 95% CI, 1.7-238; R 2 = 0.39; P = .018)	Smooth and fluent = 31 (42%) Abnormal = 43 (58%). Monotonous movements associated with better ball skills but more behavioral problems Cramped = 0
Kahraman et al50 Participants: Obstetric brachial plexus lesion and control	No significant between-group differences Infants with OBPL had slightly lower mean MOS (24.5, SD 2.21) compared with the control group (25.5, SD 2.59), P = .08	Infants with OBPL demonstrated significantly more excitement bursts (P = .03), head rotation (P = .007), hand-knee contact (P = .01), and rolling (P = .04)	No significant between-group differences OBPL mean 3.8 (SD 0.62) Control mean 3.4 (SD 0.94) P = .118	Trunk symmetry was scored as normal in a significantly greater number of infants with a normal outcome (P = .06). Variable finger postures were increased in the control group (P = .002) and infants with OBPL showed few finger postures (P = .001)	Significantly more jerky movement character observed in OBPL (P = .04)
Katušić et al39 Participants: preterm infants <32 wk	Median MOS 22 (IQR 19-26)	NR	NR	NR	NR
Kepenek-Varol et al40 Participants: Preterm infants	No significant between-group differences or correlations	No significant between-group differences or correlations	No significant between-group differences or correlations	No significant between-group differences or correlations	No significant between-group differences or correlations
Lokmanoğlu et al54 Single-case study: Infants with cri du chat	MOS = 6 in infant with cru du chat	Kicking, mouth movements and foot-to-foot contact were abnormal	Age-inadequate movement repertoire	Infants had head and body symmetry Abnormal variability of finger postures	Monotonous movement character
Luxwolda et al36 Participants: breastfed infants	NR	Significantly higher number of observed movement patterns (OMPs) in the group with higher levels of DHA (P = .04), indicating higher OMP score may discriminate between sufficient and optimal motor development	NR	NR	NR
Mebius et al38 Participants: Infants with congenital heart disease	MOS (median = 26, range 11-28) and a cut-off of 25 was used to define normal	NR	NR	NR	NR
Örtqvist et al46 Participants: Infants <27 wk and term controls	Preterm median MOS 18 (IQR 17-21) Term median MOS 26 (IQR 26-28)	Significant difference between groups (P < .001). Preterm: N > A: 25 (47%) N = A: 6 (11%) N < A: 22 (42%) Term: N > A: 53 (100%) N = A: 0 N < A: 0	Significant difference between groups (P < .001) Preterm: Present = 10 (19%) Reduced = 10 (19%) Absent = 33 (62%) Term: Present = 39 (74%) Reduced = 14 (26%) Absent = 0	Significant difference between groups (P < .001) Preterm: N > A: 6 (11%) N = A: 9 (17%) N < A: 38 (72%) Term: N > A: 47 (89%) N = A: 4 (8%) N < A: 2 (4%)	Significant difference between groups (P < .001) Preterm: Smooth/fluent = 0 Abnormal = 53 (100%) Cramped-sync = 0 Term: Smooth/fluent = 36 (68%) Abnormal = 17 (32%) Cramped-sync = 0
Peyton et al14 Participants: Infants born <31 wk and <1500 g who required oxygen at birth	MOS score has a relationship with cognitive, language, and motor outcomes even when fidgety movements component is not included in linear regression models	NR	NR	NR	Monotonous, stiff, jerky, or tremulous movement character significantly associated with specific alterations in functional connectivity
Phagava et al48 Participants: Autism spectrum disorder and controls	MOS significantly lower in the autism spectrum disorder group (P < .001). Values not provided	NR	NR	NR	Reduced MOS score in ASD the group due to lower quality of movement Values not provided
Rodijk et al49 Participants: Infants with biliary atresia and healthy controls	Significantly more infants with biliary atresia scored in the atypical range (P = .04) Controls: Median MOS 26 (21-28) Biliary atresia: MOS < 25, n = 7 (37%)	100% had N > A	15 (79%) had N > A	6 (32%) had reduced or absent age-related repertoire	7 (37%) smooth and fluent 12 (63%) abnormal all categories 8 (42%) monotonous
Salavati et al45 Participants: Preterm infants (<30 wk) and term infants (37-42 wk)	Median MOS in preterm cohort 24 (IQR 23-26) Median MOS in term cohort 26 (IQR 26-28)	No significant differences between groups Preterm: N > A = 174 (96.7%) N = A = 2 (1.1%) N < A = 4 (2.2%) Term: N > A = 176 (97.8%) N = A = 3 (1.7%) N < A = 1 (0.6%)	Significant differences between groups Preterm: Present = 84 (46.7%) Reduced = 73 (40.6%) Absent = 23 (12.8%) Term: Present = 130 (72.2%) Reduced = 36 (20%) Absent = 14 (7.8%)	Significant differences between groups Preterm: N > A = 119 (66.1%) N = A = 23 (12.8%) N > A = 38 (21.1%) Term: N > A = 161 (89.4%) N = A = 14 (7.8%) N > A =5 (2.8%)	Significant differences between groups Preterm: Smooth/fluent = 28 (15.6%) Abnormal = 152 (84.4%) Cramped-sync = 0 (0) Term: Smooth/fluent = 94 (52.2%) Abnormal = 86 (47.8%) Cramped-sync = 0 (0)
Salavati et al17 Participants: Term infants	In a multivariable analysis, a higher MOS at 5 mo was associated with better expressive language at 4 and 10 y of age (P = .007) Median MOS 3 mo = 26 Median MOS 5 mo = 24 At 3 mo, 10 infants obtained optimal scores and 9 infants were nonoptimal (<24)	NR	3 mo: Age adequate = 68% Reduced = 32% 5 mo: Age adequate = 28% Reduced = 72%	Higher scores at 5 mo were related to poorer results on one language subscale	A smooth and fluent movement character at 3 mo was associated with better language performance at 4 and 10 y of age (P = .005)
Sharp et al42 Participants: Preterm infants	Median MOS 22 (range 6-28). Infants with absent fidgety had MOS 6-11.	NR	Present in 26%	44% of infants had normal postural patterns for age	NR
Tanis et al37 Participants: Fetal growth restriction	No significant associations Median MOS 25 (IQR 23-26)	NR	NR	NR	NR
Yang et al24 Participants: Cerebral palsy both preterm and term	Moderate correlation between MOS and better mobility (ρ = 0.56) Individual domains of MOS had correlations with GMFCS in term babies but not preterm	Term-born infants had fewer normal movement patterns than those born preterm (P = .05) Quality of movements was negatively correlated with GMFCS level (ρ = −0.52; P = .001), indicating better functional mobility with a better quality of movement Repetitive opening and closing of the mouth was found significantly more often in children with lower GMFCS (P < .01)	Although no child had age-adequate repertoire, there was a significant association between the age adequacy of the motor repertoire at 9 to 20 wk post-term age and the GMFCS level later on (Kendall-Tau-c = −0.22; P < .001)	Weak negative correlation between postural patterns and later GMFCS levels (−0.19; Kendall-Tau-c; P < .05) Finger spreading and synchronized opening and closing of the fingers occurred more often in infants with later GMFCS levels IV and V (P < .01)	Moderate association between a cramped-synchronized movement character and the later GMFCS levels (P < .001) A monotonous but not cramped-synchronized movement character was also associated with GMFCS levels IV and V (P < .01)
Yuge et al28 Participants: Neurological concerns	MOS normal neurological outcome (median=24) was statistically higher than MOS with an adverse outcome (median = 14) (P ≤ .05) Median MOS 24 (IQR: P25 = 15; P75 = 26; range: 5-28)	Significant difference in number of normal movement patterns between those with normal and adverse outcomes (P < .05)	Significant difference in age adequacy between those with normal and adverse outcomes (P < .01) Reduced in CP, trisomy 21, severe developmental delay, Di-George and Smith-Magenis	Significant difference in normality of postural patterns between those with normal and adverse outcomes (P < .05) Abnormal posture in trisomy 21 (head and body not midline); and Smith-Magenis and CP	Significant difference in quality of movement character between those with normal and adverse outcomes (P < .01) Smooth and variable movement only found in infants with normal neurological development. CS found in CP. Monotonous and jerky found in trisomy 21, severe developmental delay, Di-George, developmental coordination disorder, attention deficit hyperactivity disorder, and Smith-Magenis (but also in normal)
Zang et al27 Participants: Infants with low birth weight	Lower MOS score associated with poor gross and fine motor performance at 12 mo (ρ = 0.406 for TMQ) Median MOS 24 (P25 = 22, P75 = 26; range 7-28)	NR	NR	NR	NR

Abbreviations: A, atypical; ASD, autism spectrum disorder; ATN, asymmetric tonic neck; CI, confidence interval; CP, cerebral palsy; CS, cramped synchronized; DHA, docosahexaenoic acid; FMs, fidgety movements; GM, general movements; GMFCS, Gross Motor Function Classification System; HIE, hypoxic-ischemic encephalopathy; ICC, intraclass correlation coefficient; IQR, interquartile range; LR, logistic regression; MND, minor neurological dysfunction; MOS, motor optimality score; N, normal; NR, not reported; NPV, negative predictive value; OBPL, obstetric brachial plexus lesion; OR, odds ratio; PDMS, Peabody Developmental Motor Score; PIQ, Performance Intelligence Quotient; PPV, positive predictive value; SSRI, selective serotonin reuptake inhibitors; TIQ, Total Intelligence Quotient; TMQ, Total Motor Quotient; VIQ, Verbal Intelligence Quotient.

MOS Domains

Further analyses considered significant findings in relation to each domain of the MOS. Articles reporting on this are included in Table 2 (the 2 reliability articles are not included in this table).

Overall MOS

There were 33 articles reporting the overall MOS (score from 5 to 28), with inconsistency in findings. In the cohorts of preterm and very low birth weight, the overall MOS ranged from 6 to 28.^{13,27,42–46} Studies indicate there is a correlation between the overall MOS and gross motor function in children with CP, and a trend toward children who have a diagnosis^{30,47,48,53,54} or difficulties later in life^13,17,29 having a lower MOS.

The presence of fidgety movements adds 12 points to the MOS and some studies have removed this section of scoring in their analysis^14,16 to determine whether the MOS accounts for a greater portion of the variance in long-term neurodevelopmental outcomes than by fidgety movements alone. In a large cohort of infants who went on to have CP, removing the fidgety movement score from the MOS did not add predictive value.¹⁶ In a small cohort of children born preterm (<32 weeks' gestation), removing the fidgety movement score from the MOS appeared to add predictive value in regard to cognitive and language scores on the Bayley III.¹⁴

Movement Patterns

There were 22 articles that reported on movement patterns. Two studies described movement patterns in infants born extremely preterm with 1 study⁴⁶ reporting 47% of infants demonstrating more normal (N) than atypical (A) movement patterns, in contrast to the other study⁴⁵ reporting 96.7% of infants with N greater than A.

Three studies reporting on movement patterns in infants born at term reported N greater than A (between 97.8% and 100%).^29,45,46 However, 1 study comparing healthy term infants to infants with HIE reported N greater than A in 65.8% of term infants and 84.2% in infants with HIE.²⁵ There were 7.9% of HIE infants, and no term infants, who had more atypical than normal movement patterns.

Movement patterns were reported in association with infant outcomes in 4 studies. The presence and normality of movement patterns was not associated with outcomes in infants born extremely low weight.²⁶ For infants with CP, the association between the number of normal/atypical movement patterns and the GMFCS level was inconsistent; no association using the MOS in 1 study,²³ compared with a strong negative association using the MOS in another.¹⁶ Infants exposed to the Zika virus had less normal movement patterns compared with controls.³⁰

The atypical movement patterns most commonly reported to be associated with CP were mouth movements, foot-to-foot contact, arching, and visual exploration.¹⁶ Monotonous kicking was associated with a lower GMFCS²³ and circular arm movements were associated with more severe CP (bilateral and dyskinetic CP).¹⁶

Infants who went on to have complex minor neurological dysfunction demonstrated fewer movement patterns than infants who were developing typically¹⁵ and certain movement patterns were less commonly seen in infants with trisomy 21 (eg, smiling, hand-to-hand, swipes, and leg lifting).⁴⁷ Hand-to-mouth movements and foot-to-foot postures were seen less commonly in infants following prenatal drug exposure.³³ Infants with biliary atresia were reported to always have a greater number of normal than atypical movement patterns.⁴⁹

Age-Adequate Repertoire

There were 22 articles that reported on age-adequate movement repertoire. There were 6 studies in term infants, 5 of which (n = 662) reported a normal age-adequate repertoire in 68% to 74% of infants,^{17,29,30,45,46} with 1 study (n = 38) reporting 94%.²⁵ In the preterm and very-low-birth-weight population, there were inconsistent findings across the 5 studies (n = 918) with a normal age-adequate repertoire ranging from 19%⁴⁶ to 95%.²⁶ Reduced or absent age-adequacy of the concurrent motor repertoire was reported to be significantly associated with lower gross motor outcomes at 12 months of age.²⁶

In 5 studies investigating children with CP, only 4% (28 out of 632) of infants had a normal age-adequate repertoire.^{15,16,23,24,30} In infants with syndromes and genetic conditions, the age-adequate repertoire was likely to be reduced.^28,47,54 In children with biliary atresia⁴⁹ or prenatal drug or alcohol exposure,³³ the age-adequate repertoire was more consistent with babies who were typical and born at term.

It is important to note that it is the age-adequate repertoire category that has been revised¹⁵ and therefore comparison and interpretation of these studies are difficult.

Postural Patterns

There were 24 studies that specifically reported on postural patterns. Three studies described postural patterns in the pre-term population (n = 506) and found that infants born preterm were more likely to have less normal and more atypical postural patterns than their peers born at term.^42,45,46 In infants with CP, postural patterns (specifically abnormal head centered and abnormal variability of finger movements) as well as the total number of normal postural patterns were indicative of severity and type of CP.¹⁶ Fewer normal postural patterns were observed in infants with CP compared to their peers who were neurotypical,²³ and infants who had more normal than atypical postural patterns had better intellectual ability at 7 to 11 years of age.¹³

Seven studies reported lack of variability of finger postures as a significant finding. In 1 study, variability of finger postures was associated with better cognition at 5 to 7 years of age.²⁹ Atypical finger postures were associated with more severe CP.¹⁶ Lack of variability of finger postures was observed in 51% of children with Down syndrome⁴⁷ as well as in single cases of cri du chat⁵⁴ and Smith-Magenis syndrome.⁵³ Infants who had prenatal exposure to addictive drugs and alcohol were more likely to lack variability of finger postures than controls.³³ Children who would go on to have minor neurological dysfunction were more likely to lack variability of finger postures, as well as have extended arms and a flat posture.¹⁵

Movement Character

Twenty-seven studies reported on movement character. A cramped synchronized movement character was correlated with increasing severity in studies of infants with CP^16,23,28 as well as Zika-exposed infants with microcephaly.³⁰ Moreover, the largest cohort of infants with CP yielded very few with smooth and fluent movement character.¹⁶

In cohorts of infants born prematurely, abnormal movement character was reported in at least 50% of the individual samples^43,44 with 2 studies reporting over 84% of infants having abnormal movement character.^45,46 In 1 study of infants born term, smooth and fluent movement character at 3 months was associated with better language performance at 4 and 10 years of age.¹⁷

Other high-risk cohorts, for example infants exposed to drugs³³ and an HIE cohort,²⁵ had higher proportions of infants with abnormal movement character, and the majority of infants (93.5%) with trisomy 21 had a monotonous, stiff, jerky, and/or tremulous movement character.⁴⁷ Not all infants who were neurotypical will have smooth and fluent movements, as this is considered to be the optimal condition, as opposed to normality.

DISCUSSION

This systematic scoping review explored the use of the GMA MOS at fidgety movement age (3-5 months), and the reliability and predictive value of the MOS in identifying adverse developmental outcomes. The MOS has been used to describe the movements of infants, as a predictive tool and as an outcome measure. Infants born preterm and infants diagnosed with CP are most highly represented in studies of the MOS, although the tool has been used with various other diagnostic groups. Reliability and validity of the MOS remain unclear, and future research will be important in validating the tool. The MOS offers some unique components that other assessments may not, and therefore has promising potential in the early identification of infants at risk of adverse neurodevelopmental outcomes.

Value of the MOS in Predicting Adverse Neurodevelopmental Outcomes

Assessment tools in infancy are primarily diagnostic in nature and it is not possible for 1 tool to perfectly predict neurodevelopmental outcomes.¹⁶ However, it is exciting to consider that at 3 to 5 months of age we may be able to identify children at high risk not only of CP, but other adverse neurodevelopmental outcomes. This early identification will importantly enable clinicians and families to provide targeted early intervention with the aim of maximizing long-term outcomes and ultimately maximizing the child's ability to participate in a fulfilling life.¹

Commonly used assessments for this age group measure skill development (eg, Test of Infant Motor Performance,⁵⁵ Alberta Infant Motor Scale,⁵⁶ Bayley Scales of Infant and Toddler Development,⁵⁷ or neurological function [Hammersmith Infant Neurological Examination]),⁵⁸ although the predictive validity of some of these assessments in the first 12 months remains unclear.⁵⁹ Most standardized assessments of infants at 3 to 5 months of age assess the infant in prone, supine, and upright positions and involve face-to-face interaction with the infant. The MOS does not elicit skills, and observation of social interactions and behavior may be important predictors of social outcomes. Although elicited skills are not captured by the MOS, the MOS does include items within the movement and postural pattern domains that are not assessed in other tests at this age. The MOS age-adequate repertoire (including foot-to-foot and hand-to-mouth contact, combined with number of normal movement patterns) and postural patterns (including variability of finger movement) are examples of items not measured by other assessments that may offer some predictive value.^{16,29,33,47,53,54} One major benefit of the MOS is that this assessment is scored based on the 3- to 5-minute GM video, which has become a commonly used assessment for infants with risk factors for neurodevelopmental impairments. Because this assessment is scored via a video, it has great potential as a screening tool for remote locations or during periods of social or pandemic restrictions. The use of video assessment only may also be a limitation, as it is only a small snapshot that may be significantly influenced by the infant's state at that time.

Reliability of the MOS-R

Until the publication of the MOS-R in 2019,¹⁶ there was no manual to guide scoring of this assessment, which may be 1 explanation for the variability of results in existing studies. Only 6 of the 37 studies included in this review used the MOS-R.^{16,39,45,46,49,54} There were inconsistent findings across studies of infants who were extremely preterm and low birth weight in regard to the overall MOS, as well as each MOS domain.^{13,27,42–46} In infants with a genetic condition or those who would go on to receive a neurodevelopmental diagnosis, there is a trend toward the overall MOS being lower.^28,47,53,54 In cohorts of high-risk infants (extremely preterm, low birth weight, and HIE), abnormal movement character is more prevalent (up to 100% in 1 preterm cohort); however, its association with later outcomes is not clearly established. In most studies that included controls, a proportion of infants demonstrated an abnormal movement character at fidgety age,^{25,33,43,45,46} and up to 25% of term infants did not have an age-adequate repertoire.^{17,29,30,45,46}

The MOS is a relatively new tool that appears to have promising predictive value, although there remains a lack of research regarding reliability and validity. Only 2 articles specifically reported on reliability, and both were on the earlier version of the MOS. Apart from the small amount of psychometric data presented here, we could not locate any studies that examined the content or construct validity of the assessment or test-retest reliability of any version of the MOS.

Implications for Research and Clinical Practice

Further research is needed to establish validity and reliability. Specifically, studies investigating content, construct, and concurrent validity, as well test-retest and intrarater reliability of the MOS-R, will be important to improve clinician confidence in using and interpreting results of the assessment. It is recommended that future studies clearly state in publications that MOS-R was the version used. It may be important for studies reporting on the MOS to clarify whether assessors have received direct training/certification in the MOS-R. Future studies aiming to report on the predictive value of the MOS-R should consider conducting analyses with and without the overall fidgety score.

The MOS-R provides clearer definitions of MOS items, and therefore will likely increase the reliability and consistency of scoring. To further improve reliable scoring of the MOS by clinicians and researchers, it may be important to include definitions of movement character within the MOS-R definitions and criteria for scoring. Whilst an MOS-R of equal to or more than 25 points has been reported to be considered optimal,^28,45 it would be beneficial to have consistent cut scores to guide clinicians in identifying those infants who may require closer monitoring or early intervention. This is an important aspect of clinical utility and will enable clinicians to more confidently identify infants at risk of adverse neurodevelopmental outcomes.

The MOS can only be currently learned via advanced GMA training and has not yet had widespread uptake in research or clinical practice. Once reliability of the MOS-R is further established, it may be important to consider MOS-specific training for those already advanced GM trained in the hope of this tool becoming more widely used in research and clinical practice.

Limitations

The MOS is a relatively new tool. Different versions have been used in practice and reported in research. Specifically, the age-adequate repertoire has been revised, making this category difficult to interpret and compare across studies. The MOS-R, with accompanying manual, will contribute to more consistency in the reporting of the MOS in future.

Heterogeneity of articles made it difficult to synthesize data on a quantitative level; however, this is not the focus of a systematic scoping review.

Not surprisingly, most of the research has involved authors who developed and modified the MOS and who most likely have higher levels of interrater agreement. The CP article from 2019¹⁶ is a notable exception where data were contributed from 24 different sites over multiple continents and yielded useful results for predicting CP severity. As wider groups of clinicians are trained on the MOS, more consistent data are expected across other high-risk infant groups.

CONCLUSION

The MOS-R has the potential to add value in identifying infants at risk of adverse neurodevelopmental outcomes, although the reliability and validity of this tool remains unclear. Future research will be vital in guiding the use of the MOS-R in the hope that this tool can assist with identifying high-risk infants, such that targeted early intervention can be provided.