Researchers in the Netherlands have reported that the associations of prenatal polychlorinated biphenyl (PCB) exposure with poorer cognitive performance in childhood are generally stronger and statistically significant only among non-breast-fed children. This paper reports a similar pattern of greater vulnerability among non-breast-fed children in the Michigan cohort of children whose mothers had eaten PCB-contaminated Lake Michigan fish. It is difficult to determine from prospective human longitudinal data whether the apparent decreased vulnerability among breast-fed children is due to nutrients in breast milk that might attenuate the adverse neurochemical effects of this exposure and/or the more optimal intellectual stimulation that was provided by the breast-feeding mothers in these cohorts. The latter interpretation was supported by analyses demonstrating that virtually all of the positive associations between breast-feeding and cognitive outcome can be accounted for statistically by measures of quality of parental intellectual input and that the adverse effects associated with prenatal PCB exposure were also markedly stronger in children raised by less verbally competent mothers, who presumably provided less optimal intellectual stimulation. Because there were no consistent patterns of differential vulnerability for either gender, these data did not support the hypothesis that prenatal PCB exposure may disrupt cognitive development due to the estrogenic or antiestrogenic properties of its congeners and metabolites.