Cognitive performance in 7- to 9-year-old preterm neonatal intensive-care survivors was compared with that in age-matched control children. Non-verbal memory span, spatial working-memory abilities, planning, set-shifting, and recognition memory for both spatial and patterned stimuli were assessed using the Cambridge Neuropsychological Testing Automated Battery. Relative to children in the control group, neonatal intensive-care unit (NICU) survivors demonstrated 25% more memory errors on the spatial working-memory task. Their use of strategy on this task was similar to a control group of 5-year-olds. Planning times on 'Tower of London' problems were long relative to those of term controls. NICU survivors demonstrated poorer pattern recognition as well as a shorter spatial memory span. The groups did not differ in visual-discrimination learning or in spatial-recognition memory. No specific neonatal risk factor accounted for the observed differences, although scores on the Neurobiological Risk Score (NBRS), a composite measure of neonatal risk, did predict several aspects of later task performance. Whether these data reflect a developmental delay in brain maturation in NICU survivors or the presence of a permanent information-processing deficit due to adverse neonatal events must be assessed through continued follow-up.