Maturation of excitatory synaptic connections depends on the amount and pattern of their activity, and activity can affect development of inhibitory synapses as well. In the superficial visual layers of the superior colliculus (sSC), developmental increases in the effectiveness of gamma-aminobutyric acid (GABA(A)) receptor-mediated inhibition may be driven by the maturation of visual inputs. In the rat sSC, GABA(A) receptor currents significantly jump in amplitude between postnatal days 17 and 18 (P17 and P18), approximately when the effects of cortical inputs are first detected in collicular neurons. We manipulated the development of these currents in vivo by implanting a drug-infused slice of the ethylene-vinyl acetate copolymer Elvax over the superior colliculus of P8 rats to chronically release from this plastic low levels of N-methyl-D-aspartate (NMDA). Sham-treated control animals received a similar implant containing only the solvent for NMDA. To examine the effects of this treatment on the development of GABA-mediated neurotransmission, we used whole cell voltage-clamp recording of spontaneous synaptic currents (sPSCs) from sSC neurons in untreated, NMDA-treated, and sham-treated superior colliculus slices ranging in age from 10 to 20 days postnatal. Both amplitude and frequency of sPSCs were studied at holding potentials of +50 mV in the presence and absence of the GABA(A) receptor antagonist, bicuculline methiodide (BMI). The normal developmental increase in GABA(A) receptor currents occurred on schedule (P18) in sham-treated sSC, but NMDA treatment caused premature up-regulation (P12). The average sPSCs in early NMDA-treated neurons were significantly larger than in age-matched sham controls or in age-matched, untreated neurons. No differences in average sPSC amplitudes across treatments or ages were present in BMI-insensitive, predominantly glutamatergic synaptic currents of the same neurons. NMDA treatment also significantly increased levels of glutamate decarboxylase (GAD), measured by quantitative western blotting with staining at P13 and P19. Cell counting using the dissector method for MAP 2 and GAD(67) at P13 and P19 indicated that the differences in GABAergic transmission were not due to increases in the proportion of inhibitory to excitatory neurons after NMDA treatment. However, chronic treatments begun at P8 with Elvax containing both NMDA and BMI significantly decreased total neuron density at P19 ( approximately 15%), suggesting that the NMDA-induced increase in GABA(A) receptor currents may protect against excitotoxicity.