Extreme preterm infants are at risk for developmental problems, because of perinatal difficulties. Stimulation at home may be important in alleviating such developmental problems. Developmental course from birth to 512 years of a group of initially 38 infants born before 32 weeks gestational age, was studied in a multidisciplinary, longitudinal study. The preterm group was examined at ages corrected for prematurity and compared to a group of 35 term reference infants on repeated measures of growth and health, neurological and motor development, cognitive and socio-emotional development. The children are now examined again at the age of 10 years.Over the first 512 years the preterm and the reference group most clearly differed in cognitive development, with more preterms repeatedly showing delay. Concerning the amount of stimulation at home only a few group differences were found. Mother-infant interaction was observed during a feeding when the infants were one month old, during play at nine months and during play in the morning as well as dinner with the fanily in the evening at 18 months. Only one facet of maternal behavior differed the groups during the feeding (mothers of preterms talked more to their children) and during play at nine months (reference mothers played more structured children's games). At 18 months no group differences were found in parental behavior during the evening dinner; during play in the morning mothers of preterms were found to stimulate their infants less (verbally, by the use of toys or by demands or restrictions). The HOME interview, which was done at the age of 6 and 18 months, only differed the groups in the dimension ‘involvement’ at 18 months, which was greater for the mothers of preterms.For the preterm group the HOME total score and the amount of verbal stimulation of the mothers during play at 18 months correlated positively with results of cognitive assessments from 18 months until 512 years. In cognitive development gradually different pathways became clear. The preterm infants could be divided into a subgroup (GOOD) of 19 children who repeatedly had cognitive assessments fitting their age levels, and a subgroup of 17 children who frequently had results lower than one standard deviation below the norm (SUSPECT/BAD) at the Bayley mental scales at 6, 12, 18, 24 and 30 months, the SON nonverbal intelligence test at 312, the Reynell language scales at 4, and the RAKIT inrelligence test at 412 and 512 years. For these preterm subgroups GOOD and SUSPECT/BAD a different pattern of correlations is found between the amount of stimulation at home and the infants' cognitive development. In the subgroup GOOD the children's results at 18 and 24 months correlate positively with the amount of their mothers' verbal stimulation and the HOME total score, whereas in the subgroup SUSPECT/BAD such correlations are not found until the age of 312 and 4 years.To conclude: our results suggest that stimulation at home measured when the infants were 18 months of age, is important for the extreme preterm infants' mental development. Those preterms who manifest cognitive delay, may not be able to show this before preschool age.