Aim: To examine the correlation between neonatal cranial ultrasound and school age magnetic resonance imaging (MRI) and neurodevelopmental outcome. Methods: In a prospective 2 year cohort study, 221 children (gestational age ≤ 32 weeks and/or birth weight ≤ 1500 g) participated at a median age of 8.1 years (inclusion percentage 78% ). Conventional MRI, IQ (subtests of the WISC), and motor performance (Movement Assessment Battery for Children) at school age were primary outcome measurements. Results: Overall, there was poor correspondence between ultrasound group classifications and MRI group classifications, except for the severe group (over 70% agreement). There was only a 1% chance of the children with a normal cranial ultrasound having a major lesion on MRI. Mean IQ (standard deviation) was significantly lower in children with major ultrasound or MRI lesions, but was also lower in children with minor lesions on MRI compared to children with a normal MRI (91 ± 16, 100 ± 13, 104 ± 13 for major lesions, minor lesions, and normal MRI, respectively). Median total impairment score (TIS) was significantly higher in children with major lesions on ultrasound or MRI as well as in children with minor lesions on MRI (TIS 4.0 and 6.25 for normal and minor lesions on MRI, respectively; p < 0.0001). Conclusions: A normal neonatal cranial ultrasound excluded a severe lesion on MRI in 99% of cases. MRI correlated more strongly with mean IQ and median TIS than ultrasound. Subtle white matter lesions are better detected with MRI which could explain the stronger correlation of MRI with IQ and motor performance. Aim: To examine the correlation between neonatal cranial ultrasound and school age magnetic resonance imaging (MRI) and neurodevelopmental outcome. Methods: In a prospective 2-year cohort study, 221 children (gestational age ≤ 32 weeks and / or birth weight ≤ 1500 g) Results of a median age of 8.1 years (inclusion percentage 78%). Conventional MRI, IQ (subtests of the WISC), and motor performance (Movement Assessment Battery for Children) at school age were primary outcome measurements. Results: Overall, there was poor correspondence between ultrasound group classifications and MRI group classifications, except for the severe group (over 70% agreement). There was only a 1% chance of the children with a normal cranial ultrasound having a major lesion on MRI. Mean IQ (standard deviation ) was significantly lower in children with major ultrasound or MRI lesions, but was also lower lower in children with minor lesions on MRI compared to children with a normal MRI (91 ± 16, 100 ± 13, 104 ± 13 for major lesions, minor lesions, and normal MRI, respectively). Median total impairment score (TIS) was significantly higher in children with major lesions on ultrasound or MRI as well as in children with minor lesions on MRI (TIS 4.0 and 6.25 for normal and minor lesions on MRI, respectively; p <0.0001). Conclusions: A normal neonatal cranial ultrasound excluded a severe lesion on MRI in 99% of cases. MRI correlated more strongly with mean IQ and median TIS than ultrasound. Subtle white matter lesions are better detected with MRI which could explain the stronger correlation of MRI with IQ and motor performance.