Hippocampal Network Dynamics during Rearing Episodes

Abstract

Animals build a model of their surroundings on the basis of information gathered during exploration. Rearing on the hindlimbs changes the vantage point of the animal, increasing the sampled area of the environment. This environmental knowledge is suggested to be integrated into a cognitive map stored by the hippocampus. Previous studies have found that damage to the hippocampus impairs rearing. Here, we characterize the operational state of the hippocampus during rearing episodes. We observe an increase of theta frequency paralleled by a sink in the dentate gyrus and a prominent theta-modulated fast gamma transient in the middle molecular layer. On the descending phase of rearing, a decrease of theta power is detected. Place cells stop firing during rearing, while a different subset of putative pyramidal cells is activated. Our results suggest that the hippocampus switches to a different operational state during rearing, possibly to update spatial representation with information from distant sources.

Keywords: CA1; cross-frequency coupling; dentate gyrus; gamma oscillation; independent component analysis; motion capture; navigation; place cell; silicone probe; theta rhythm.

Graphical abstract

Figure 1

Rearing Events Are Accompanied by Theta Oscillation Alterations (A) Arrangement of the 3D behavioral observation system. Four cameras recorded the position of markers on the mouse’s head implant. The motion capture software reconstructed the exact position and orientation of the animal’s head in real time. (B) Sample trajectory from an animal in the open arena during one trial (top right, side views, bottom left, top view). Red circles indicate rearing events. (C) Example rearing event: black, vertical head position; gray, head pitch angle; green vertical line, start; red vertical line, peak; blue vertical line, stop of rearing event. Inset shows the average of n = 132 rearing events of one animal in black; the overlaid individual events are in gray. (D) Example LFP recording in the CA1 pyramidal cell layer (bottom trace) during one rearing episode (vertical head positions, middle trace) and corresponding wavelet spectrogram of the LFP (top). Note the elevated frequency of theta oscillation during the rearing event. (E) Rearing peak-triggered average vertical head position (bottom) and the corresponding average wavelet spectrogram (top) of all rearing events of an animal. (F) Theta oscillation frequency (top diagram) and power (bottom) as a function of running speed and vertical head position. Note the high theta frequency values at low running speed and high head position. (G) Mean power in different frequency bands around rearing events calculated for all mice. A prominent increment of high theta power is paralleled by the reduction of low theta power. Brown traces indicate the shuffled control (n = 12 animals; shaded gray area corresponds to SEM; beige area represents the 99% band of the shuffled control). (H) Timing of changes in the delta and theta bands during rearing. Horizontal bars indicate the duration of the LFP and behavioral events; error bars represent SEM. Red, power increment; blue, power reduction. Note that power alterations of high and low theta tended to precede the start of head lift-up but nearly coincided with change of pitch angle (^∗p < 0.05, paired t test; n = 12 animals). See also Figure S1.

Figure 2

Enhancement of the DG Sink during Rearing (A) Histological image shows the linear silicone probe track through all layers of the hippocampus. Oriens, stratum oriens; pyr, pyramidal layer; rad, radiatum; lac-mol, lacunosum-moleculare; mol, molecular layer; gran, granule cell layer; poly, polymorphic layer. (B) Perievent binned (bin size = 1 s) average power in the low and high theta bands in the main layers of the hippocampal formation. Brown traces indicate the shuffled control (n = 4 animals; shaded area represents the 99% band of the shuffled control). Note the high theta power elevations in the lacunosum-moleculare and in the hilus of the DG. (C) CSD maps from an animal of averaged LFP profile during running (left) and rearing (right). White traces at the top indicate theta oscillation in the pyramidal layer. (D) Comparison of current sink magnitude during running and rearing (^∗p < 0.05, ANOVA with post hoc Dunnett’s test; n = 4 animals; values are mean ± SEM). Note the prominent increase of dentate gyrus sink during rearing compared to running. Pr, pyramidal layer; Rd, radiatum; LM, lacunosum-moleculare; DG, dentate gyrus; Hl, hilus. See also Figure S2.

Figure 3

Robust Enhancement of Theta-Gamma Coupling in the Dentate Gyrus during Rearing Events (A) Anatomical sketch that depicts the main inputs of CA1 pyramidal and dentate granule cells. Sch, Schaffer; TA, temporoammonic; LPP, lateral perforant; MPP, medial perforant pathways. (B) CSD wavelet spectrograms indicate the characteristic gamma bands and their theta phase along the hippocampal formation. Pr, pyramidal layer; Rd, radiatum; LM, lacunosum-moleculare; DG, dentate gyrus. (C) Independent-component analysis (ICA) of hippocampal LFPs resulted in five main components (ICs), three restricted to the CA1 area and two to the dentate gyrus. Top row: voltage loadings of the five identified LFP generators in the hippocampal formation (from left to right: CA1pyr, CA1 pyramidal layer; rad, radiatum; LM, lacunosum-moleculare; LPP, lateral perforant path; MPP, medial performant path generators; from top to bottom: Pr, pyramidal layer; LM, lacunosum-moleculare; GCL, granule cell layer). Bottom row: CSD loadings of the same generators. In the dentate gyrus one component had maximal negative amplitude in the outer molecular layer (LPP) and the other (MPP) in the middle of it and both had positive polarity through the granular layer and the hilus. (D) Average cross-frequency-amplitude comodulograms for the five ICs during running (averages of n = 5 animals). (E) Average theta phase-gamma amplitude coupling matrices for the five ICs during running (averages of n = 5 animals). Two theta cycles are shown for visualization purposes. Top dashed trace indicates theta oscillation in the CA1 pyramidal layer. Note the opposite theta-phase preference of LM and MPP gamma oscillations, matching the phase distribution of entorhinal layer II and III spiking and local gamma oscillations (Fernández-Ruiz et al., 2017). (D) and (E) were calculated from IC segments during which the animal movement speed was above 0.04 m/s for at least 2 s. (F) Fast gamma in the dentate gyrus. First row: filtered (5–12 Hz) CA1 pyramidal layer LFP; second row: 120–150 Hz filtered lacunosum-moleculare LFP; third row: 120–150 Hz filtered dentate gyrus LFP; bottom row: wavelet spectrogram of the dentate gyrus LFP. Red arrows highlight fast gamma transients on the dentate gyrus recordings and corresponding wavelet spectrogram. Vertical dashed lines indicate the pyramidal layer theta troughs. Fast gamma components were not detected in the lacunosum-moleculare LFP. (G) Average comodulograms of the five ICs for different time intervals around rearing peak (averages of n = 5 animals). Plots were constructed by averaging individual comodulograms from five animals followed by the subtraction of baseline (−10 to −5 s to rearing peak) comodulograms. Red rectangle highlights comodulograms at rearing peak (0 s). Note the robust enhancement of the theta coupling of the medial perforant path generator (MPP). (H) Average perievent modulation index values of the five ICs (rows) in three different gamma frequency bands (columns; averages of n = 5 animals, brown traces indicate the shuffled control, shaded area represents the 99% band of the shuffled control). Vertical dashed lines at zero indicates rearing peak. The dentate fast gamma (MPP generator) has broad frequency range appearing in both 50–100 Hz and 120–150 Hz frequency bands. (I) The coupling of LM mid gamma and MPP dentate fast gamma to theta oscillation increases with running speed. However, rearing-coupled MPP theta-gamma coupling is about two times larger than that is observed during running (^∗p < 0.05, paired t test; n = 5 animals; values are mean ± SEM). See also Figure S3.

Figure 4

Single-Unit Activity during Rearing Behavior (A) Rearing-on units and place cells show contrasting rearing-dependent activity. Bottom row: vertical head position during rearing events. Middle row: blue vertical lines, sample place cell spikes. Top row: black vertical lines, sample rearing-on unit spikes. The rearing-on unit increased and the place cell decreased activity during rearing epochs. (B) Rearing peak-triggered perievent firing histograms corresponding to the units in A. Left: rearing-on unit; middle: place cell infield (within place field); right: place cell outfield (out of place field). (C) Spatial firing maps of a rearing-on (top) and place cell (bottom, black dots indicate the infield, white dots the outfield rearing events). (D) Rearing-on units and place cells have different firing preference for vertical head position and horizontal speed. Two-dimensional histograms show the horizontal speed and vertical head position dependence of a rearing-on unit (left top) and a place cell (left bottom). Units are the same as in (C). Horizontal gray dashed lines mark the threshold of rearing events. Average vertical head position (right top, overlaid histograms) and horizontal speed distributions (right bottom, overlaid histograms) highlight the differences between rearing-on units (gray) and place cells (blue). Vertical head positions are measured form the rearing threshold. (E) Rearing peak-triggered perievent firing histograms of rearing-on units (above horizontal dashed line) and infield perievent histograms for place cells (below horizontal dashed line). The Z-scored perievent histograms were ordered by their average value between −0.25 and +0.25 s around rearing peaks. (F) Average rearing peak-triggered perievent histograms of rearing-on units and place cells: rearing-on units (top; n = 11 units), place cells infield (middle; n = 48 units), place cells outfield (bottom; n = 48 units). Green traces indicate the shuffled data. Shaded area around the histogram bars corresponds to SEM. (G) Summarized data showing the firing rate during running and rearing for rearing-on units and place cells. In the case of place cells, infield and outfield firing data are shown separately (^∗p < 0.05, Wilcoxon signed-rank test; n = 11 rearing-on units, n = 48 place cells; error bars represent SEM). See also Figure S4.