Furthermore, cross-frequency coherence analysis demonstrated that gamma and theta oscillations were nested (Figure 4C), as reported previously (Bragin et al., 1995, their Figure 1). Finally, LFP power in the theta range was reduced by thermoinactivation of the ipsilateral entorhinal cortex (Figure S5), consistent with the results of previous lesion experiments (Bragin et al., 1995). We then examined the relations between EPSC and LFP signals (Figures 4D–4F). Analysis of the coherence

between the two signals revealed a high level of coherence in the theta frequency range but a low coherence in the gamma frequency Estrogen antagonist band (Figure 4E). The main peak had a mean coherence of 0.40 ± 0.04, corresponding to a frequency of 6.2 ± 1.0 Hz (13 cells in awake rats; Figure 4D). Control analysis with shuffled data showed that the coherence was significant (p < 0.05; Figure 4E). Furthermore, phase analysis demonstrated that EPSCs were significantly phase locked to theta cycles of the LFP (p < 0.005). The angular lag for the theta activity was –39° ± 12° (321°; 13 cells), implying that EPSCs coincided with the descending phases and the troughs of the theta oscillations (Figure 4F). Consistent with these results, EPSC power spectra showed a peak at theta frequency, with a maximum at 4.3 ± 0.3 Hz in anesthetized rats and 6.1 ± 0.4 Hz in awake rats (15 and 13 cells,

respectively; Figures Quisinostat molecular weight S6A and S6B; see Klausberger et al., 2003). Furthermore, autocorrelation analysis of EPSC traces revealed regular peaks at a mean period of 204.6 ± 27.8 ms in anesthetized and 179.1 ± 18.8 ms in awake rats (Figures S6C and S6D). Taken together, these results indicate that EPSCs represent a global synaptic input signal, which is spatially coherent over the dentate gyrus and mainly operates at theta frequency. While the high coherence in the theta frequency range Resminostat is consistent with the excitation model, the lower coherence in the gamma frequency range

seems inconsistent with this model. If EPSCs are strongly theta coherent but only weakly gamma coherent, what are the synaptic mechanisms underlying gamma oscillations in the dentate gyrus (Bragin et al., 1995)? To address this question, we recorded IPSCs in GCs and examined the coherence with the LFP in awake rats (Figure 5). IPSCs were isolated under voltage-clamp conditions at a holding potential of 0 mV, close to the reversal potential of AMPAR-mediated currents. Surprisingly, the frequency dependence of coherence of IPSCs was markedly different from that of EPSCs. Analysis of coherence between the IPSC signal and the LFP indicated a highly significant peak in the gamma frequency range (five cells in awake rats; p < 0.05). The main peak had a mean coherence of 0.35 ± 0.07, corresponding to a frequency of 76.2 ± 5.2 Hz. Additional coherence peaks of lower amplitude were present in both the theta (coherence 0.14 ± 0.03, frequency 3.4 ± 0.4 Hz) and the supragamma frequency ranges (coherence 0.26 ± 0.