Fast Gamma Rhythms in the Hippocampus Promote Encoding of Novel Object-Place Pairings

Abstract

Hippocampal gamma rhythms increase during mnemonic operations (Johnson and Redish, 2007; Montgomery and Buzsáki, 2007; Sederberg et al., 2007; Jutras et al., 2009; Trimper et al., 2014) and may affect memory encoding by coordinating activity of neurons that code related information (Jensen and Lisman, 2005). Here, a hippocampal-dependent, object-place association task (Clark et al., 2000; Broadbent et al., 2004; Eacott and Norman, 2004; Lee et al., 2005; Winters et al., 2008; Barker and Warburton, 2011) was used in rats to investigate how slow and fast gamma rhythms in the hippocampus relate to encoding of memories for novel object-place associations. In novel object tasks, the degree of hippocampal dependence has been reported to vary depending on the type of novelty (Eichenbaum et al., 2007; Winters et al., 2008). Therefore, gamma activity was examined during three novelty conditions: a novel object presented in a location where a familiar object had been (NO), a familiar object presented in a location where no object had been (NL), and a novel object presented in a location where no object had been (NO+NL). The strongest and most consistent effects were observed for fast gamma rhythms during the NO+NL condition. Fast gamma power, CA3-CA1 phase synchrony, and phase-locking of place cell spikes increased during exploration of novel, compared to familiar, object-place associations. Additionally, place cell spiking during exploration of novel object-place pairings was increased when fast gamma rhythms were present. These results suggest that fast gamma rhythms promote encoding of memories for novel object-place associations.

Keywords: CA1; gamma oscillations; gamma rhythms; hippocampus; memory; place cells.

Figure 1.

Verification of target recording sites and behavioral effects in object–place association task. A, Histologic sections showing example recording sites in CA1 and CA3. B, A schematic explaining the object–place association task is shown. The behavioral task consisted of 3 familiarization days (F; Object A) and 3 d in which novel object–place pairings were presented. The novel-object place pairings included an object identity and location change (NO+NL; Object C), a location change only (NL; Object A′), and an object identity change only (NO; Object B). Each day consisted of three 10 min exploration sessions (S1, S2, S3) separated by 10 min rest periods, and the order of the conditions was randomly assigned for each animal. C, The discrimination index for the familiarization and novelty conditions, as well as control conditions in which no objects were presented. Grey dashed line indicates chance level. For NO+NL conditions, rats explored the novel object–place pairings significantly more than the familiar object–place pairings and significantly more than they explored the same locations when no objects were present. Because novel object–place pairings were presented in the second session, familiarization measures were also computed using the second session of familiarization or re-familiarization days (F) in this figure and all subsequent figures. D, The amount of time rats spent exploring familiar object–place pairings in Session 1 (S1) of the familiarization condition and the different novelty conditions. E, The amount of time rats spent exploring familiar (light blue bars; Object A indicated in white text) and novel (dark blue bars; Objects C, A′, and B indicated in white text) object–place pairings in Session 2 (S2) of the familiarization condition and the different novelty conditions. *p < 0.05, **p < 0.01. Data are presented as mean ± SEM in this figure and all subsequent figures.

Figure 2.

Changes in slow and fast gamma power in CA1 in response to exploration of novel object–place pairings. A–C, Color-coded power across gamma frequencies in CA1 as a function of running speed, plotted during time periods of familiarity exploration versus novelty exploration, averaged across all CA1 tetrodes and rats. The time periods of exploration of familiar object–place pairing A in F conditions were time-matched with those during exploration of familiar object–place pairing A (top row) and novel object–place pairings C, A′, and B (bottom row) in NO+NL (A), NL (B), and NO (C) conditions, respectively. Note that x- and y-axes are shown in log scale. D, Changes in fast and slow gamma power between time-matched periods in the F condition and the three novelty conditions (NO+NL, NL, and NO), during exploration of familiar (A) and novel (ie, C, A′, and B) object–place pairings. Data from individual rats are shown in gray. *Indicates significantly (p < 0.05) different changes in gamma power from familiarization session to novelty session for exploration of novel object–place pairings compared to exploration of familiar object–place pairings; # and ## indicate that the change in gamma power between N and F sessions was significantly (#p < 0.05, ## p < 0.01) greater than zero.

Figure 3.

No significant changes in slow and fast gamma power in CA3 during exploration of novel object–place pairings. A–C, Same as in Figure 2A-C , except for CA3 recordings instead of CA1. D, Changes in fast and slow gamma power in CA3 between time-matched periods in the F condition and the three novelty conditions (NO+NL, NL, and NO) during exploration of familiar (A) and novel (ie, C, A′, and B) object–place pairings. Data from individual rats are shown in gray.

Figure 4.

Changes in slow and fast gamma phase synchrony between CA3 and CA1 during exploration of novel object–place pairings. The difference in CA3–CA1 slow and fast gamma phase synchrony between exploration periods for novel and familiar object–place pairings in NO+NL (A), NL (B), and NO (C) conditions. The differences in slow and fast gamma interregional phase synchrony between the explorations periods for the two familiar object–place pairings in the F condition are also shown (D). Data from individual rats are shown in gray. **p < 0.01.

Figure 5.

Phase-locking of CA3 and CA1 place cell spikes to CA1 slow and fast gamma during exploration of novel object–place pairings. A–C, Mean vector lengths of CA1 slow and fast gamma phase distributions were estimated for spike times of CA3 and CA1 place cells with place fields close to either familiar or novel object–place pairings in the novelty conditions. For the NO+NL condition, place cell spikes were significantly more phase-locked to fast gamma during exploration of the novel object–place pairing than during exploration of the familiar object–place pairing. D, Mean vector lengths of CA1 slow and fast gamma phase distributions were estimated for spike times of CA3 and CA1 place cells with fields near either of the familiar object–place pairings in the familiar condition. E, Example spike time-gamma phase distributions from individual place cells. Spike counts were normalized (ie, number of spikes in each bin/total spike count). A representative place cell from each cell category is shown for fast gamma (top row, red) and slow gamma (bottom row, blue). Grey lines indicate moving average (moving size = 2 bins). **p < 0.01.

Figure 6.

CA1 place cell spiking increased selectively in fast gamma periods during exploration of novel object–place pairings. A–C, Examples of color-coded rate maps of CA1 place cells that exhibited place fields close to the novel object–place pairs in the NO+NL (A), NL (B), and NO (C) conditions. Red indicates peak firing rate, dark blue represents no firing, and white pixels indicate unvisited areas. Rate maps constructed from spikes across the entire exploration session are shown in the left columns. Rate maps constructed from spike times during slow and fast gamma episodes are shown in the middle and right columns, respectively. Black dots indicate the defined place fields. Each map is shown scaled to the peak firing rate of the cell across the entire session, which is shown to the left. D, Mean in-field firing rates of CA1 place cells during slow and fast gamma episodes that occurred during exploration of novel or familiar object–place pairings. In these plots, slow and fast gamma episodes were detected from the same tetrodes on which the cells were recorded. E, The same as D, except that slow and fast gamma were detected using different tetrodes than the ones on which cells were recorded. *p < 0.05, **p < 0.01.

Figure 7.

Changes in theta power, CA3–CA1 phase synchrony, and place cell firing patterns during exploration of novel object–place pairings. A–C, Color-coded theta power in CA1 (top rows) and CA3 (bottom rows) as a function of running speed during exploration of familiar and novel object–place pairings, averaged across all recordings for each region. As in Figures 2 and 3, the familiar object–place pair exploration periods in the F condition (first and third columns) were time-matched with those during exploration of familiar object–place pairs (second column) and novel object–place pairs (fourth column) in NO+NL (A), NL (B), and NO (C) conditions. D, E, No significant changes in CA1 (D) and CA3 (E) theta power occurred between time-matched periods in the F condition and the three novelty conditions (NO+NL, NL, and NO) during exploration of familiar and novel object–place pairings. Data from individual rats are shown in gray. F, CA3–CA1 theta phase synchrony did not significantly change between novel and familiar object–place pair exploration in NO+NL, NL, and NO conditions, nor between explorations of the two familiar object–place pairs in the F condition. Data from individual rats are shown in gray. G, Mean vector lengths of CA1 theta phase distributions for CA3 and CA1 place cell spike times in novelty and familiarization conditions. Theta phase-locking was higher during novel object exploration compared to familiar object exploration for the NO+NL condition. *p < 0.05.