Previous - Next
Abstract
1. Background
2. A theoretical formulation
3. Purpose
4. Methods5. Results
6. Discussion
References
Discussion (summarized by C.M. Super)
E. POLLITT and K. GORMAN *
* Department of Applied Behavioral
Sciences, University of California, Davis, CA 95616, U.S.A.
Physical activity as a component of energy expenditure and physical activity as a determinant of child development are not congruent concepts. For the study of functional long-term effects of nutrition in infancy on later development, the age at which motor milestones are reached, conceptually, seems to be a more appropriate intermediary variable than physical activity per se. Data from a longitudinal study in Guatemala were used to test the hypothesis that early energy deficiency, as evidenced by growth retardation, is associated with delays in the reaching of motor milestones in the first 15 months of life and that these, in turn, are associated with measures of intellectual ability and educational achievements in childhood and adolescence. Birthweight did not predict anthropometric data or scores of motor and mental development tests at 15 months. Anthropometric data at 15 months, however, predicted mental and, even to a greater extent, motor test scores. Motor test scores and height, but not mental test scores, predicted results of cognitive tests during the preschool years. Motor test scores, but neither mental test scores nor anthropometric indicators, predicted cognitive performance in adolescence. These results suggest that one practical way of looking at possible long-term effects of chronic energy deficiency in infancy and early childhood may be by assessing its impact on milestones of motor development.
Researchers investigating relationships between nutrition and function face formidable challenges in the search for an answer to the question of whether the maintenance of energy balance at low levels of energy intake in infancy and early childhood has adverse behavioral or psychological effects in adolescence and adulthood. In addition to requiring longitudinal data which are costly, the nature of the task precludes tightly controlled experiments with infants and children. Moreover, the question calls for a close integration of concepts and methods from, at least, nutrition, physiology and developmental psychology. Difficulties notwithstanding, the question must be faced as it is a crucial aspect of the problem of human capital formation in many developing countries (POLLITT and AMANTE, 1984).
The question as generally formulated (Figure 1; BEATON, 1984) suggests that physical activity as a component of energy expenditure and physical activity as a determinant of development are equivalent. Such a formulation, however, is equivocal. Theoretically, it is to be expected that the nature of physical activity as a dependent variable in the equation of energy balance, and as an independent variable in the equation of developmental effects must differ substantively.
Physical activity as a component of
energy expenditure includes all body movements (directed e.g., grasping and
non-directed actions). Measurements of energy expenditure resulting from physical activity
yield information on a single dimension of an action; their behavioral or adaptive
significance is irrelevant in such a context. On the other hand, physical activity of
developmental significance is a more complex construct which does not lend itself to a
simple unidimensional expression, particularly when we consider the multiple systems and
subsystems involved. In addition to energy expenditure, developmentally meaningful motor
actions imply the involvement of critical dimensions related to the organization of motor
schemes, information processes, and motivational and affective elements, among others. The
time-space configuration of a behavioral action goes beyond the simple summation of how
energy was spent. At issue is how to identify developmentally meaningful motor actions.
The position adopted in this paper is that the reciprocal interaction that exists between motor maturation and organized physical activity suggests a way of addressing such an issue. We propose that, in a population at risk of chronic energy deficiency, the study of the relationships between motor milestones in early life - obviously dependent on the availability of energy - and later functioning is a test of the functional significance of reduced physical activity. To test this hypothesis we have taken advantage of a longitudinal study in rural Guatemala that allows for correlational analyses of motor maturation during the first 15 months of life and performance on a series of functional tests in adolescence.
We first assume that physical activity contributes to motor maturation which, in turn, is a codeterminant of organized action. Physical activity mobilizes and strengthens the skeletal and neuromuscular systems and allows for the onset (i.e., milestones) and development of increasingly complex movements. Motor maturation, in turn, is a prerequisite of activity involving directionality, as organization is dependent on selective movements. For example, before a child creeps for the first time it must have reached a stage of maturation of the skeletal and neuromuscular system that makes the action mechanically possible. Conjointly, creeping Implies that the infant has reached a developmental stage that allows for the onset of that particular directed action. Once a child has incorporated creeping into its repertoire, its view of the physical environment changes as it is then exposed to visual and tactile stimuli which were previously beyond reach or unavailable.
The onset of a developmentally significant action (e.g., creeping, sitting, standing), dependent on the reciprocal interactions of physical activity and motor maturation, is not necessarily the result of a directive from a central control at a particular instant. From a systems view of growth and development (KUGLER, SCOTT KELSO and DURVEY, 1982; THELEN, 1987; THELEN and FOGEL, 1989) the action is viewed as a result of the simultaneous presence and relationship of various elements in different developmental domains. The convergence of the availability of energy, the presence of necessary antecedent movements, the maturity of the skeletomuscular system, drive, and the appropriate environmental circumstances may all be needed.
It follows that data on motor maturation, particularly derived from the assessment of motor milestones, should provide useful information for the study of developmentally meaningful actions. Based on the arguments already formulated, we propose that if physical activity is reduced under conditions of chronic energy deficiency, then the onset of developmentally meaningful actions will be delayed. It is already known that the motor maturation of malnourished children is delayed (MALINA, 1984) and that nutrition supplementation has beneficial effects on the performance of nutritionally at risk infants on motor developmental scales (Jogs and POLLITT, 1987). What is not known is whether motor maturation predicts functional performance in adolescence and adulthood.
If the propositions in the previous paragraphs are correct, then the maturation of motor actions (i.e., milestones) that lead to developmentally meaningful behaviors should be correlated with cognitive test scores in later life in populations with chronic energy deficiency in early childhood.
The attractiveness of this proposal is enhanced by the solid evidence already available that in well-nourished populations the scores on scales of motor and mental development in the first 24 months of life maintain a zero correlation with later cognitive test scores (McCALL, 1979, 1981; LEWIS, JASKIR and ENRIGHT, 1986). Thus, the search for a statistically significant correlation, even in a population which is nutritionally at risk, goes against common wisdom in developmental psychology. Evidence that motor test scores in infancy and cognitive test scores in childhood and adolescence covary in such a population would strengthen the theoretical argument presented previously. The argument would be even stronger if the correlations in the same nutritionally at risk population between mental development test scores and childhood and adolescent scores were, as in the case of the well-nourished populations, not different from zero.
In particular, we hypothesize that
energy deficiency in infancy affects the timetable of motor maturation and delays the
onset of behaviors which are instrumental in the reciprocal interaction between the infant
and its environment. The acquisition of behaviors such as sitting alone, rolling,
creeping, standing alone, taking steps, or walking alone, broadens the infant's view of
the world and allows for engagement in more diverse motor activities than before the onset
of such behaviors (BAYLEY, 1936). We also conjecture that the delay in the acquisition of
such behaviors and the consequent interference with the interaction between the infant and
its environment has long-term developmental consequences. By this we mean that delayed
motor maturation due to chronic energy deficits interferes with the acquisition of basic
abilities and skills that are necessary to function in society. School-related behaviors
are examples of such abilities.
The purpose of this paper is, first,
to assess whether early energy deficiency, as evidenced by growth retardation, is
associated with delays in the reaching of motor milestones in the first 15 months of life.
The second purpose is to determine whether the variability in the timetable of those
milestones is associated with the variability in intellectual ability and educational
achievements in childhood and adolescence.
Figure 2 is a scheme of the longitudinal analysis presented. The arrows throughout the scheme point to the correlations that have been calculated. The first set of analyses examines the relationship between birth weight, and height- and weight-for-age at 15 months, as well as between birth weight, and mental and motor development at the same age. A second set includes calculations of the interrelationships between anthropometry and mental and motor assessments at 15 months. These two sets of analyses test the proposition that energy deficiency in infancy is related to a comparative delay in motor maturation. It also tests whether the same predictive factors correlate with mental development.
The third set examines the associations between mental and motor development test scores at 15 months and functional performance in the early childhood period (36, 48, 60, 72 and 84 months of age). The final set includes calculations of the correlations between scores on the developmental scales and tests of functional performance in adolescence. These last two sets test the proposition that variability in motor maturation at 15 months will account for a significant portion of the variance in functional performance at a later age period.
This developmental analysis relates
data from the study on nutrition supplementation and mental development conducted at INCAP
from 1969 to 1977 with data collected in 1988 and 1989 on the same subjects.
The study population consisted of the children born between 1969 and 1977 in one of four Ladino villages in Eastern Guatemala. They participated in a longitudinal study of the effects of nutritional supplementation on growth and development. The villages were selected because of similar demographic, social and economic characteristics. All children up to 7 years of age in 1969 and all children born into the villages from the initiation of the project composed the sample.
Subjects were included in this study
if the data required for the four sets of analyses were available. For the first two sets
of analyses, data on birth weight, body growth measurements and developmental assessments
at 15 months were required. For the third set the criterion was availability of
measurements at 15 months and of psychological test scores at 36, 48, 60, 72 or 84 months.
For the last set the criterion was the availability of the same early assessment and
function test data in adolescence. The sample size for the calculations of the different
correlation coefficients ranged from 78 to a maximum of 306 subjects.
4.2.1. Anthropometry
The methods used to collect birth weight and anthropometric data at 15 months have been reported in detail in other publications (MARTORELL, et al. 1980).
4.2.2. Motor development and cognition
4.2.2.1. Infancy
A Composite Infant Scale was used for the assessments of mental and motor development at 15 months (Instituto de Nutricion de Centro America y Panama, 1974). The infants were tested in a building in the villages where most of the research-related activities were conducted. The infants were tested at 6, 15 and 24 months (±15 days). Reliability and predictive validity data of the infant assessments have been published (LASKY et al., 1981).
The Composite Infant Scale was developed in Guatemala for the specific purposes of the supplementation study. Test items were selected from most of the popular developmental scales available (e.g., Bayley, Cattell, Gesell). The items were then classified according to their apparent weight in measuring mental or motor domains of development. The scores corresponded to the percent of items passed on the respective scale.
The 15-month motor scale was used for the present analysis because it included more items assessing motor milestones that are critical for physical activity than the 6- or 24-month scales. The following four motor categories of items are included: (1) sitting: ranging from tilts forward and is able to sit back, to moves from sitting to prone position, and rotates on a vertical axis; (2) stands on two feet: ranges from resting on pole to stands alone; (3) walks: from needs to be sustained with two hands to seldom falls; (4) walking up stairs (INCAP, 1974).
4.2.2.2. Preschool period
The original battery consisted of 22 tests which assessed a variety of skills such as verbal reasoning, verbal processes, learning, memory and perceptual-analytical skills. The tests were administered to all children once a year on their birthday (±15 days) from 3 to 7 years. At 36 and 48 months, ten scales were given to the children. At older ages, additional tests were used. In an attempt to maintain consistency across ages, for the present analysis only the following ten tests which were administered to children at all ages were included.
Embedded Figures: A card with a picture (probe) of an animal, flower, or house was shown to the child. A series of additional cards of figures were then presented, one at a time, and the subject was instructed to find the probe in each picture. An example with corrective feedback was administered prior to testing. The score represented the total number of correct identifications.
Embedded Figures, Time Delay of Response: The time (in seconds) elapsed between the presentation of the card and the response in the Embedded Figures test was measured.
Memory for Digits: Twenty-six series of numbers, varying in length from 2 to 8 digits, were read slowly, and the subject was instructed to repeat the series in the same order it was read. One series was read first as an example. The score consisted of one point for each correctly recalled series.
Memory for Objects: Twelve objects which could be clustered into three conceptual categories (animals, clothes, utensils) were presented on a tray. After a period of observation the objects were removed from the subject's view. The subject was then asked to recall as many objects as possible. The score consisted of the number of objects correctly recalled.
Memory for Sentences: Twenty-six sentences of increasing length from 2 to 14 two-syllable words were included in this test. The sentences were read slowly for subsequent recall. The score consisted of one point for each word correctly recalled plus a point if the sentence was recalled without additions, omissions, or transformations.
Picture Naming: Ten cards containing either 4 or 5 figures were included in this test. The subject was instructed to name the figure as the examiner pointed to it. A point was given for each figure correctly named.
Picture Recognition: The same 10 cards shown in the Picture Naming test were used for picture recognition. All figures which were named incorrectly were presented again and the subject was instructed to point to the object in the card named by the examiner. The score was the total number of items correctly identified.
Discrimination Learning: A candy was hidden under one of two visual stimuli (large or small house, large or small car) and the subject had to learn which attribute (e.g., large) of the stimulus was associated with the hidden reward. Criterion was five successive correct responses. The score was the number of trials to criterion.
Draw a Line: The purpose of this test was to evaluate the ability to inhibit motor control. Subjects were instructed to draw one line slowly and another line quickly as demonstrated by the examiner. After drawing the first line, subjects were instructed to draw a second line even more slowly than the first. The score was the length of the line divided by the time it took to draw it.
Impossible Puzzle: The purpose of this test was to measure the degree of persistence and resistance of the child when faced with a difficult task. Subjects were presented with puzzles designed to be impossible to resolve. Over a 3-minute period, subjects scored a point for every 10-second interval they remained on task. At the end of this period, subjects were given the option of continuing to play with the puzzle or choosing another activity. Over a one-minute period, subjects scored a point for every 10-second interval they continued to play with the puzzle. The score was the number of time intervals on task.
Data Reduction: In order to interpret the data from the numerous scales, factor analysis was performed with the results from each of the tests of the Preschool Battery at the five time points. The first factor to emerge in each of the factor analyses contained a strong verbal component after orthogonal rotation. For example, at 36 months the highest loadings on this factor were vocabulary recognition and naming, memory for objects, and embedded figures. At 48 months, factor one represented verbal ability with the highest loadings for recognition, naming, and recalling objects. At 60 months, the highest loadings on the first factor were picture naming, picture recognition and draw a line. At 72 and 84 months, a more general factor emerged with the highest loadings both on the naming and recognition tests as well as on the digit and sentence span tests.
4.2.2.3. Adolescence
A literacy test, standardized
Guatemalan tests of reading and vocabulary (i.e., Inter Americana), an IQ test (i.e.,
Raven Progressive Matrices) and the maximum grade attained in primary school were the five
measures of functional performance included in the present analysis. These measures
reflect the competence of an adolescent in key social and educational domains in
Guatemalan society. The literacy test was scored on a one-to-four-point scale; only those
who obtained a score of three or four points took the reading and vocabulary achievement
tests. For maximum grade attained, only those children with some schooling were included
in the analysis.
Table 1 presents the results from the first two sets of analyses: (1) correlations between birth weight and the developmental test scores at 15 months of age; (2) correlations between the physical growth measurements at 15 months and the developmental test scores at 15 months. Correlations which include height and weight measures at 15 months, partialling out the covariance with birth weight, are also included. The table also presents correlations between anthropometric data and motor scores, partialling out the variance shared with the mental scale scores. The correlation between the mental and motor scales was 0.426 (N = 365; p < .000l).
Table 1. Correlations: Prenatal (birth weight) and postnatal growth measurements with development test scores (mental and motor) at 15 months
Measurement |
Motor |
Mental |
Motor (mental partialled out) |
Birth weight (BW) (N = 274) |
.067 |
.113 |
.021 |
Z-Weight-for-age 15 mo (N = 325) |
.314 ** |
.188 * |
.317 ** |
Z-Weight (BW partialled out) |
.307 ** |
.142 * |
.272 " |
Z-Height-for-age |
.357 ** |
.177 ** |
.362 ** |
Z-Height (BW partialled out) |
.352 ** |
.139* |
.323 ** |
Z-Weight-for-height |
.047 |
.079 |
.084 |
Z-Weight-for-height (BW partialled out) |
.023 |
.036 |
.008 |
* p < .05
* p < .001
The correlations between birthweight and the mental and motor test scores were not statistically significant. Likewise, the correlations between weight-for-height at 15 months and the mental and motor scores were not significantly different from zero. Conversely, the correlations between weight-for-age and height-forage (with or without the shared variance with birthweight) and the mental and motor scores were statistically significant. However, the magnitude of the correlations differ between those involving the mental and motor scores. Among the correlations involving the mental test scores the highest coefficient was 0.188 with weight-forage; on the other hand, the lowest coefficient including the motor score was 0.307, with weight (partialling out birthweight effects). These two coefficients were significantly different (p < .05) from each other. Height-for-age at 15 months, the anthropometric indicator most highly correlated with motor development, explained about 13% of the motor scale variance.
Figure 3 presents the set of correlations between the measurements at 15 months of age and the verbal factor from the Preschool Battery at 36, 48, 60, 72 and 84 months of age.
Measurements at 15 months include height-for-age, weight-forage, and the mental and motor scale scores. The motor score at 15 months was a better predictor of verbal scores at 36 and at 48 months of age than the mental score at 15 months. However, those differences in predictive power between the mental and motor scores disappeared at 60, 72 and 84 months. Height-for-age and weight-for-age at 15 months were equally good if not better predictors of performance from 36 to 84 months than the motor scores at 15 months.
In connection with the assumptions of this paper regarding the importance of early motor maturation for later development in a population which is nutritionally at risk, the final set of analyses are of critical importance. These analyses specifically focus on the predictive power of early motor maturation on adolescent functional performance. As already indicated, it is well established that in well-nourished populations the correlations between early motor maturation and adolescent functional performance are not significantly different from zero.
Figures 4-8 are histograms reporting the magnitude of the correlations between the mental and motor scores and anthropometry at 15 months, and scores on tests of functional performance in adolescence. In all except one case (i.e., Raven Progressive Matrices), the motor test scores at 15 months predicted the scores on the outcome variable in adolescence. In particular, after controlling for mental test scores, the motor test scores predicted literacy (r = .16; p < .05; N = 220), reading scores (r = .22, p < .01; N = 173), vocabulary scores (r = .18; p < .05; N = 173), and the maximum grade attained in school (r = 14; p < .05; N = 231) in adolescence. Conversely, the mental test scores, weight-for-age and height-for-age at 15 months failed to predict any of the outcome variables in adolescence.
A basic premise of the analyses presented in this paper was that data on motor maturation among infants who were nutritionally at risk have the potential of providing insightful information on long-term effects of chronic energy deficiency. The correlational analyses presented were rooted in the proposition that if the maturation of motor actions (e.g., creeping) that lead to developmentally meaningful behaviors (e.g., environmental exploration) are delayed by chronic energy deficiency, then motor maturation test scores of infants who are nutritionally at risk should be correlated with their cognitive test scores in later childhood and adolescence. The analyses were done in the context of a large body of developmental data which conclusively show that, among well-nourished subjects, the scores on developmental scales of motor and mental development in the first 24 months of life maintain a zero correlation with later cognitive test scores.
In this study, intrauterine growth as reflected by birth weight neither predicted anthropometric data nor mental and motor test scores at 15 months. On the other hand, at 15 months of age, weight and height predicted both mental and motor developmental test scores. However, both measurements of body growth were more closely associated with the motor than with the mental scores. This finding is consistent with the available evidence which suggests that, in infancy, motor maturation is more sensitive to nutritional insult than those behaviors that are generally interpreted as reflecting mental function (Jogs and POLLITT, 1984).
The third set of analyses showed that, in comparison to the mental scores, the motor scores were better predictors of the factor scores that emerged from the preschool cognitive test battery at 36 and 48 months. However, linear growth measures at 15 months were as good predictors as motor test scores of performance in the preschool period. Thus, these analyses provided only weak support for our theoretical formulations.
On the other hand, the fourth set of correlational analyses between motor scores at 15 months and the functional performance measures in adolescence were in full agreement with our formulations. Motor test scores at 15 months predicted at a statistically significant level (p < .05) literacy, reading and vocabulary scores, and maximum grade attained in school. On the other hand, as is the case of many other studies among well-nourished children, the present analyses showed that neither the mental test scores nor anthropometry in infancy predicted cognitive performance in adolescence. It must be underscored that, despite our familiarity with the literature, we do not know of any other longitudinal study in the world among well-nourished subjects that has conducted analyses similar to those reported in this paper and obtained similar results. Similar analyses with chronically energy-deficient populations are not available.
An explanation which could be an alternative to the thesis presented in this paper is that both motor maturation and cognitive test performance in this sample were dependent on a third variable. Social and economic variables are obvious candidates for such an explanation. However, this alternative is seriously weakened by the fact that there is no obvious reason why social and economic variables would affect motor and not mental test scores. If the underlying mechanisms were social environmental factors then the mental development scores should have been as strong predictors of the adolescent test scores as the motor score. This was not the case.
Figure 9 depicts in diagrammatic form the pattern of relationships that we propose exist between energy intake, motor maturation, physical activity and developmental test scores in two populations adapted to two different levels of energy intake. In one, the supply of energy meets the needs of the population which is well-nourished, and motor maturation and physical activity are independent of energy intake. Exceptions to these zero relationships between intake, maturation and activity are likely to be found in cases of obesity. Obese children mature faster (DIETZ, 1987) and are less active (DURNIN, 1984) than children whose weight falls within the normal range. Further, for most of the population, functional performance in adolescence is also independent of motor maturation and activity. Exceptions to these cases may be found in the presence of neurodevelopmental pathology and gross motor maturational delay in infancy and early childhood. These have been associated with poor cognitive test and educational achievement performance in the school-age period (SILVA, 1980).
In contrast, in a population where the supply of energy fails to meet the physical needs (Figure 1) and accommodations must be made to reach energy balance (i.e., in a chronically energy-deficient population), the relationships between the variables in question would be significantly different from zero. As the results from the analyses that have been presented suggest, low intake (inferred from retarded growth) is related to motor maturation (i.e., motor test scores) which, in turn, is related to developmental test scores in adolescence. The associations between physical activity and maturation are only inferred from the theoretical formulations in the introduction.
It cannot be overemphasized that the results from this study must be interpreted with caution. However, they do suggest that one practical way of looking at the possible long-term effects of chronic energy deficiency is by assessing its impact on motor milestones which are important determinants of cognitive and socioemotional development.