Epidemiological studies and case reports show that even a relatively minor degree of maternal hypothyroxinemia during the first half of gestation is potentially dangerous for optimal fetal neurodevelopment. Our experimental approach was designed to result in a mild and transient period of maternal hypothyroxinemia at the beginning of corticogenesis. Normal rat dams received the goitrogen 2-mercapto-1-methyl-imidazole for only 3 d, from embryonic d 12 (E12) to E15. Maternal thyroid hormones decreased transiently to 70% of normal serum values, without clinical signs of hypothyroidism. Dams were injected daily with 5-bromo-2'-deoxyuridine (BrdU) during 3 d, from E14-E16 or E17-E19. Their pups were tested for audiogenic seizure susceptibility 39 d after birth (P39) and killed at P40. Cells that had incorporated BrdU were identified by immunocytochemistry, and quantified: numerous heterotopic cells were found, whether labeled at E14-E16 or E17-E19, that were identified as neurons. The cytoarchitecture and the radial distribution of BrdU-labeled neurons was significantly affected in the somatosensory cortex and hippocampus of 83% of the pups. The radial distribution of gamma-aminobutyric acidergic neurons was, however, normal. The infusion of dams with T4 between E13 and E15 avoided these alterations, which were not prevented when the T4 infusion was delayed to E15-E18. In total, 52% of the pups born to the goitrogen-treated dams responded to an acoustic stimulus with wild runs, followed in some by seizures. When extrapolated to man, these results stress the need for prevention of hypothyroxinemia before midpregnancy, however moderate, and whichever the underlying cause.