The Institute of Medicine (IOM) recently revised gestational weight gain (GWG) guidelines, yet little is known about the association between extremes of GWG and child cognition.The authors examined this association using U.S. Collaborative Perinatal Project data (1959-73). Women with a calculable GWG and term, singleton children were eligible for inclusion. Children (31 968 individuals, including 8704 siblings) were assessed at 4 (Stanford-Binet Intelligence Scale, Graham-Ernhart Block Sort Test) and 7 years of age [Wechsler Intelligence Scales for Children, Wide Range Achievement Test (WRAT)]. Multiple regression with restricted cubic splines and generalized estimating equations (GEEs) was used. Fixed-effects (FE) models were also applied to control for confounders shared within sibships.GWG displayed non-linear inverted U-shaped associations with most outcomes when the data were analysed using GEE, suggesting a detrimental influence of low and high GWG on cognition. After controlling for familial factors (FE), all associations were null except for a positive linear association between GWG above the IOM-recommended ranges and WRAT spelling (β = 1.41, 95% confidence interval: 0.13, 2.70).Familial factors are important confounders of the association between GWG and cognitive development at 4 and 7 years of age. After controlling for such factors, GWG was generally unassociated with child development.