Aversive stimulus-tuned responses in the CA1 of the dorsal hippocampus

Abstract

Throughout life animals inevitably encounter unforeseen threatening events. Activity of principal cells in the hippocampus is tuned for locations and for salient stimuli in the animals' environment thus forming a map known to be pivotal for guiding behavior. Here, we explored if a code of threatening stimuli exists in the CA1 region of the dorsal hippocampus of mice by recording neuronal response to aversive stimuli delivered at changing locations. We have discovered a rapidly emerging, location independent response to innoxious aversive stimuli composed of the coordinated activation of subgroups of pyramidal cells and connected interneurons. Activated pyramidal cells had higher basal firing rate, more probably participated in ripples, targeted more interneurons than place cells and many of them lacked place fields. We also detected aversive stimulus-coupled assemblies dominated by the activated neurons. Notably, these assemblies could be observed even before the delivery of the first aversive event. Finally, we uncovered the systematic shift of the spatial code from the aversive to, surprisingly, the reward location during the fearful stimulus. Our results uncovered components of the dorsal CA1 circuit possibly key for re-sculpting the spatial map in response to abrupt aversive events.

Fig. 1. Aversive air puff (AP)-triggered responses of dorsal CA1 pyramidal cells.

a Schematics of the recording configuration. Linearized belt indicates the different AP locations during AP epochs. Fluorescent micrograph of a sagittal section showing the tracks of the silicon probe. Mol, molecular layer of the dentate gyrus; Or, stratum oriens; Rad, stratum radiatum; S, subiculum. Scidraw. (2020). mouse running. Zenodo. 10.5281/zenodo.3925913. b Representative location-lap number raster (upper plot) and mean tuning curves in no stimulation and AP epochs (lower plots) of an AP-activated pyramidal cell. White arrowheads and violet bars mark AP locations. Light grey dashed lines separate consecutive no stimulation and AP (4) epochs. c Peri AP firing histograms of putative pyramidal cells from all session (n = 851 units, 18 sessions). d Normalized spike number around the AP stimulation calculated in the 1^st, 2^nd and 3^rd AP epochs. Light (pre) and dark (post) grey lines indicate the control epochs, violet line represents the AP epochs, shaded area corresponds to s.e.m. Red line indicates significant differences in spike count between control and AP epochs (n = 54, 12 sessions, Wilcoxon signed rank test, p < 0.05). e Percentages of AP responsive putative pyramidal cells. aAP, air puff activated; nAP, non-AP responsive; iAP, decreased activity upon air puff stimulation; PC, place cell; nPC, non-place cell. f Two representative examples demonstrating the onset of air puff responses (bin size 20 ms). Inset, bin size 5 ms. Note the remarkably rapid response onset of the second aAP-Pyr cell. g Distribution of the duration (left panel) and latency (right panel) of AP-evoked activation. Latency was quantified as time from air puff TTL onset to the 1^st spike in the 1^st significant bin at logarithmic timescale. h Firing rate during quiet wakefulness (Q) and movement periods (M). Circles represent individual neurons. 2-way repeated measures ANOVA: Factor A (quiet-mov) F(1779) = 75.94, p < 10⁻¹⁷, Factor B (cell groups) F(3779) = 28.65, p = 1.4*10⁻¹⁷, interactions F(3779) = 13.69, p = 1.03*10⁻⁸. Post hoc Tukey test: all cell groups are significantly different from each other except of the aAP-nPC vs nAP-nPC. Box and whiskers correspond to median, quartile and 10-90% range. Note, that we broke the y axis for better visualization. Source data are provided as a Source Data file.

Fig. 2. Interneuron response to aversive air puff (AP) stimuli.

a Peri AP firing histograms of putative interneurons from all sessions (n = 216, 18 sessions). b Representative location-lap number raster of interneurons with increasing (upper panel, 2 AP epoch) and decreasing activity (lower panel, 4 AP epoch), in response to AP. Arrowheads mark AP locations. Dashed lines separate consecutive no-stimulation and air puff epochs. c Firing histograms of interneurons, with activation (upper panel) and suppression of activity (lower panel) in response to AP (bin size 20 ms). Inset, bin size 5 ms. d Distribution of the duration of AP-evoked activation. e Distribution of the latency of air puff-evoked activation quantified as time from air puff TTL onset to the 1^st spike in the 1^st significant bin at logarithmic timescale. f Percentage of putative interneurons with different response types. aAP-IN, increasing activity to air puff (n = 55); iAP-IN, decreasing activity to air puff (n = 68); aiAP-IN, biphasic response to air puff (n = 15); nAP-IN – no response upon air puff stimulation (n = 78). g Theta phase preference of air puff responsive putative interneurons. aAP-IN, increasing activity to air puff (1.3°; 0.2); iAP-IN, decreasing activity to air puff (10.4°; 0.3); aiAP-IN, biphasic response to air puff (356.7°; 0.1); nAP-IN – no response upon air puff stimulation (2.4°; 0.4). (mean angle, circular variance). h Proportion of air puff responsive and non-responsive interneurons in theta peak and theta trough phase groups of interneurons. i Representative example of air puff-activated pyramidal cells converging on an air puff-activated putative interneuron. On each subpanel, peri-air puff firing histogram of the respective neuron is presented. j Summary data about the number of postsynaptic interneurons targeted by air puff-activated PC cells (aAP-PC, n = 87), air puff-activated non-place cells (aAP-nPC, n = 48), non-air puff responsive place cells (nAP- PC, n = 364), and nonair puff responsive, non-PC (nAP-nPC, n = 339). Kruskal-Wallis test H(3) = 124.53, p < 0.05, Dunn-Holland-Wolfe post hoc test: aAP-PC vs aAP-nPC p < 0.05, aAP-PC vs nAP-PC p < 0.05, aAP-PC vs nAP-nPC p < 0.05, aAP-nPC vs nAP-PC p > 0.05, aAP-nPC vs nAP-nPC p < 0.05, nAP-PC vs nAP-nPC p < 0.05. Box and whiskers correspond to median, quartile and 10–90% range. Source data are provided as a Source Data file.

Fig. 3. Amount of spatial information differentially modulated the effect of aversive stimulation on pyramidal cells versus interneurons.

a Peri AP firing histograms of putative pyramidal cells (n = 361 from 5 sessions) on predisc sphere (pre-Sph), disc and postdisc sphere (post-Sph). Average peri-AP firing histograms of significantly activated and suppressed pyramidal cells are shown below the histograms, shaded areas correspond to s.e.m. b Proportion of AP-activated pyramidal cells (upper panel, 5 session, mean ± SD) on pre-sphere, disc and postsphere. Pre-sphere: 0.31 ± 0.086, disc: 0.43 ± 0.10, postsphere: 0.23 ± 0.06; one-way-ANOVA F(2,12) = 6.6, *p = 0.012, Tukey’s post hoc test: disc vs. pre-sphere: p = 0,11, disc vs. postsphere: p = 0.009. Suppressed pyramidal cells on pre-sphere: 0.007 ± 0.01, disc: 0.12 ± 0.07, postsphere: 0.010.016; one-way-ANOVA F(2,12) = 13.3, p = 0.0009, Tukey’s post hoc: disc vs presphere: **p = 0.002, disc vs post-sphere: **p = 0.002. Lower panel: Cumulative distribution of pyramidal cell response magnitudes on pre-sphere (black line), disc (violet) and postsphere (grey). Inset highlights the increased probability of AP-evoked suppression in pyramidal cells’ activity on the disc. Kolmogorov-Smirnov test pre-Sph vs disc, **p = 0.003; disc vs post-Sph **p = 6*10⁻⁶ c Distribution of the duration of AP-evoked activation of pyramidal cells on pre-sphere (black, upper panel), disc (violet, middle panel) and postsphere (grey, lower panel). For visualization purposes time axis was truncated at 2 s also at (f) panel. d Peri AP firing histograms of putative interneurons (n = 88 from 5 sessions) on presphere (left), disc and post-sphere (right). Average peri AP firing histograms of significantly increased and suppressed interneurons are shown below the histograms, shaded areas correspond to s.e.m. e Proportion of air puff-activated interneurons (upper panel, 5 session, mean ± SD) on presphere: 0.50 ± 0.12, disc: 0.35 ± 0.15 and postsphere: 0.38 ± 0.10; one-way-ANOVA F(2,12) = 2.2, p = 0.15. Suppressed interneurons on presphere: 0.10 ± 0.11, disc: 0.36 ± 0.17, postsphere: 0.09 ± 0.10; one-way-ANOVA F(2,12) = 6.8, p = 0.01, Tukey’s post hoc: disc vs pre-sphere: *p = 0.021, disc vs postsphere: p = 0.018. Lower panel: Cumulative distribution of air puff-responsive putative interneurons’ response magnitude on presphere (black line), disc (green) and postsphere (grey). Kolmogorov-Smirnov test pre-Sph vs disc, **p = 0.0006; disc vs. post-Sph ** p = 0.0007 (f) Distribution of air puff-response durations of interneurons on presphere (black, upper panel), disc (green) and post-sphere (grey, lower panel). Source data of panel b and e are provided as a Source Data file.

Fig. 4. Selectivity of air puff-responsive pyramidal cells to the type and saliency of the stimulus.

a Peri air puff (AP, n = 336, 5 session, left panel), peri tail shock (TS, n = 336, 5 session, middle panel) and peri tone (TN, n = 273, 4 session, right panel) firing histograms of putative pyramidal cells and average histograms of significantly activated cells. Shaded areas correspond to s.e.m. b Representative responses of two cells, cell1 responded to air puff and not to tail shock, cell2 responded to tail shock and not to air puff. c Percentages of AP and TS responsive putative pyramidal cells (n = 336, 5 session). nTS-aAP-Pyr, n = 61 air puff-activated but tail shock non-responsive; TS-aAP-Pyr, n = 26 both TS and AP activated putative pyramidal cell; TS-nAP-Pyr, n = 80 tail shock responsive but air puff nonresponsive putative pyramidal cell; nTS-nAP-Pyr n = 169 nonresponsive cell to either tail shock or air puff. d Representative location-lap raster (upper plot) and mean tuning curves during no stimulation and stimulation epochs (lower plots) of a reward (RW) cell. Blue bars mark reward locations, arrowhead and violet bars mark air puff locations. White dashed lines separate consecutive control and air puff (4) epochs. Peri-RW firing histograms in control and air puff epochs are plotted below the location-lap number raster. e Average peri-reward (1^st reward upper panel, 2^nd reward middle panel) and peri-air puff (lower panel) firing histograms of putative pyramidal cells exhibiting significant activation before (n = 5 from 10 sessions, left panel) or after reward delivery (n = 5 from 10 sessions, right panel). Shaded areas correspond to s.e.m. Markanday, Akshay. (2020). Reward Water Drop. Zenodo. 10.5281/zenodo.3925935.

Fig. 5. Phase precession and ripple participation of aAP- and nAP-Pyr cells.

a Theta phase precession of two representative aAP-Pyr cells (left panels) and two representative place cells (right panels). Dots mark individual spikes. b Paired plots showing individual phase ranges of theta phase precession for aAP-Pyr cells exhibiting phase presession (left panel) and place cells (right panel). Left and right sides of the diagrams are corresponding to start and stop phase distributions, respectively. Two consecutive theta cycles are presented for illustrative purposes. Left side represents start phase and right side represents stop phase. Start phase: aAP-Pyr cells’ vs. place cells’ 29.6 ± 34.1° vs. 146.9 ± 32.2°, Watson-Williams-test p < 10⁻³. Stop phase: a-AP-Pyr cells’ vs. place cells’ 283.8 ± 28.3° vs. 326.1 ± 30.9°, Watson-Williams-test p = 7.8 ×10−⁴. c Phase range of aAP-Pyr cells exhibiting phase precession (n = 46, aAP) and nAP-PC cells (n = 130, PC). Phase range, aAP cells: 126.2 ± 9.0° vs. place cells: 183.4 ± 7.2° Wilcoxon-Mann-Whitney-test **p = 1.38×10⁻⁵. d Cumulative distribution of ripple participation of aAP-Pyr cells (violet line) and air puff-nonresponsive pyramidal cells (grey line). Kolmogorov-Smirnov test: p = 0.004. e Summarized ripple participation of aAP-PC cells (n = 82), aAP-nPC cells (n = 46), nAP-PC (n = 349) and nAP-nPC cells (n = 306), Kruskal-Wallis test H(3) = 145.7, p < 0.005, There is significant difference between each group except of aAP-nPC vs. nAP-PC with Dunn-Holland-Wolfe post hoc test. f Spike number ratio (total number of spikes during ripple / spike numbers during quiet wakefulness) of an aAP-PC cells (n = 82), aAP-nPC cells (n = 46), nAP-PC (n = 349) and nAP-nPC cells (n = 300), Kruskal-Wallis test H(3) = 17.6, p < 0.005. Post hoc Dunn-Holland-Wolfe test, significant differences are labeled by *, note that for better visualization y axis was broken. On panels c,e,f box and whiskers correspond to median, quartile (25−75%) and 10–90% range. Source data of panel c,e and f are provided as a Source Data file.

Fig. 6. Aversive air puffs suppress place cell spiking within collocating place fields and augments out of field activity while triggers remapping.

a Peri air puff z-scored firing histograms of putative pyramidal cells if air puff was given inside place field (Infield, left panel, n = 133) or outside place field (Outfield, right panel, n = 133). The color bar between the firing histograms indicates the identity of the corresponding cell: red, aAP, air puff-activated place cell; green, nAP-PC, air puff-non-responsive place cell; blue, iAP, air puff-inhibited place cell. Below the PSTH plots are the averages of the corresponding matrices, shaded error bars correspond to s.e.m. b Z-scored outfield air puff responses as a function of infield air puff responses of putative pyramidal cells (upper panel). Circles represent putative pyramidal cells. Red line marks z-score value of 2 (2 SD). c Summarized data showing z-scored infield and outfield air puff responses (infield: 0.11 ± 1.63, outfield: 0.28 ± 1.1 Wilcoxon signed rank test **p = 0.0005, n = 133 units; middle, box and whiskers correspond to median, quartile and 10-90% range). d Place field distances from the location of air puff stimulus before the stimulus (Pre) versus after the stimulus (Post). Place fields along the diagonal correspond to stable place fields, off diagonal circles are shifted place fields (remapped). Circles along the x axis are the vanishing and along the y axis are the emerging place fields. Air puff responses at different locations are pooled together. Grey histograms show the distribution of emerging, vanishing fields, violet histograms correspond to stabile and remapping place fields. e Correlations of place maps (n = 364 nAP-PC, 3 AP locations) calculated between first and second halves of pre stimulation (Pre), air puff stimulation (AP) and post stimulation epochs (Post). Kruskal-Wallis test H(5) = 126.6, p < 0.005. Dunn-Holland-Wolfe: there are significant differences: * AP-AP vs Pre-Pre; Post-Post and Pre-Post vs Pre-Pre or Post-Post. Box and whiskers correspond to median, quartile and 10-90% range. Source data of panel b-e are provided as a Source Data file.

Fig. 7. Air puff selective assembly patterns.

a Representative assembly reactivation location – lap number raster with place (left panel) and air puff preference (right panel, from same session as example neuron on Fig. 1b). Arrowhead mark air puff locations. White dashed lines separate consecutive control and air puff (4) epochs. Assembly activation strength are normalized by the total bin count. b Summary data of the number of aAP-Pyr cells (left panel, 3 ± 1.4 vs 0.7 ± 0.9, p**=1.7 ×10−⁶) and weight of aAP-Pyr cells (right panel, 2.6 ± 1.1 vs 0.9 ± 0.4, p**=3.4 ×10−⁷) in air puff assemblies (AP-a, n = 7) and in non-air puff assemblies (nAP-a, n = 84), for testing significance Wilcoxon-Mann-Whitney test was used. c Representative air puff-assembly identified by using the whole session (left panel) or only the non-air puff epochs (right panel). Weights of simultaneously recorded units in the two assemblies arranged in descending order. aAP-Pyr cells are color coded in violet. d Similarity of a representative place (upper panel) and air puff assembly (lower panel) to a synthetic air puff assembly. Histograms correspond to similarity index distributions generated by shuffling unit IDs. Red line indicates similarity threshold (1%) and filled circles mark similarity index of the example place (red dot) and air puff assembly (violet dot). e Weight of AP-Pyr cells and non-AP-Pyr cells in air puff assemblies (n = 4 from 4 sessions) detected in pre- and post-air puff sessions on the polystyrene ball and in the non-air puff epochs of the air puff session on the disc. For testing significance 2-way repeated measures ANOVA was used. Factor A (aAP – nAP Pyrs), F(1,6) = 0,006, p = 0.94; for Factor B (Pre, Disc and Post) F(2,12) = 5.778, *p = 0.017; Interaction F(2,12) = 1,21, p = 0.33. There was no significant difference with Tukey post hoc test between Pre, Disc and Post conditions. On panels b,e box and whiskers correspond to median, quartile (25%−75%) and 10–90% range. Source data of panel b,e are provided as a Source Data file.

Fig. 8. Reactivation of air puff selective assemblies.

a Occurrence of a representative air puff assembly identified by using both AP and non-AP epochs (violet, upper panel) during the session and similar air puff assembly identified by opting out air puff epochs (grey, lower panel) from the same session. Violet rhombi indicate air puff stimulation. For visualization purposes, assembly activation during air puff epochs was removed from the lower panel. b Assembly activation strength before air puff stimulus (Pre) between air puff stimuli (3 air puff stimuli, Inter Event Interval 1 and 2 (IEI)) and after air puff stimuli (Post). Violet circles mark air puff assemblies (n = 6 from 6 sessions) and grey circles correspond to non-air puff assemblies (n = 84 from 6 sessions). (2-way repeated measures ANOVA: Factor A (AP-a – nAP-a), F(1,43) = 0,627, p = 0.433; for Factor B (Pre, IEIs and Post) F(3,129) = 8.679, p = 2.74*10⁻⁵; Interaction F(3,129) = 4,784, p = 0.003). There are significant differences between Pre AP-a strength and Post AP-a strength with Tukey post hoc test, *p = 0.021. c Representative air puff assembly reactivation (grey, middle panel) in comparison to the occurrence of ripples (red, upper panel) and speed (orange, lower panel). d Reactivation occurrence of air puff-assemblies (AP-a, n = 7 from 6 sessions) and non-air puff assemblies (n = 84 from 6 sessions) during movement. For testing significance Wilcoxon-Mann-Whitney two-sample rank-sum test was used, p > 0.05. e Ratio of assembly reactivation strength of air puff and non-air puff assemblies during immobility versus movement of air puff and non-air puff assemblies (AP-a, 2.36 ± 1.2, n = 6 from 6 sessions; nAP-a, 1.49 ± 1.26, n = 84 from 6 sessions). For testing significance Wilcoxon-Mann-Whitney two-sample rank-sum test was used, *p = 0.012. f Ripple-triggered average of normalized assembly reactivation strength for both air puff- and non-air puff assemblies, shaded area corresponds to s.e.m. Inset shows the statistically similar normalized reactivation strength of air puff- and non-air puff assemblies in the ±10 ms peri-ripple peak window (AP-a, 0.032 ± 0.025, n = 7 from 6 sessions; nAP-a, 0.030 ± 0.02, n = 84 from 6 sessions Wilcoxon-Mann-Whitney two-sample rank-sum test, p = 0.73). On panels b,d,e and inset of f, box and whiskers correspond to median, quartile and 10-90% range. Source data of panel b,d,e are provided as a Source Data file.

Fig. 9. Air puff disrupts the population place code and causes its systematic shift.

a Real versus decoded positions in control (left panel) and air puff (right panel) epochs of a representative session. b Decoding error in control (black) and air puff (violet) epochs of all sessions used for position estimation (n = 9 sessions). Red indicates significant difference (paired t-test, p < 0.05). c Probability of estimating the position of mice to be at the reward location in pre-air puff control and first inter-air puff (blue traces upper and lower left plot, respectively) and first and second air puff epochs (violet traces, upper and lower right plot, respectively). Grey area corresponds to repeating 1000 times the estimation of random locations outside the reward zone. Light blue circles mark reward locations. Light violet line corresponds to probability without aAP-Pyr cells. Insets right to the air puff-epoch estimations zoom in on the −5 cm to +10 cm segment around the start location of air puff-delivery to better show the significantly higher probability of locating the animal to the reward zone following the air puff. Mean probability of decoding to the reward location in the 0–35 cm from air puff delivery: 0.09 ± 0.14 in 1^st no stimulation epoch vs. 0.28 ± 0.13 in 1^st air puff epoch, Wilcoxon signed rank test: *p = 0.01; 0.11 ± 0.09 in 2^nd no stimulation epoch vs. 0.25 ± 0.08 in 2^nd air puff epoch, Wilcoxon signed rank test **p = 0.008. d Z-scored averaged tuning curve of putative pyramidal cells firing during reward coding theta cycles from a 3 sec segment after air puff stimulation. Pyramidal cells were sorted into quintiles based on within cycle firing rate and their tuning curves were averaged in each quintile (shaded areas correspond to s.e.m.). Peri-reward elevation of place field tuning curve is detectable only in the uppermost quintile collecting the most active place cells.