One would think that the oscillation regulating sequence reactivation across the hippocampus would be the high frequency (∼150–200 Hz) ripple oscillation that accompanies sharp waves.

However, high-frequency ripples are not correlated between CA3 and CA1 (Csicsvari et al., 1999). This is problematic because reactivation in CA1 requires properly timed input from CA3 (Nakashiba et al., 2009). Moreover, the large majority of replay events include neuronal activity from both CA1 and CA3 (Carr et al., 2012). In this issue of Neuron, Carr et al. (2012) propose a solution GS-7340 in vivo to this problem. Their results indicate that low frequency (“slow,” ∼20–50 Hz) gamma oscillations regulate the precisely timed reactivation of neuronal sequences in CA3 and CA1. They report that SWRs are accompanied by increases in CA3 and CA1 slow gamma activity. In contrast to ripples, SWR-associated slow gamma oscillations occurred synchronously across CA3 and CA1. Moreover, CA3-CA1 slow gamma synchrony was stronger

during SWRs than when no SWRs were present. Concurrent increases in CA3-CA1 synchrony were not seen in other frequency bands. CA3 slow gamma oscillations entrained spiking GDC-0068 cell line of neurons in both CA3 and CA1, and CA3 slow gamma entrainment of CA1 spiking was stronger during SWRs than when no SWRs were present. The new findings by Carr et al. (2012) also imply that slow gamma oscillations in the hippocampus serve as an internal clock during sequence reactivation. The authors measured slow gamma phase intervals between spikes from pairs of place cells. They found that slow gamma phase intervals across successive gamma cycles were significantly correlated with distance between the neurons’ place fields. Considering that distinctive places like cue-containing walls (Hetherington and Shapiro, 1997) and goal locations (Hollup et al., 2001) are heavily represented by place cell activity, the new findings raise the possibility that discrete locations are

reactivated on separate slow gamma cycles. Replay occurring during pauses in exploratory activity matches activation patterns from earlier experiences more accurately than replay occurring during extended periods of quiescence Thalidomide (Karlsson and Frank, 2009). Carr et al. (2012) found that quiescent SWR replay (i.e., relatively low-quality replay) was not associated with increases in slow gamma entrainment of cell spiking, a finding that supports the conclusion that enhanced slow gamma entrainment is necessary for high-fidelity replay. This conclusion received further support from their finding that large increases in CA3-CA1 slow gamma synchrony during SWRs were predictive of high fidelity replay events. Why would slow gamma entrainment of place cell spikes increase during some SWRs (i.e., waking SWRs) but not others (i.e.