Analyze the relation of gestational diabetes and maternal blood glucose levels to early cognitive functions in the first two years of life. In a prospective Singaporean birth cohort study, pregnant women were screened for gestational diabetes at 26-28 weeks gestation using a 75-g oral glucose tolerance test. Four hundred and seventy three children (n = 74 and n = 399 born to mothers with and without gestational diabetes respectively) underwent neurocognitive assessments at 6, 18, and/or 24 month, including electrophysiology during an attentional task and behavioral measures of attention, memory and cognition. Gestational diabetes is related to left hemisphere EPmax amplitude differences (oddball versus standard) at both six (P = 0.039) and eighteen months (P = 0.039), with mean amplitudes suggesting offspring of mothers with gestational diabetes exhibit greater neuronal activity to standard stimuli and less to oddball stimuli. Associations between 2-hour maternal glucose levels and the difference in EPmax amplitude were marginal at 6 months adjusted β = -0.19 (95% CI: -0.42 to +0.04) μV, P = 0.100 and significant at 18 months adjusted β = -0.27 (95% CI: -0.49 to -0.06) μV, P = 0.014, and the EPmax amplitude difference (oddball-standard) associated with the Bayley Scales of Infant and toddler Development-III cognitive score at 24 months β = 0.598 (95% CI: 0.158 to 1.038), P = 0.008. Gestational diabetes and maternal blood glucose levels are associated with offspring neuronal activity during an attentional task at both six and eighteen months. Such electrophysiological differences are likely functionally important, having been previously linked to attention problems later in life.