, 2006; Campbell and Gillin, 1987; Gierz et al., 1987; Walker, 2010).

Neurons in the hippocampal cortex display distinct firing patterns during different behaviors (O’Keefe, 2007). Waking exploration and REM sleep are characterized by theta oscillations and neural firing “episodes” in which individual cells sustain elevated firing rates for several hundreds of milliseconds (Buzsáki, 2002; Louie and Wilson, 2001; Montgomery et al., 2008). HIF inhibitor In contrast, during immobility and non-REM sleep, hippocampal neural firing is concentrated in short (∼120 ms) sharp-wave ripple events, which synchronize activity across much of the network and have been suggested to reflect reactivation of learned firing patterns (Buzsáki, 1989; Wilson and McNaughton, 1994). Between ripples, neural firing is sparse and asynchronous for RG7204 supplier hundreds of milliseconds (Buzsáki et al., 1992; Csicsvari et al., 1999; Sullivan et al., 2011; cf. Carr et al., 2011). During sleep, hippocampal ripples are weakly correlated with neocortical slow oscillations (Steriade et al., 1993), although hippocampal activity is often dissociated from that of the neocortex (Hahn et al., 2007; Wolansky et al., 2006; Isomura et al., 2006). We examined the evolution of population firing patterns in the CA1 hippocampal region during sleep. Our findings show that discharge rates of both pyramidal cells and interneurons gradually ramp up during non-REM episodes,

interrupted by larger rate decreases during the interleaving REM epochs. This “sawtooth” pattern of rate changes across non-REM and REM episodes results in an overall downscaling of discharge rates over the course of sleep. In contrast, synchrony during non-REM ripple events increases from the early to late stages of sleep. The concurrent decrease of firing rates and increased population synchrony from one non-REM episode to the next are correlated with the power of theta oscillations during the intervening REM sleep. Our findings, therefore, suggest

a central role of REM sleep in regulating both discharge rates and synchrony in the hippocampus. Local field potentials (LFPs) and spiking activity of isolated CA1 putative pyramidal cells and putative interneurons were recorded Non-specific serine/threonine protein kinase in the home cage while the rat was immobile and assumed a characteristic sleep posture. The ratio of theta (5–11 Hz) and delta (1–4 Hz) power was used to identify non-REM and REM episodes (Figure 1A; see Supplemental Experimental Procedures available online), as described previously (Montgomery et al., 2008). Twenty-two sleep sessions (38.2 ± 5.8 min, SEM) with at least one non-REM-REM-non-REM cycle were recorded in five rats. Mean firing rates of pyramidal cells (n = 618) were similar between non-REM and REM episodes, whereas firing rates of interneurons (n = 111) were significantly higher during REM (p < 0.00018; sign-rank test; Csicsvari et al., 1999).