To narrow down the pool of candidates, we first applied a moderately low concentration of Ba2+ (200–250 μM, for 10 min in the bath but not in the puffer pipette) that should have no significant effect on delayed rectifier, Ca2+-activated and transient D-type K+ currents, or the hyperpolarization-activated cation channel Ih but blocks inward rectifier K+ (IRK) currents and partially inhibits transient A-type (Gasparini et al., 2007 and Losonczy et al., 2008) and certain “leak” K+ currents (Coetzee et al., 1999). Fast NMDA spikes were unanimously prolonged by Birinapant mouse 200–250 μM Ba2+, rendering them slow (half-width ctr: 50.5 ± 3.5 ms, Ba2+: 114.7 ± 12.6 ms, n = 6, p < 0.05, Wilcoxon test, Figure 6A, see also Figures

S4A and S4B for effect of K+ channel modulators on Vm and input resistance), whereas the kinetics of slow NMDA spikes was not affected by the same treatment (half-width ctr: 90.8 ± 7.4 ms, Ba2+: 106.4 ± 5.3 ms, n = 5, p = 0.138, Wilcoxon test). As a time control, repeated activation of NMDA spikes in 10 min induced only a slight, nonsignificant increase in the half-width of fast NMDA spikes (half-width ctr: 47.6 ± 4.4 ms, LY294002 after 10 min: 55.0 ± 5.7 ms, n = 8, p = 0.092, Wilcoxon test). Transient A-type K+ currents regulate dendritic Na+ spikes and are partially inhibited by 200–250 μM Ba2+ in CA1PCs (Gasparini et al., 2007 and Losonczy

et al., 2008). To more specifically test the involvement of A-type current in the regulation of NMDA spike decay in CA3PCs, we applied 4-aminopyridine (2 mM 4-AP +1 μM TTX to control network activity, see Experimental Procedures). 4-AP applied in the presence of TTX prolonged the half-width of fast NMDA spikes (TTX only: 52.7 ± 3.2 ms, TTX + 4-AP: 71.7 ± 6.5 ms, n = 8, p < 0.05, Wilcoxon test, Figures 6B and S4), but the effect was nearly half that of Mephenoxalone 200–250 μM Ba2+ (fractional change: 4-AP: 1.37 ± 0.12; 200–250 μM Ba2+: 2.29 ± 0.24, p < 0.05, Mann-Whitney tests with Bonferroni correction for fractional change comparisons

after Kruskal-Wallis test, p < 0.01), indicating that transient voltage-gated K+ currents play only a partial role in the regulation of this parameter. This result is consistent with the notion that the majority of transient K+ channels are expected to inactivate during the slow time course of NMDA spikes. Among K+ currents, inward rectifiers (IRKs) have the highest sensitivity to external Ba2+ (Hibino et al., 2010). We next tested a lower concentration of Ba2+ (30 μM), which should completely block IRK channels, while having only a minimal effect on other types of K+ currents. We found that 30 μM Ba2+ reversibly increased the half-width of fast NMDA spikes to a similar degree as that measured with 200–250 μM Ba2+ (Figures 6C and S4, control: 51.1 ± 1.3 ms, Ba2+: 106.0 ± 17.0 ms, n = 7, p < 0.05, Wilcoxon test; comparison of fractional change with 200–250 μM Ba2+: p = 0.