Theta oscillations regulate the speed of locomotion via a hippocampus to lateral septum pathway

Abstract

Hippocampal theta oscillations support encoding of an animal's position during spatial navigation, yet longstanding questions about their impact on locomotion remain unanswered. Combining optogenetic control of hippocampal theta oscillations with electrophysiological recordings in mice, we show that hippocampal theta oscillations regulate locomotion. In particular, we demonstrate that their regularity underlies more stable and slower running speeds during exploration. More regular theta oscillations are accompanied by more regular theta-rhythmic spiking output of pyramidal cells. Theta oscillations are coordinated between the hippocampus and its main subcortical output, the lateral septum (LS). Chemo- or optogenetic inhibition of this pathway reveals its necessity for the hippocampal regulation of running speed. Moreover, theta-rhythmic stimulation of LS projections to the lateral hypothalamus replicates the reduction of running speed induced by more regular hippocampal theta oscillations. These results suggest that changes in hippocampal theta synchronization are translated into rapid adjustment of running speed via the LS.

Figure 1. Optogenetic control of hippocampal theta oscillations.

(a) Injections of Cre-dependent ChR2 in MS of PV-Cre mice and light-induced stimulation of MS–Hip projections. Expression of AAV2/1.CAGGS.flex.ChR2.tdTomato.WPRESV40: neuronal somata in MS (1,2), fibre tracts, fimbria (fi)-fornix (f), nucleus of the horizontal limb of the diagonal band (HDB) (3,5,6) and axons in the Hip (4,5). Scale bars, 500 μm (1,3,4) and 50 μm (2,5,6). (b) Left: LFP PSD (colour coded, computed for 10-s epochs) for all control recordings, as well as optostimulation at 7 and 10 Hz (N=9 mice). Power spectra marked with arrows are shown on top. The rows are ordered according to entrainment fidelity (optogenetically entrained theta), or dominant theta frequency (spontaneous theta). Right: cumulative distribution of theta entrainment fidelity for various optostimulation frequencies. (c) Laminar LFP profiles of spontaneous and optogenetically controlled theta oscillations; str. or., stratum oriens; str.rad., stratum radiatum. It is worth noting that the light pulse marked by an arrowhead (at 7 Hz) resets theta phase, thus adjusting the rhythm frequency to the stimulation frequency. (d) Phase–amplitude coupling of theta and gamma oscillations (left) compared across theta amplitudes (right) between spontaneous and optogenetically entrained theta (P=0.42, N=3 mice, 8 and 19 recordings, respectively). Data are presented as mean±s.e.m. (e) Intact bilateral coordination of optogenetically entrained theta oscillations (P<0.05, Pearson's correlation, N=5 mice). Left: signal traces recorded during ipsi- (top) and contralateral (bottom) hippocampal stimulation; right: respective LFP power spectra during optostimulation. (f) Left: histograms of preferred discharge phases of CA1 pyramidal cells (opt. entrainment: blue, 30 neurons; spontaneous theta: black, 29 neurons). Grey shaded bar: a theta phase bin when timestamps of laser pulse were preferentially recorded. Right: histograms of preferred discharge phases of fast firing interneurons (n=28 neurons). Preferred theta phases did not differ (pyramidal cells, P=0.79, Watson–Williams test; fast firing interneurons, P=0.97, Watson U2 permutation test).

Figure 2. Optogenetic entrainment of hippocampal theta oscillations regulates locomotion speed and speed variability.

(a) Correlation of running speed and frequency of theta oscillations during spontaneous (r=0.82, Pearson's correlation, P=0.013, N=6 mice, n=42 recording sessions) but not optogenetically entrained theta oscillations (r=0.04, P=0.7, N=8 mice, n=72 recording sessions). (b) Scheme illustrating a bidirectional influence between theta and locomotion as proposed earlier. Optogenetic control of theta removes the influence of speed-correlated afferents on oscillation frequency (blue cross), as oscillation frequency is controlled by the laser pulse frequency. (c) Optogenetic entrainment of theta oscillations at 7 Hz (P=0.0002, Bonferroni test, N=5 mice) or 9 Hz (P=0.0002, N=3 mice) reduced speed in running mice. (d) Running speed during 10-s epochs of optogenetically entrained (blue) and spontaneous (black) theta oscillations. (e) Example of optogenetically elicited theta oscillations during immobility. Hippocampal LFP signal traces (2–250 Hz band-pass filtered, middle), recorded simultaneously with running speed measurement (red trace, bottom). Blue shade marks time of optostimulation. Excerpt, top: LFP signal trace shown at a higher time resolution. (f) Number of experiments when resting mice stayed immobile (Imm.) or moved (Run) during 15 s after stimulation onset (P=0.13, χ²-test, N=8 mice, optogenetic entrainment; N=5 mice, control light stimulation). (g) Speed (deviation from group mean) averaged for 20 s before and after the onset of optogenetic theta entrainment (left, N=8 mice) or control light stimulation (right, N=13 mice). Speed was less variable during optogenetic theta entrainment compared with baseline (P=0.0062, Bonferroni test, n=18 recording sessions, N=8 mice) and control light stimulation recordings (P=0.0107, n=42 recording sessions, N=13 mice). (h) Representative traces before and after the onset of stimulation (7 Hz) with a low (<0.3; left) or high (>0.8; right) entrainment fidelity; grey shadows mark 10%–90% ranges of speed distributions during stimulation. (i) Higher entrainment fidelity was associated with lower coefficients of variation of theta amplitude (r=−0.84, P=0.0046) and running speed (r=−0.84, Pearson's correlation, P=0.0051, n=79 recording sessions, N=8 mice). *P<0.05, **P<0.01, ***P<0.001. Data are presented as mean±s.e.m.

Figure 3. Variability of theta amplitude mediates speed variability.

(a) Distribution of theta oscillation epochs according to variability of theta cycles amplitude (spontaneous theta, black, n=384 recording epochs, N=8 mice; optogenetic stimulation, blue, n=818 recording epochs, N=8 mice). (b,c) Both in control and during optogenetic theta entrainment, theta amplitude variability predicted changes of speed variability (b, polynomial fit, R²=0.89, spontaneous theta; R²=0.90, optogenetic stimulation) and of running speed (c, R²=0.96, spontaneous theta; R²=0.92, optogenetic stimulation). (d) Changes of theta amplitude correlated with changes of firing probability in CA1 pyramidal cells during spontaneous and optogenetically entrained theta (25 and 12 single units, respectively; polynomial fit, R²=0.90, spontaneous theta; R²=0.88, optogenetic stimulation). Data are presented as mean±s.e.m.

Figure 4. Hip–LS-LH pathway supports theta-rhythmic speed regulation.

(a) Hip–LS–LH pathway. (b) Coordinated theta oscillations in Hip and LS (LFP, 1–200 Hz, grey trace; 5–15 Hz, brown trace, band-pass filtered). (c) Coherence of the Hip and LS LFP during hippocampal theta oscillations (N=3 mice). (d) Histograms of preferred LS theta discharge phases of 21 out of 73 significantly theta-modulated LS units (black line, P<0.05, Rayleigh test). Grey bars, firing probability of an example unit; dotted line, reference oscillation cycle. (e) Top: DREADDs (AAV-CaMKIIa-hM4D(Gi)-mCherry) expression in hippocampal pyramidal cells. Bottom: cannula implantations for CNO/vehicle injections and axonal immunofluorescence in LS. Scale bars, 500 μm (left) and 50 μm (right). (f) Intra-LS CNO, in comparison with intra-LS vehicle, prevented reduction of running speed during optogenetic theta entrainment (P<0.00001, analysis of variance (ANOVA)). Left bars show baseline running speed. (g) Running speed (P<0.00001, ANOVA; see also Supplementary Fig. 6f) and speed variability (P=0.15, N=6 mice) after intra-LS CNO or vehicle during spontaneous theta. (h) Top: eNpHR3.0 (AAV2/1.CamKIIa.eNpHR3.0-EYFP.WPRE.hGH) expression in hippocampal pyramidal cells. Bottom: bilateral optic fibres implantation, axonal immunofluorescence in LS. Scale bars, 500 μm (left) and 50 μm (right). (i) PSD (colour coded) for all recordings where optogenetic theta entrainment at 9 Hz was combined with eNpHR3.0 stimulation in LS. Rows are ordered according to entrainment fidelity. (j) Reduction of speed variability during optogenetic theta entrainment (grey, P=0.0073, ANOVA, N=8 mice) was prevented by simultaneous LS yellow light (593 nm) delivery (yellow bar, baseline white bar). (k,l) Locomotion-dependent firing of LH cells. (k) Top: examples of running onset—triggered rastergrams (50 epochs, 11 units, epochs are ordered according to firing rate); bottom: spike count (grey bars) and average running speed (black line). (l) Changes of firing rate according to running speed (P<0.05 for each cell, Pearson's correlation, 11 cells). (m) Top: injections and expression of Cre-dependent ChETA (AAV2/5.Ef1a.DIO.ChETA(E123T/H134R)-EYFP.WPRE.hGH) in LS in Vgat-Cre mice. Bottom: bilateral fibre implantation and axonal fluorescence in LH. Scale bars, 500 μm (left) and 50 μm (right). (n) Optogenetic theta-frequency activation of LS–LH pathway decreased running speed (P=0.003, Bonferroni test, N=7 mice). Data are presented as mean±s.e.m.