(introductory text...)

New understanding of the role of nutrition in child development points to suggestions for future research.

Biological and behavioral mechanisms

Research efforts are gradually revealing a finer-grained picture of the processes linking nutrition and observable behavior. Various mechanisms involving both biological and behavioral aspects of development are implicated and bear investigating.

Biological. Earlier research linking PEM and behavior suggested that undernutrition interfered with the development of the central nervous system. Undernutrition reduced brain weight and the number of brain cells, which in turn were seen as the cause of irreversible detriments in cognitive and motor performance (NAS, 1973). This emphasis on brain growth focused attention almost exclusively on the period of maximal brain growth, seen as the period of greatest vulnerability.

New findings indicate, however, that periods before and after that of maximal brain growth may be equally important. It is now understood that critical aspects of central nervous system development - for example, gliogenesis, macroneurogenesis, and early glial and neuronal migrations - precede the period of maximal brain growth. Other Later processes, such as synaptogenesis and myelinization, may also be sensitive to insult and remediation (Levitsky & Strupp, 1995; Strupp & Levitsky, 1995).

Research during the last decade has shown that the effects of undernutrition late in gestation are similar to those occurring early. In the case of previously undernourished animals, for instance, the period of brain growth can be extended, and remarkable recovery has been observed. Such evidence is leading researchers to consider a broader range of possible biological mediators, including brain differentiation and changes in neuroreceptor sensitivity. For example, perturbations at the sub-cellular level, as suggested by alterations in sensitivity to pharmacological challenges, persist after periods of early undernutrition and nutritional rehabilitation (Levitsky & Strupp, 1995).

Behavioral. In the mid 1970s the concept of functional isolation, referring to restricted interaction with the environment, was proposed to explain the long-term behavioral effects of early undernutrition (Levitsky, 1979; Levitsky & Barnes, 1972). It was hypothesized that it is the differential experience of the organism rather than disrupted brain growth that mediates the effects of early undernutrition over time. The child who is undernourished attempts to maintain energy balance by reducing energy expenditure and withdrawing from environment stimulation. Such withdrawal limits the child's capacity to take in environmental information and thereby acquire the skills and knowledge necessary for normal behavioral development (Levitsky & Strupp, 1984). This concept arose from evidence suggesting that the behaviors affected by early undernutrition were similar to those produced by early environmental isolation.

Although the functional isolation hypothesis was initially developed as an alternative to a purely biological explanation of nutrition mediated behavioral deficits, the two explanations may in fact be compatible rather than conflicting. Functional isolation may actually influence both central nervous system and behavioral development. While not all aspects of the CNS may be sensitive to environmental influences, and the extent of effects may be relatively small (Bedi & Bhide, 1988), evidence from behavioral neuroscience studies illustrates how restricting experience may adversely influence development (Diamond, 1988; Greenough & Black, 1992) and efficiency (Stone, 1987) of specific brain structures and processes. In addition to the influence of functional isolation on both brain and behavior, subsequent neural changes may further accentuate the effects of functional isolation on ultimate development.

Recently, the functional isolation proposition has been elaborated to explain in greater detail some of the mechanisms that may contribute to long-term adverse effects of undernutrition on cognitive development (Pollitt et al., 1993). This revised proposal hinges on the well-documented effects of undernutrition on body size, neuromotor development, and physical activity. If the child is small and physically underdeveloped and inactive, he or she may

(1) induce behaviors and social responses from caretakers that would generally be reserved for younger children;

(2) undertake less exploration of the environment; and

(3) consequently lag in acquiring the motor skills, cognitive abilities, and social behaviors that typify normal development.

These patterns can operate independently and interactively, with cumulative effects, such that the child ultimately falls behind in competencies attained by well-nourished children. Investigating how nutrition promotes individual differences in children's motor skills, exploration, and play behavior would provide a welcome test of this hypothesis.

Assessing nutrition

Focusing on individual nutrient deficiencies is particularly problematic within populations in which undernutrition is a major public health concern. With the possible exception of populations where this is not a problem, studies of single nutrients are no longer adequate (Golub et al., 1995; Pollitt, 1995) - as indicated by the coexistence of multiple nutrient deficiencies (Schürch, 1995), complex interactions in the absorption and utilization of nutrients, and the demonstrated effects of different nutrients on CNS function. Moreover, failure to account for the relationship between PEM and micronutrient deficiencies has led to inconclusive findings regarding the causal role of distinct nutrients on cognitive outcomes (Pollitt et al., 1993).

Too often studies are marred by a lack of information on the overall nutritional status of the population in question. For example, studies testing the functional consequences of particular nutritional deficits (e.g., energy) have floundered, because the prevalence of the deficit in the populations under study was negligible (Allen, 1993). It is important, therefore, to obtain survey data on nutritional indicators before implementing field studies. Further, if limits in knowledge and technology preclude the use of laboratory measures to determine nutrient status (as in the case of zinc), then alternative procedures (e.g., response to treat meet) should be used to establish baseline values. While this approach could be costly, the yield to science would be rewarding.

Measuring outcome

Just as the field is moving beyond an emphasis on single nutrients, so the focus on cognitive development to the exclusion of biological and psychosocial development no longer suffices in the investigation of undernutrition effects. Research is expanding to encompass the broader context and the multiple risks that interact with nutrition in determining outcomes for the undernourished child.

Assessments of affective characteristics, e.g., temperament, reactivity to stress, self-regulation, and emotional regulation, stand to shed light on the effects of undernutrition on the behavioral adjustment and psychological functioning undernourished children. Although such attributes are typically treated as innate, increasing knowledge of the CNS processes that underlie individual differences in temperament and link temperament and cognitive processes makes research in this domain promising. One issue is whether the nutritional environment could modify the genotype of temperament.

Likewise, studying the effects of undernutrition on the simultaneous or sequential relationships between developmental systems is also critical. The link, for example, between neuromotor and cognitive delays in the undernourished infant merits attention in light of new information showing that self-locomotion is an antecedent to perceptual development (Bertenthal & Campos, 1990; Lockman & Thelen, 1993).

We must understand more about the neural mechanisms through which undernutrition and related factors translate into individual differences in behavior and development. Recent advances in biomedical methodology promise a more direct assessment of critical nutrition-sensitive CNS processes. For example, nuclear magnetic resonance spectroscopy allows noninvasive assessment of changes in CNS energy metabolism (Holtzman, McFarland, Jacobs, Offut, & Neuringer, 1991) and has been used with some success to distinguish the CNS metabolic concentrations of at-risk and normal infants (Cady et al., 1983).

Finally, the possible effects of undernutrition- across generations must be identified. At issue is how biological and behavioral mechanisms may contribute to the transfer of a burden of undernutrition from one generation to the next (Susser & Stein, 1994). Several longitudinal studies of severe and mil-do-moderate undernutrition in early life, launched in the past, offer unique opportunities for the follow- of new generations.

Accounting for factors that modify effects

Studies are needed of the effects of undernutrition over the lifespan, from the earliest stages, including the prenatal, to old age. But, while age may be related to the outcome, we must identify what factors related to increasing age act to modify outcomes. Obviously, effects are not necessarily mediated only by changes in the CNS occurring during the specific period of maximal brain growth (Levitsky & Strupp, 1995). How, for instance, is the educational achievement and progress of children hampered by undernutrition (Soemantri et al., 1985)? Or how does Atonal education limit the adverse effects of early undernutrition?

Future studies should also attend to the role of intra-individual and environmental factors that protect against or accentuate the risks of undernutrition. Research designs that would, for example, track the chain of relationships - of effects of undernutrition on the child's behavior, the child's behavior on caregiver behavior and vice versa, and possible buffering or deleterious factors on these relationships - would go a long way toward clarifying how undernutrition affects behavioral outcomes and development.

Finally, what is needed are studies that capture the broader context, the "human condition." Undernutrition among children in economically impoverished populations is likely compounded by multiple risk factors-by conditions besides undernutrition. With few exceptions (Chafes & Martinez, 1984), the recognition that such interactions do exist has failed to lead to studies that attempt to disentangle them or that seek to identify the factors that either increase or decrease the risks. It is, for example, important to understand why, in one study, poor, undernourished children experienced significant developmental delays, while middle-class children who suffered from severe undernutrition secondary to medical problems showed no such effects (Pollitt, 1987).