It is noteworthy that each of the studies mentioned above described effects in the ~25–45 Hz range, i.e., the lower frequency end of the band traditionally defined as gamma. This ‘slow gamma’ activity synchronizes CA1 with inputs from CA3 [14–17], a region that is believed to play a key role in memory retrieval [18–20]. Thus, it is possible that slow gamma reflects a state in which the entorhinal-hippocampal network is optimally primed to retrieve, rather than encode, memories. If the hippocampus exhibits slow gamma during memory retrieval, then entorhinal inputs may need to synchronize with slow gamma in order for sensory cues to trigger memory retrieval, as occurred in the Igarashi et al. study [12]. These interactions across regions are likely to involve both theta and slow gamma, with slow gamma activating cell assemblies that represent a particular memory within a given theta cycle. Consistent with this notion, co-modulation of theta and ~30–50 Hz slow gamma power was several times higher during spatial memory retrieval than during exploration prior to memory encoding in a study by Shirvalkar and colleagues [21]. Moreover, co-modulation of theta and slow gamma power was higher during successful trials than during error trials.
