To our knowledge, this study provides the first evidence that regional differences in cerebral activity may be influenced persistently by early developmental experiences. We found inverse linear associations between markers of fetal growth and RH activity 8–9 years later, which strengthened when fetal growth was considered in relation to placental growth. The associations seen at rest were markedly stronger in the period immediately following the TSST-C, suggesting that stress unmasks inherent differences in cerebral laterality. Thus, the activity of brain regions involved in the processing and response to stressful stimuli may be particularly susceptible to altered lateralization in fetal life. The associations were continuous across the normal ranges of birth weight and placental weight and not dependent on a threshold effect such as might be seen if they resulted only from medically significant intrauterine growth retardation. As the children were born at or near to term and birth size measures were adjusted for gestational age, the findings are not the result of prematurity.
We studied young children who are unlikely to have been affected by the psychological disorders or neurodegenerative processes that usually appear in later life. Our findings were robust to adjustment for current weight and maternal socioeconomic status and smoking, increasing the likelihood that they represent the direct effect of altered neurodevelopment.
We found that hand preference was related to TMT differences indicative of greater baseline activity in the hemisphere that was not dominant for motor functions. While this finding may seem counter-intuitive, it agrees with long-established evidence that overall hemispheric activity is usually greater in the hemisphere that is not dominant for motor activities 
. However, the reasons for this remain a matter of speculation. Because of this finding, further analyses of TMTL were adjusted for hand preference.
We used birth weight as a summary measure of fetal growth. As the heritable component of birth weight is small 
, it may be considered a non-specific reflection of adverse intrauterine influences, to which fetal growth is highly sensitive 
. As these influences may be disparate, including variations in placental function and maternal nutrition, health and stress, our findings point to a need for future investigations to establish which particular developmental influences may alter later laterality of cerebral activity.
In healthy pregnancies, fetal and placental weights generally increase proportionately. In childhood, we found that birth weight adjusted for placental weight was a stronger determinant of increased RH activation following stress than birth weight alone. This suggests that disproportionate growth of the fetus and placenta is more important than fetal growth restriction alone in determining cerebral laterality. In our data, a relatively small baby with a relatively large placenta was associated with a neurodevelopmental outcome that might be associated with impaired neurocognitive function in later life. This is not the first time that such disproportion has been linked to later pathology. For example, increased placenta to fetus weight ratio has been linked to risk of hypertension in adulthood 
. Interestingly, however, so has a decreased placenta to fetus weight ratio 
and other components of the metabolic syndrome also appear to have a ‘U’-shaped relationship with placenta to fetus weight ratio 
. Thus, although fetal growth in disproportion to that of the placenta is likely to indicate a departure from normal, healthy growth at some stage during pregnancy, underlying causes may differ. Therefore, this indicator of fetal adversity, like birth weight, remains non-specific but may be more sensitive, given its stronger associations with asymmetry of cerebral activity in our study.
In our study, a relatively small fetus with a relatively large placenta might be considered detrimental. Although a precise pathogenesis cannot be determined, existing data allows for some speculation on the type of placental adaptation that may have occurred. Maternal undernutrition during pregnancy may stimulate placental overgrowth as an adaptation to poor nutrient supply in an effort to improve fetal nutrition in later gestation 
. In early gestation, the placenta is less developed and has the greatest growth potential. Therefore, adaptation to poor nutrient supply might be best achieved through placental hypertrophy. In later gestation, placental enlargement may be limited by maternal constraint and diminished developmental plasticity and adaptations to improve nutrient transfer usually take the form of altered placental structure and function, rather than size. Thus, the gestational timing of adverse influences on the pregnancy, such as maternal malnutrition, may be a determinant of the direction of feto-placental growth disproportion. There is animal 
and human 
data that supports this, with undernutrition in early pregnancy being associated generally with increased placental weight relative to that of the fetus and there is some evidence that undernutrition in late pregnancy can be associated with a decreased placenta to fetus weight ratio.
Recently, maternal nutritional status has been implicated as a moderating factor that might determine the polarity of feto-placental growth disproportion 
. In the offspring of short, low socioeconomic status mothers, a relatively small placenta was associated with later hypertension, whilst in the offspring of tall, wealthier mothers, a relatively large placenta was associated with hypertension. This was interpreted as evidence that compensatory placental enlargement is more likely in mothers who are more likely to have good nutrition in later pregnancy. As none of the parents in our study fell into the ‘routine work’ or ‘never worked/long-term unemployed’ categories at the lower end of the socioeconomic scale, this may explain why only feto-placental disproportion with a relatively large placenta was related to an adverse outcome in our study and no ‘U’-shaped relationship was found.
There is evidence that prenatal influences other than maternal nutrition may play an important role. In a recent large-scale study of the influence of maternal stress on human pregnancy outcomes 
, increased maternal stress levels were associated with placental enlargement. Although the extent to which maternal diet was altered by stress was not examined, it is more likely that maternal stress mediators rather than nutrient availability altered placental growth. This raises the interesting possibility that the findings in the present study could be the result of combined effects of maternal stress mediators on both placental growth and the development of brain regions involved in offspring stress responses. However, this is speculative and requires further investigation.
In vertebrates, neuroanatomical asymmetries appear very early in gestation as the result of complex expression of genetic signaling pathways 
. Of particular interest is that many of the genes involved in these pathways are susceptible to epigenetic regulation 
and, thus, to the influence of external environmental regulation. Thus, future studies might investigate epigenetic programming of these genes as a possible explanation for our findings.
Our study has a number of limitations. Our markers of fetal developmental experiences are non-specific and further studies should address specific influences on neurodevelopment in both early and late gestation. Our measure of cerebral activation is indirect, is likely to have limited sensitivity and cannot specify which regions of the brain are involved. Future work with fMRI or PET has the potential to address these issues in selected groups of subjects. Finally, due to the disruptive nature of our measurement technique, we were unable to measure TMT during the TSST-C. Thus, our first measure was several minutes after the termination of stress. It is unknown how rapidly changes in cerebral activation are reflected in TMT changes and it is therefore possible that our findings represent a reduced ability to return RH activation to baseline following acute stress rather than differences in acute RH response to stress. However, the presence of similar albeit weaker associations at rest argues against this. Further work with TMT measurement could explore continuous measures to fully assess the response to stress.
We have provided evidence that early development of the fetus influences the lateralization of cerebral activity, particularly following stress, in early childhood. Risk of depression and enhanced stress responsivity, with their consequent risks for ill health including cardiovascular disease, is increased in people who were small at birth. Our findings suggest that persistent differences in cerebral lateralization determined during gestation offer a potential explanation for this. Brain regions involved in the processing and response to stressful stimuli may be particularly susceptible to these effects.