Feedforward inhibition mediated by GABA_A receptors has been shown to regulate the timing of postsynaptic action potentials in layer 4 cells of rodent somatosensory (barrel) cortex (Gabernet et al. 2005). However, the role of immature GABAergic transmission, which is widely reported to be depolarising (Ben-Ari, 2002), has not been investigated. We have used gramicidin perforated patch-clamp recordings from layer 4 cells in mouse thalamocortical (TC) slices to measure the reversal potential of pharmacologically isolated GABA_A-mediated currents during development from postnatal day (P) 3-P9. The reversal potential over this period changed from -53 ± 3 mV (P3, n = 3) to -71 ± 2 mV (P9, n = 3). We also used whole cell recordings to measure the relative contributions of GABA_A and glutamate receptors to TC synaptic currents over the same age range. The contribution of GABA_A receptors increased greatly with age (GABA_A charge transfer as a percentage of glutamate receptor-mediated charge: P4 = 10 ± 6%, n = 10; P9 = 343 ± 93%, n = 8, p < 0.001 (unpaired t test). In the same cells the latency of the peak GABA_A-mediated synaptic response decreased (latency from stimulus: P4 = 47 ± 30 ms, P9= 5 ± 1 ms, p < 1 x 10-4), as did the variability in this latency (S.D. of latency: P4 = 22 ± 6 ms, P9 = 0.9 ± 0.4 ms; p < 1 x 10^-5). Thus GABAergic transmission shows a profound change in its properties during the first postnatal week. To investigate the consequences of this, we studied the effects of feedforward GABAergic transmission on action potentials evoked by TC input using perforated patch recordings. In P3-5 animals application of the GABA_A antagonist, bicuculline methobromide (BMB) had no effect on the probability or latency of action potential firing (probability: control 34 ± 5%, BMB 34 ± 10%, n = 7, p = 1; latency: control 12 ± 2, BMB 14 ± 2 ms, n = 5, p = 0.3). In P8-9 animals application of BMB caused an increase in firing probability (control = 31 ± 7%, BMB = 89 ± 6% , n = 14, p < 1 x 10-5 paired t test) but had no effect on action potential timing (latency: control = 5 ± 1 ms, BMB = 6 ± 1 ms, n = 10, p = 0.5; jitter (S.D. of latency): control = 2 ± 1 ms, BMB = 2 ± 1 ms, n = 10, p = 0.8). However, when stimulus intensity was decreased in the presence of BMB to return firing probability to baseline values a robust change in action potential timing was observed (latency: control = 5 ± 1 ms, BMB = 10 ± 1 ms, n = 5; p < 0.01, jitter: control = 1 ± 1 ms, BMB = 4 ± 1 ms, n = 5, p < 0.05). Our findings demonstrate a dramatic functional switch in the role of GABAergic input to layer IV barrel cortex during development. Moreover, our data indicate that mature, hyperpolarising GABAergic transmission, rather than causing high timing fidelity of action potential timing per se, also restricts action potential firing only to synaptic input of sufficient amplitude to result in high timing precision.