Abstract
OBJECTIVE: To determine the stability of the scores obtained on tests of motor development from birth until school age in healthy, term singletons and to determine if early motor scores are associated with more complex cognitive functions at school age, such as attention and memory.
PATIENTS AND METHODS: This longitudinal, prospective cohort study included 77 infants. The motor development of these infants was assessed during the neonatal period with Prechtl's neurologic examination; in early infancy with Touwen's neurologic examination and general movement assessment; at toddler age with Hempel's neurologic examination and the Psychomotor Developmental Index from the Bayley Scales of Infant Development; and at school age with the Movement Assessment Battery for Children. Cognition was determined at toddler age with the Mental Developmental Index from the Bayley Scales of Infant Development; and at school age with an intelligence test and attention and memory tests.
RESULTS: The mean absolute difference in standardized motor scores for all time points was 1.01 SD (95% confidence interval: 0.91–1.11). Only the explained proportions of variance of maternal socioeconomic status and verbal intelligence were significant for sustained attention and verbal memory (r2 = 0.104, P = .030 and r2 = 0.074, P = .027), respectively. The children's scores on early motor tests added little value for their motor and cognitive development at school age.
CONCLUSIONS: In healthy children the stability of motor development from birth until school age is low. Maternal socioeconomic status and verbal intelligence rather than the infants' scores on early motor tests signified added value for complex cognitive functions at school age.
WHAT'S KNOWN ON THIS SUBJECT:
In high-risk infants, such as those with extremely low birth weight, early neurologic test scores have a strong predictive value for motor functioning later in life. Evidence has indicated that early motor impairment may also be related to later cognitive problems.
WHAT THIS STUDY ADDS:
In term-born children, stability of motor development from birth until school age was found to be low. Maternal SES and verbal intelligence rather than the infants' scores on early motor tests signified added value for complex cognitive functions at school age.
During the past decades there has been a growing interest in the prediction of developmental outcomes in children from early infancy to performance at school age, particularly in infants at high risk for developmental deficits. Gestational age, for example, is a strong predictor of later motor and cognitive development.1,2 In high-risk infants, such as those with extremely low birth weight, early neurologic tests scores have strong predictive value for function later in life. In other words, children with early developmental delays, such as motor impairments, have relatively stable developmental trajectories until school age. This finding indicates that early alterations in brain development have an important impact on the integrity and maturation of the central nervous system during childhood.
Less is known about the stability of developmental trajectories of healthy, term infants. Results of previous studies have suggested that in a healthy population the age at which children reach developmental milestones is predictive for development later in life.3,–,5 Other investigators have questioned whether a child's developmental trajectory can be readily predicted, because development is influenced by a number of biological and environmental factors, such as socioeconomic status (SES), as well as medical, social, and developmental interventions.6,–,8
To gain insight into the developmental trajectories of healthy children, we first aimed to determine the stability of scores on motor development tests from birth until school age in healthy, term singletons.
There is growing evidence that early motor impairment may be related to later cognitive problems. Therefore our second aim was to determine if the scores on tests of motor development in a population of healthy children were associated with performance at school age, particularly complex cognitive functions such as attention and verbal memory.
PATIENTS AND METHODS
Participants
The cohort of this study consisted of 90 white, healthy, randomly selected pregnant women and their healthy, term singleton infants, all of whom lived in 1 of the 3 northern provinces of the Netherlands. All women who registered with midwives and gynecologists between October 2001 and November 2002 in the province of Groningen were invited to participate in the study. This cohort was originally established for a study on the influence of prenatal exposure to environmental compounds on the development of healthy children.9
Follow-up
The children were invited prospectively to participate in an extensive longitudinal follow-up program. This program entailed the assessment of motor performance and cognition from birth until school age (5–7 years). Before the study, mothers gave their informed consent for themselves and both parents gave their informed consent for their children to participate in the follow-up program. The study was approved by the medical ethics committee of the University Medical Center Groningen.
Neonatal Period
When the infants were 10 days old we administered Prechtl's neonatal neurologic examination.10 This assessment tool contains 60 items for which an optimal range is defined. For example, this test assesses excitability (eg, apathy), motility and tonus (eg, hypotonia, dystonia), asymmetries, and visuomotor integration. We noted the presence of possible neonatal neurologic abnormalities and made a clinical interpretation of the infant's neurologic condition. The results of this assessment were used to classify the neurologic condition of the infants as normal, suspect, or abnormal. We also gave each infant a total neurologic optimality score (Prechtl's optimality score, range: 0–60).
Early Infancy
When the infants reached the age of 12 weeks, we performed Touwen's neurologic examination to evaluate the infants' neurologic condition.11 With this test we assessed in an age-specific context neurologic items similar to those examined with the neonatal neurologic examination. Again, we obtained a total neurologic optimality score for each infant (Touwen's optimality score, range: 0–60), and classified the neurologic condition of the infants as being normal, suspect, or abnormal.
When the infants were 12 weeks old we also assessed their quality of spontaneous general movements according to the method of Prechtl et al.12,–,14 This is the age at which fidgety movements emerge, movements characterized by small amplitude, moderate speed, and variable acceleration. We assessed the quality of fidgety movements and scored them as normal, abnormal (ie, amplitude, speed, and jerkiness were exaggerated), or absent. In addition, we calculated the infants' motor optimality score (range: 5–28 points) by assessing the quality of their fidgety movements, the quality and age-adequacy of their concurrent motor repertoire, and the presence and normality of their motor patterns.14,15
Toddler Age
When the children reached the age of 18 months, we assessed their motor and cognitive development with the Bayley Scales of Infant Development, Second Edition, Dutch Version (BSID-II-NL).16 The BSID-II-NL contains items that address the mental and psychomotor development of children, and assessment with the BSID-II-NL generates a Mental Developmental Index (MDI) and Psychomotor Developmental Index (PDI). Furthermore, we administered Hempel's age-specific neurologic examination to evaluate the children's neurologic condition. This neurologic examination contains items on posture, coordination of trunk and extremities, gross and fine motor function, visuomotor integration, and sensory function.17 For each infant we obtained a total neurologic optimality score (Hempel's optimality score, range: 0–57) and we classified their total neurologic condition as normal, suspect, or abnormal.
During this period we measured the mothers' verbal intelligence with the Wechsler Adult Intelligence Scale.18
School Age
When the children reached the age of 5 to 7 years, we assessed their motor and cognitive development. We determined their motor outcome with the Movement Assessment Battery for Children (Movement ABC), a standardized test of motor skills for children aged 4 to 12 years.19 This test, which is widely used in both practice and research, yields a score on total movement performance that is based on separate scores for manual dexterity (fine motor skills), ball skills, and static and dynamic balance (coordination). Tasks children perform during the Movement ABC include, for example, posting coins in a bank box, drawing a line between 2 existing lines on a figure, catching a bean bag, and jumping over a rope. The tasks on the Movement ABC are representative of the motor skills that are required of children attending elementary school and are adapted to the children's ages.
The children's cognitive development was assessed by a shortened form of the Wechsler Preschool and Primary Scale of Intelligence, revised.20 We used the subtests for vocabulary, picture completion, and reproduction of block designs to measure total, verbal, and performance intelligence.
In addition, we measured the children's attention and verbal memory. Sustained attention and selective attention were measured with the 2 subtests “Score!” and “Sky Search” of the Test of Everyday Attention for Children.21 Sustained attention involves maintaining attention over an extended period of time. Selective attention refers to the ability to select target information from an array of distractors.22 We assessed verbal memory with a standardized Dutch version of Rey's Auditory Verbal Learning Test.23 This test consists of 5 learning trials, and tasks include immediate recall of words (tested after each presentation), a delayed recall trial, and a delayed recognition trial.
To gain insight into maternal SES, we determined the mother's highest level of education during the first year after birth.
A summary of the different tests that were used, including time period, abbreviations, and range of scores, is provided in Table 1.
Longitudinal Follow-up Assessments
Statistical Analyses
IQ scores were calculated according to the mean of scores on the verbal and performance subtests. We classified scores on the BSID-II-NL, Movement ABC, and cognitive tests into normal (>15th percentile), suspect (>5th to ≤15th percentile), and abnormal (≤5th percentile). For the neurologic examination we performed the same score classification on the basis of criteria from the manuals. We used Pearson's and Spearman's correlation coefficients to explore the relationship between the scores on the developmental tests, when appropriate. For the BSID-II-NL, Movement ABC, and intelligence scores we used the normal scores as defined in the manuals, whereas for the Prechtl's, Touwen's, and Hempel's neurological optimality scores and examination optimality scores and the motor optimality score, we used the raw scores, because normal score criteria were unavailable. The correlations were calculated both before and after correction for maternal SES, because SES may act as a confounding factor for both motor and cognitive development.24,25
To investigate the stability of the scores obtained on motor development tests, we first calculated the z scores for each test so we could compare the different tests. Subsequently, we calculated the differences in z scores between the different time points (ie, between Prechtl's optimality score and the motor optimality score, the motor optimality score and Touwen's optimality score, Touwen's optimality score and the PDI, the PDI and Hempel's optimality score, and Hempel's optimality score and the Movement ABC). We then determined the mean absolute difference over all time points.
To compare the added value of the maternal characteristics, we calculated the proportion of variance (r2) in development at school age explained by the scores on the early tests and maternal characteristics. For normally distributed variables we used a linear regression model. For variables that were not normally distributed we used a logistic regression model in which we compared the school age scores of the children whose scores were classified as normal with those whose scores were classified as suspect and abnormal. Throughout the analyses P < .05 was considered statistically significant. This threshold of significance was uncorrected for multiple comparisons.
We used SPSS 16.0 software for Windows (SPSS Inc, Chicago, IL) for all the analyses.
RESULTS
Of the 90 children whose mothers were invited to participate in the study, 77 (86%) participated in the follow-up program at school age. Eight sets of parents declined the invitation to take part in the follow-up and 4 sets of parents could not be traced. One girl had to be excluded because she suffered severe cognitive impairment of unknown origin and, as a result, could not be tested. The demographics of the children are shown in Table 2.
Participant Demographics
Follow-up
The mean age at follow-up in the neonatal period was 10 days (range: 8–12). The mean for Prechtl's optimality score derived from the neurologic examination (n = 75) was 50 (range: 39–59).
During early infancy the mean age at follow-up was 12 weeks (range: 11–13). The mean for Touwen's optimality score (n = 77) was 45 (range: 35–51). Regarding the quality of spontaneous general movements (n = 72), 69 children showed normal fidgety movements, 2 showed abnormal fidgety movements, and in 1 child fidgety movements were absent. The mean motor optimality score was 25 (range: 10–28).
At toddler age, the mean age at follow-up was 18 months (range: 17 months 3 weeks to 18 months 1 week). The mean Hempel's optimality score (n = 75) was 43 (range: 31–54). According to the BSID-II-NL, the mean for the PDI (n = 75) was 93 (range: 69–118) and for the MDI (n = 75) the mean was 96 (range: 66–125).
The mean age at follow-up at school age was 6 years and 1 month (range: 5 years and 8 months to 7 years and 6 months). The mean total IQ of the children was 102 (range: 82–125; SD: 9), the mean verbal IQ was 100 (range: 78–130; SD: 10), and the mean performance IQ was 103 (range: 73–133; SD: 12). In Table 3 we present an overview of the longitudinal test results classified as normal, suspect, and abnormal. On the whole, the test scores were comparable to a reference population. The neurologic outcome in early infancy and toddler age, mothers' intelligence, and intelligence of the children at school age were slightly better than one would expect in the total population. Furthermore, selective attention was slightly better than sustained attention, although comparable to a reference population.
Longitudinal Test Results
Stability of the Developmental Scores Over Time
Figure 1 shows the correlations between the developmental scores at the various ages and the total score on the Movement ABC. Hempel's optimality score at toddler age correlated significantly with motor outcome at school age both before and after correction for SES (r = 0.244, P = .035 and r = 0.273, P = .018, respectively). The MDI at toddler age correlated significantly with motor outcome at school age after correction for SES (r = 0.272, P = .019). Additional analysis revealed that the MDI and Hempel's optimality score correlated significantly, in particular with the manual dexterity subtest (both r = 0.276, P = .019) of the Movement ABC.
The correlation coefficient (r) between developmental tests at various ages and motor outcome (Movement ABC) at school age (aP < .05).
Figure 2 shows the correlations between the developmental test scores and total intelligence at school age. None of these correlations reached significance. Nevertheless, we found a trend toward correlation between the MDI at toddler age and intelligence at school age before correction for SES (r = 0.194, P = .098).
The correlation coefficient (r) between developmental tests at various ages and intelligence at school age.
We also correlated the test scores in the neonatal period, early infancy, and toddler age with one another. After correcting for SES, we found that the motor optimality score in early infancy correlated with the PDI at toddler age (r = 0.262, P = .030). At toddler age we found a correlation between the PDI and Hempel's optimality score (r = 0.284, P = .015) and the MDI and Hempel's optimality score (r = 0.299, P = .011). The correlation between the MDI and PDI was also significant (r = 0.417, P < .001). At school age, we found that the children's motor development correlated significantly with intelligence and sustained attention (r = 0.267, P = .021 and r = 0.337; P = .004). Intelligence correlated significantly with selective attention and verbal memory (r = 0.253, P = .034 and r = 0.268, P = .024).
After transforming the test scores to z scores and calculating the mean absolute difference in z scores between the time points, we observed the largest difference in the test scores between Prechtl's optimality score (neonatal period) and the motor optimality score (early infancy), and Touwen's optimality score (early infancy) and the PDI (toddler age), with mean absolute differences of 1.03 SD (95% confidence interval [CI]: 0.81–1.24) and 1.15 SD (95% CI: 0.96–1.35), respectively. The mean absolute difference over all the time points was 1.01 SD (95% CI: 0.91–1.11).
Figure 3 shows the distribution of the neurologic test scores at the different time points, divided into 3 categories (<33rd, ≥33rd to ≤66th, and >66th percentiles) of motor performance at school age. As is apparent from Fig 3, there is little consistency in the motor trajectories of the children over time. One can see in the second column that in early infancy, the distribution of motor performance was approximately equivalent across all 3 neurologic test score groups. With increasing age, however, the tests do enable us to better identify those children whose motor scores will be below the 33rd percentile at school age.
The distribution of neurologic test scores at different time points (x-axis), divided into 3 equal categories (<33rd, ≥33rd to ≤66th, and >66th percentiles; y-axis) according to motor performance at school age. At each time point, the children were classified into 3 equal groups according to the percentiles of their neurologic test scores at that time. The slices of the pies represent the number of children of the groups that scored <33rd percentile (black), ≥33rd to ≤66th percentile (gray), and >66th percentile (white) at school age. For example, one can see that of all children with scores below the 33rd percentile in the neonatal period, only one-third had scores that remained below the 33rd percentile at school age (black slice), and nearly one-fourth had scores in the highest category at school age.
Early Predictors of Motor and Cognitive Development at School Age
The Movement ABC percentiles, IQ, and verbal memory were normally distributed, and were therefore included in a linear regression model. Attention was not normally distributed and was therefore included in a logistic regression model.
Table 4 gives the proportion of variance in motor and cognitive development at school age explained by early tests scores and maternal SES and verbal intelligence. Furthermore, we present the r2 for the complete model. Table 4 shows that the r2 values for the complete model on all outcome variables are low. We found significant but low r2 values for the MDI results on motor development at school age (r2 = 0.074, P = .018), SES on sustained attention (r2 = 0.104, P = .030), and mothers' verbal intelligence on verbal memory of the children at school age (r2 = 0.074, P = .027).
Explained Proportions of Variance in Motor and Cognitive Development at School Age
DISCUSSION
The results of our study demonstrated that the stability of motor development from birth until school age was limited in healthy, term-born singletons. Moreover, in healthy children the added value of developmental tests used at early ages to predict intelligence and more complex cognitive functions at school age was low. Only the contribution of maternal SES and verbal intelligence, although limited, was significant to children's sustained attention and verbal memory at school age.
Previous investigators reported contrasting results for studies of the stability of motor development in healthy children. Results of a study by Ridgway et al26 demonstrated that the age at which children reach certain milestones (ie, walking and unaided standing) has prognostic value for grades in school physical education performance at 14 years old. Another study showed that early fine and gross motor development in children at 4 to 48 months, as measured with a parental questionnaire, is not related to fine and gross motor development at 6 to 12 years.25 Our study showed that in healthy children even the standardized assessments of motor and neurologic performance during the neonatal period, early infancy, and toddler age had limited predictive value for motor development at school age. This finding indicates that results for outcome studies based on a single measurement at an early age should be interpreted with utmost caution.
A possible explanation for these conflicting results is that different tests at various ages may or may not measure the same function. Also, previous studies have used many different measures at different ages. In addition, the nature of the investigated relationships may vary. In certain studies the investigators may have searched for small effects that may not be relevant in everyday clinical practice but may be clues to neurodevelopment.
Several theories on motor development have been formulated to describe the mechanisms involved in the development of motor functions. These theories have shifted from the idea that motor development is a gradual unfolding of predetermined patterns in the central nervous system toward the notion that environmental factors and sensory information play important roles in motor development.27,–,29 The neuronal group selection theory states that cortical and subcortical systems in the brain are dynamically organized into variable neuronal networks, the structure and function of which are selected by development and behavior. Neuronal network selection performed on the basis of afferent information plays a significant role in this process. According to this theory, variation in developmental parameters, such as motor performance, developmental sequence, and the duration of developmental stages, is the key to normal development.29 Our findings that motor development stability is limited until school age may reflect this large variability.
In this study, however, we did find some correlations between test scores obtained at consecutive ages. We found correlations between the scores for motor optimality at early infancy and the PDI at toddler age, and between Hempel's optimality score at toddler age and the Movement ABC at school age. These correlations may be attributable to the fact that these measurements were obtained at subsequent ages. It is striking to note in this respect that the MDI at toddler age, which is considered a cognitive measurement, correlated with motor development rather than cognitive development at school age. Indeed, the MDI contains items that require the use of fine motor skills. This feature may account for our finding that the MDI correlated with manual dexterity on the Movement ABC.
Results of various reported studies have indicated that the trajectory of motor development in preterm-born children is more stable than the trajectory we observed for the term-born children in our study. For example, Erikson et al found that a majority of preterm children show stable motor development that could be predicted by low birth weight and the presence of cerebral damage such as periventricular leukomalacia.30 Other investigators have also found that early neurologic assessments are predictive of later neurodevelopmental outcome in preterm infants.31,32 The results of these studies seem to indicate that certain risk factors have impacts on neurologic development that are so strong that they cannot be counteracted by environmental factors such as SES, experience, and training. The development of healthy children, by contrast, seems to be more susceptible to variations in these factors.
With regard to the relationship between motor and cognitive development, we did not find any early motor test score that contributed significantly to the prediction of outcomes related to intelligence, attention, and verbal memory at school age. This finding contrasts with the finding of Piek et al,25 who stated that gross motor functioning, measured with a parental questionnaire, accounted for a significant proportion of the variance for cognitive performance. Results of other studies have suggested that specific items on motor tests, such as those that address aspects of the quality of spontaneous movements, or the age of reaching milestones, may have predictive value for later cognitive development.5,33 Our study showed that there was no clear relation of overall motor performance of children in early life with cognitive development at school age. The only factors that contributed significantly to the variance in complex cognitive functions were maternal SES and verbal intelligence. These findings are in accordance with those of previous studies in which investigators found that mother-infant interactions and environmental quality play important roles in intellectual development at preschool age.34,35 Our results concur with these observations, particularly for complex cognitive functions, ie, attention and memory.
Several considerations should be taken into account in the interpretation of the present results. One consideration is the question of whether the different tests at the various ages measured the same functions. The neurologic examinations measure the integrity of the central nervous system and enable investigators to identify infants at risk for minor and major neurologic dysfunction, whereas the Movement ABC performed at school age is a rather quantitative test of daily motor skills. Infants identified as having minor neurologic dysfunction, however, are at risk for development of clumsy motor behavior, which may lead to abnormal Movement ABC scores.36 Furthermore, we found that at toddler age, for example, the results of the neurologic examination correlated highly with the results of the children's PDI and MDI. This correlation may reflect the similar and interrelated functions and aspects of development that are measured by these tests. It is a fact that the BSID-II-NL and the neurologic examination both include a number of similar items. A second consideration is that the children's behavioral state might have had an impact on the test results. During the first year of life in particular, the children's behavioral state may have led to an increased variability in test results that may have made it difficult to find a relationship with development at school age.
We aimed to report our results transparently; therefore, we did not adjust our threshold of significance (P = .05) for multiple comparisons. In this explorative study we assessed motor function at 4 time points, and cognition at 2 time points. A multiple comparison correction would thus set our significance level at approximately P = .01–.02. Because the number of significant findings in our study was already limited, this adjustment would not have changed the main findings of our study. A final consideration is that the SES of our study population was relatively high and our sample size was thus biased to the upper classes. This limitation may also have influenced our findings.
As a suggestion for future studies, it would be interesting to see how developmental trajectories evolve beyond school age, because little is known about motor stability into adulthood.
CONCLUSIONS
The results of our study showed that in a cohort of healthy children motor developmental trajectories varied considerably. Moreover, the added value of early assessments of motor development for later cognitive function was limited. These results lead us to believe that a single abnormal test result at a certain age in an individual child at risk for developmental delay should be interpreted cautiously. The effect of the risk factor on development must be quite large to come to light at every assessment until school age. If a certain risk factor has only a moderate impact on development, it may soon be counteracted by factors that are known to have an important impact on development. Cases in point are maternal SES and verbal intelligence and possibly other factors that have yet to be discovered.
ACKNOWLEDGMENTS
The COMPARE project (Toolbox for Improving the Comparability of Cross-National Survey Data With Applications to the Survey of Health, Ageing and Retirement in Europe) was supported financially by the European Committee (Life Science Program, QLK4-CT-2000-0261). This study was part of the program of the research school of Behavioral and Cognitive Neurosciences, University of Groningen.
We greatly acknowledge the help of Dr V. Fidler for statistical advice, Dr T. Brantsma van Wulfften Palthe for correcting the English manuscript, and E. Drent for help with the figures.
Footnotes
- Accepted July 9, 2010.
- Address correspondence to Elise Roze, BSc, Division of Neonatology, Beatrix Children's Hospital, Hanzeplein 1, 9713 GZ Groningen, Netherlands. E-mail: e.roze{at}bkk.umcg.nl
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- SES =
- socioeconomic status •
- BSID-II-NL =
- Bayley Scales of Infant Development, Second Edition, Dutch Version •
- MDI =
- Mental Developmental Index •
- PDI =
- Psychomotor Developmental Index •
- Movement ABC =
- Movement Assessment Battery for Children •
- CI =
- confidence interval
REFERENCES
- Copyright © 2010 by the American Academy of Pediatrics
In this issue
