Medial septum regulates the hippocampal spatial representation

Abstract

The hippocampal circuitry undergoes attentional modulation by the cholinergic medial septum. However, it is unclear how septal activation regulates the spatial properties of hippocampal neurons. We investigated here what is the functional effect of selective-cholinergic and non-selective septal stimulation on septo-hippocampal system. We show for the first time selective activation of cholinergic cells and their differential network effect in medial septum of freely-behaving transgenic rats. Our data show that depolarization of cholinergic septal neurons evokes frequency-dependent response from the non-cholinergic septal neurons and hippocampal interneurons. Our findings provide vital evidence that cholinergic effect on septo-hippocampal axis is behavior-dependent. During the active behavioral state the activation of septal cholinergic projections is insufficient to evoke significant change in the spiking of the hippocampal neurons. The efficiency of septo-hippocampal processing during active exploration relates to the firing patterns of the non-cholinergic theta-bursting cells. Non-selective septal theta-burst stimulation resets the spiking of hippocampal theta cells, increases theta synchronization, entrains the spiking of hippocampal place cells, and tunes the spatial properties in a timing-dependent manner. The spatial properties are augmented only when the stimulation is applied in the periphery of the place field or 400-650 ms before the animals approached the center of the field. In summary, our data show that selective cholinergic activation triggers a robust network effect in the septo-hippocampal system during inactive behavioral state, whereas the non-cholinergic septal activation regulates hippocampal functional properties during explorative behavior. Together, our findings uncover fast septal modulation on hippocampal network and reveal how septal inputs up-regulate and down-regulate the encoding of spatial representation.

Keywords: hippocampus; medial septum; optogenetics; theta rhythm; transgenic rat.

Figure 1

ChR2-YFP expression in the medial septum of ChAT::Cre rats. (A) Atlas schematic of our experimental setup investigating the functional relation of septal activity to the hippocampal formation (MS, medial septum; HIP, hippocampus). The medial septum of ChAT::Cre rats was injected with cre-inducible ChR2-EYFP. Chronically-implanted headstage with optic fiber and microdrive allowed parallel application of blue laser light and measurement of single unit activity in medial septum. Additionally, the implantation of bipolar concentric electrode allowed electric stimulation. Concurrently, recording tetrodes were implanted in hippocampal area CA1 to measure neuronal and local field activity. Coronal atlas schematic (below) and histological section (right) show where optic fiber and eight tetrodes were implanted and subsequently lowered in medial septum. The black arrow indicates the location of tetrodes tip. (B) Colocalization of ChAT staining and ChR2-YFP expression in the medial septum. High-magnification views of ChR2-YFP expression in ChAT-positive septal cell bodies after injection of cre-dependent virus in the medial septum of ChAT::Cre rats.

Figure 2

Physiology of multiunit optical responses in medial septum of ChAT::Cre rats. (A) Sample scatterplot, showing all signals recorded on a given tetrode. Right: sample waveforms of septal units, corresponding to neurons from the ChAT, inhibition (slow-spiking), inhibition (fast-spiking), potentiation, re-inhibition and re-potentiation groups. (B) Sample size of the main groups of neurons, which responded to optogenetic activation of septal cholinergic neurons. (C) Raster plot from 40 repetitions (above) and spike count of 120 repetitions (below) of optically evoked time-locked responses of ChAT cell. Time 0 indicates the delivery of the first train of the stimulation protocol (50 Hz, 5 ms pulse duration, 12 pulses, 473 nm). (D) Raster plot from 40 repetitions (above) and spike count of 120 repetitions (below) of slow-spiking unit from the inhibition group. (E) Comparison of the firing rate (spikes/s) 250 ms before, 250 ms after and 500 ms after the stimulation protocol for ChAT, slow-spiking inhibition, fast-spiking inhibition, potentiated, re-inhibition and re-potentiation neurons. Error bars represent ± sem, Wilcoxon signed-rank test; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. Multiunit activity in response to optical activation of septal cholinergic neurons: sample raster (above) and spike count (below) plots recordings of fast-spiking inhibition (F), potentiation (G), re-inhibition (H), and re-potentiation (I) units.

Figure 3

Dependence of the septal network response on the behavioral state. (A) Raster plot and spike count of representative ChAT cell during inactive (left) and active (right) behavioral state after 10 Hz stimulation protocol. (B) Raster plot and spike count of representative potentiation cell during inactive (left) and active (right) behavioral state after 50 Hz stimulation protocol. (C) Raster plot and spike count of representative inhibition cell during inactive (left) and active (right) behavioral state after 10 Hz stimulation protocol. (D) Raster plot and spike count of representative re-inhibition cell during inactive (left) and active (right) behavioral state after 50 Hz stimulation protocol. (E) Ratio of the spiking firing rate for inactive over active behavioral state for the ChAT unit shown in (A). (F) Ratio of the spiking firing rate for inactive over active behavioral state for the inhibition unit shown in (C). (G) Ratios of the spiking firing rate for inactive over active behavioral state for baseline background activity (left) and stimulation-induced spiking (right) for ChAT, inhibition, potentiation, re-inhibition, and re-potentiation groups. Error bars represent ± sem, paired t-test; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. The horizontal red dotted line indicates the ratio level of 1.

Figure 4

Hippocampal neuronal responses after septal ChAT stimulation. (A) High-magnification views of septal axons expressing ChR2-YFP in hippocampal CA1 area after injection of cre-dependent virus in the medial septum of ChAT::Cre rats. (B) Raster plot from and spike count of optically evoked responses from hippocampal re-potentiation unit. Time 0 indicates the delivery of the first train of the stimulation protocol (50 Hz, 5 ms pulse duration, 12 pulses, 473 nm). (C) Average values of the stimulation-evoked firing rate increase (as percent of the baseline firing rate) for the potentiation and re-potentiation hippocampal units for stimulation frequency vs. behavioral state: 50 Hz inactive, 50 Hz active, 10 Hz inactive and 10 Hz active, respectively. Error bars represent ± sem, Wilcoxon signed-rank test; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. Raster plots from and spike count of optically evoked responses from hippocampal inhibition (D) and potentiation (E) units. (F) Ratio of the firing rate for inactive over active behavioral state for baseline background activity (left) and stimulation-induced spiking (right) for hippocampal interneurons. (G) Ratio of the firing rate for stimulation over baseline for 250 ms (left) and 500 ms (right) after septal stimulation for hippocampal place cells. Error bars represent ± sem, paired t-test; ^*P < 0.05.

Figure 5

Hippocampal local field responses after septal ChAT stimulation. (A) Sample event related potentials (ERP) recorded in dorsal CA1 after 10 Hz septal optogenetic stimulation during inactive (left) and active (right) behavioral state. Upper red traces show ERP from medial septum, middle green traces represent ERP from hippocampus band-pass filtered (4–15 Hz) and lower blue traces represent the same hippocampal ERP after low-pass filtered (0–15 Hz). Time 0 indicates the delivery of the first train of 10 Hz stimulation protocol to medial septum. (B) Color-coded power spectrograms of hippocampal low-frequency oscillations after 10 Hz septal stimulation protocol during inactive (left) and active (right) behavioral state. Vertical dotted black line indicates time 0. (C) Representative samples of phase-locking value for 10 Hz inactive (top, left), 50 Hz inactive (top, right), 10 Hz active (below, left), and 50 Hz active (below, right) state. Blue traces show the observed data, while the green values represent shuffled data. (D) Average values of the phase-locking value for the same groups. Error bars represent ± sem, two-tailed t-test; ^*P < 0.05, ^**P < 0.01. (E) Frequency histogram of band-passed local field potential for 10 Hz (red), 50 Hz (blue) and baseline epoch counts (green) during inactive behavioral states. (F) Frequency histogram of band-passed local field potential for 10 Hz, 50 Hz and baseline epoch counts during active behavioral states. Error bars represent ± sem, Newman–Keuls test, ^*P < 0.05.

Figure 6

Theta-burst stimulation of medial septum resets the spiking of hippocampal interneurons and theta oscillations. (A) Raster plot and spike count of representative re-potentiation hippocampal cell after electric theta-burst stimulation protocol (TBS, four bursts with inter-train interval of 100 Hz and inter-train interval of 8 Hz) to medial septum. Time 0 indicates the delivery of the first train. The vertical arrows indicate the delivery time of four trains. (B) Sample of 1000 ms autocorrelogram and spike waveform (right) of hippocampal theta unit. (C) Averaged frequency histogram for all theta cells. The vertical arrows indicate the delivery time of four trains. The horizontal arrow indicates the post-stimulation period. Black error bars represent ± sem. (D) Averaged firing frequency from −60 ms before to 60 ms after the first train for re-potentiated (theta), potentiated, and inhibited units, respectively. Error bars represent ± sem, Mann–Whitney test; ^*P < 0.05, ^***P < 0.001. (E) Left: Sample band-pass filtered and low-pass filtered ERP recorded in dorsal CA1 after septal TBS. Right: Sample color-coded power spectrogram of the same recording. (F) Power histogram of hippocampal theta power before (blue) and after (red) TBS. Error bars represent ± sem, Two-Way ANOVA, ^*P < 0.05.

Figure 7

Septal TBS effect on hippocampal place cells. (A) Sample place field maps of CA1 pyramidal neurons after exploration of rectangular-shaped linear track. The non-stimulated place cells (blue, purple, and green) are denoted as controls, while the target place cell (yellow) is stimulated with septal TBS (red). TBS pulses are positioned on each side of the field depending on the direction (clockwise and counter-clockwise) of the animal in the track. Bottom: spike waveforms and spike clusters of the same place cells. (B) Sample of periphery-stimulated place cell (cell #1); Left: map of animal trajectory with spikes (green) and TBS pulses (red). Firing rate maps during baseline (middle) and TBS session (right). Below: sample recording of the place cell spikes (green) and the preceding TBS pulses (red). (C) Sample of center-stimulated place cell (cell #2); map of animal trajectory with spikes (green) and TBS pulses (red). Note that the TBS pulses targeted the center of the place field in the course of clockwise direction. Firing rate maps during baseline (middle) and TBS session (right). Below: sample recording of the place cell spikes (green) and TBS pulses (red). (D) Scheme of place field binning for clockwise direction. (E) Comparison of the place field properties between TBS applied in the center of the place field (−3 to 0), in the periphery of the place field (−9 to −6), or intermediately (−6 to −3). Center rate, mean firing frequency, place field size and centroid difference (E), and spatial coherence and spatial information content (F) are represented as ratios of the measured values from baseline session over the values of the TBS session. (G) Average intra-field firing rate per 3 cm bins during baseline (left) and TBS (right) sessions, for periphery-stimulated (green) and center-stimulated (blue) place fields. Error bars represent ± sem, Newman–Keuls test; ^*P < 0.05, ^** P < 0.01.