Screening programs for congenital hypothyroidism (CH) dramatically improved the neuropsychological prognosis in affected children. However, mild impairments in cognitive performances, poorer motor skills, defective language abilities, and learning problems have been reported in some studies of early-treated CH children. The occurrence of these defects makes neuropsychological follow-up mandatory. The aim of the present study was to identify those neuropsychological functions that are more frequently affected in early-treated CH children and that might require prompt rehabilitation treatment to prevent permanent defects. The study group involved 24 CH children. Levothyroxine (LT4) treatment (initial dose 8-10 microg/kg per day) was started at mean age of 28 days (range 15-45) and was then adjusted with the goal to keep thyrotropin (TSH) and free thyroid hormone levels in the normal range. Cognitive evaluation was performed at 3, 5, and 7 years of age and did not significantly differ from that of controls. Mean neurological scores were lower in children 5 years of age than in controls. Children with severe neonatal hypothyroidism (serum thyroxine [T4] < 2 microg/dL) had significantly lower neurological scores compared to less affected CH children and normal controls. The most affected functions were balance, extremity coordination, fine motricity, quality of movements, associated movements, and head movements. Language disorders were observed in half of CH children at 3 and 5 years of age, but moderately severe defects were restricted to those with severe neonatal hypothyroidism. In conclusion, a problem-oriented, simplified neuropsychological follow-up of early-treated children with CH should not systematically include the frequent repetition of time-consuming and expensive psychometric tests because individual IQ scores are in the normal range of tests in almost all CH children and can be differentiated from those of normal controls only on a population-statistic basis. Selected tests of motor proficiency are indicated at 3 and 5 years of age to detect those defects in motor skills that appear to be more specifically affected in CH children. Language performances are at particular risk in CH children, and should be always checked at 3 and 5 years of age. Children with even mild language disorders or delayed language achievements should be regularly reevaluated at 6-month intervals and, if no spontaneous improvement is observed, they should receive specific rehabilitation treatment. No further motor and language evaluation is warranted in CH children with normal tests at age 5 years.