OBJECTIVE: An association between thyroid function during pregnancy or infancy and neurodevelopment in children has been demonstrated. We aimed to investigate whether newborn TSH concentrations are related to subsequent neurocognitive development. DESIGN: We conducted a longitudinal study on 178 children from a general population birth cohort in Granada (Spain) born in 2000-2002. METHODS: TSH concentrations were measured in umbilical cord blood, and cognitive functions were assessed at 4 years of age using the McCarthy's scales of children's abilities (MSCA). Organochlorine (OC) compound concentrations and the combined oestrogenicity (total effective xeno-oestrogenic burden (TEXB)) were also determined in the placentae. RESULTS: Mean newborn TSH was 3.55  mU/l (range=0.24-17 mU/l). In multivariate regression analyses, adjusting for maternal and child characteristics, higher newborn TSH concentrations showed a decrease of 3.51 and 3.15 points on the MSCA general cognitive and executive function scores respectively and were associated with a higher risk of scoring below the 20th percentile (P20) on the quantitative score (odds ratio (OR)=2.64). Children with TSH in the upper quartile (4.19-17.0 mU/l) were at higher risk of scoring <P20 on span memory (OR=5.73), whereas children with TSH in the second quartile (2.05-2.95 mU/l) were at lower risk of scoring <P20 on the verbal scale (OR=0.24). Neonatal TSH status was also associated with general cognitive and executive function outcomes when controlling for prenatal exposure to OCs or placental TEXB. CONCLUSIONS: Newborn thyroid hormone status expressed by TSH in cord blood may adversely affect later cognitive function. A more thorough screening for neonatal thyroid deficiency is warranted.