Background Preterm birth may leave the brain vulnerable to dysfunction. Knowledge of future neurodevelopmental delay in children born with various degrees of prematurity is needed to inform practice and policy. Objective To quantify the long‐term cognitive, motor, behavioural and academic performance of children born with different degrees of prematurity compared with term‐born children. Search strategy PubMed and Embase were searched from January 1980 to December 2016 without language restrictions. Selection criteria Observational studies that reported neurodevelopmental outcomes from 2 years of age in children born preterm compared with a term‐born cohort. Data collection and analysis We pooled individual estimates of standardised mean differences (SMD) and odds ratios (OR) with 95% confidence intervals using a random effects model. Main results We included 74 studies (64 061 children). Preterm children had lower cognitive scores for FSIQ (SMD: −0.70; 95% CI: −0.73 to −0.66), PIQ (SMD: −0.67; 95% CI: −0.73 to −0.60) and VIQ (SMD: −0.53; 95% CI: −0.60 to −0.47). Lower scores for preterm children in motor skills, behaviour, reading, mathematics and spelling were observed at primary school age, and this persisted to secondary school age, except for mathematics. Gestational age at birth accounted for 38–48% of the observed IQ variance. ADHD was diagnosed twice as often in preterm children (OR: 1.6; 95% CI: 1.3–1.8), with a differential effect observed according to the severity of prematurity (I2 = 49.4%, P = 0.03). Conclusions Prematurity of any degree affects the cognitive performance of children born preterm. The poor neurodevelopment persists at various ages of follow up. Parents, educators, healthcare professionals and policy makers need to take into account the additional academic, emotional and behavioural needs of these children. Tweetable abstract Adverse effect of preterm birth on a child's neurodevelopment persists up to adulthood.

The functional selectivity of human primary visual cortex (V1) for orientation and motion direction is established by around 3 months of age [1-3], but there have been few studies of the development of extrastriate visual areas that integrate outputs from V1 [4-8]. We investigated sensitivity and topographical organization for global form and motion with high-density visual event-related potentials (VERPs) in 4- to 5-month-old infants and adults. Responses were measured to transitions between concentrically organized elements (short arc segments for form, dot trajectories for motion) and random arrangements. Adults showed topographically separate responses, with midline motion and more lateral form responses. Of 26 infants, 25 showed significant motion responses but only 13 showed form responses, suggesting more advanced development for extrastriate motion areas than form. Infants' form and motion responses were topographically distinct but contrasted with the corresponding adult topographies, with infants' motion responses more lateral than form responses. These results imply distinct neural sources at both ages and raise the possibility of substantial reorganization of extrastriate networks between infancy and adulthood. We speculate that global motion responses arise from area V5 in infants but are dominated by more medial areas such as V3/V3A and V6 in adults.

During the first 3 months, infants develop visual evoked potential (VEP) responses that are signatures of cortical orientation-selectivity and directional motion selectivity. Orientation-specific cortical responses develop in early infancy. This study compared these responses directly in the same infants, to investigate whether the later appearance of direction selectivity was intrinsic, or a function of the spatio-temporal characteristics of the stimuli used. Steady-state orientation-reversal (OR-) VEPs and direction-reversal (DR-) VEPs were recorded in infants aged 4-18 weeks. DR-VEPs were elicited with random pixel patterns and with gratings spatially similar to those used for OR-VEPs, at velocities of 5.5 and 11 deg/s, and reversal rates of 2 and 4 reversals/s. Infants throughout the age range showed significant responses to orientation-reversal. Direction-reversal responses appeared in less than 25% of infants under 7 weeks of age, rising to 80% or more at 11-13 weeks, whether tested with dots or gratings and for both speeds and reversal rates. However, 2 reversals/s elicits the DR-VEP on average about 2 weeks earlier than 4 reversal/s stimulation. We conclude that human cortical direction selectivity develops separately from orientation-selectivity and emerges at a later age, even with tests that are designed to optimise the former.