Olfactory memory representations are stored in the anterior olfactory nucleus

Abstract

The anterior olfactory nucleus (AON) is the initial recipient of odour information from the olfactory bulb, and the target of dense innervation conveying spatiotemporal cues from the hippocampus. We hypothesized that the AON detects the coincidence of these inputs, generating patterns of activity reflective of episodic odour engrams. Using activity-dependent tagging combined with neural manipulation techniques, we reveal that contextually-relevant odour engrams are stored within the AON and that their activity is necessary and sufficient for the behavioural expression of odour memory. Our findings offer a new model for studying the mechanisms underlying memory representations.

Fig. 1. Silencing a context-specific odour memory.

a Method for activity-dependent labelling and cellular inhibition using hM4D-mCherry. b Injections of 4-OHT induces AAV-mediated gene expression restricted to the medial AON as seen in representative coronal sections of a Fos^CreER mouse [each group, n = 4 mice, two-tailed independent-samples t-test, t(6) = 3.549, *P = 0.0121). Anterior/Posterior (AP); Dorsal/Ventral (DV). Coordinates are relative to bregma. Scale bar represents 500 µm. c Behavioural protocol for tagging and inhibiting context-driven odour engrams. Context (Ctx). d Behavioural expression of contextually-cued odour memory as measured by total investigation time of the odour source [each group, n = 10 mice, two-way ANOVA, main effect of group F(2, 54) = 11.44, ****P < 0.0001; main effect of day F(1, 54) = 13.38, ***P = 0.0006; interaction between group and day F(2, 54) = 8.421, ***P = 0.0007]. e Context-driven reactivation of tagged AON neurons in representative confocal images. Scale bar represents 50 µm. Similar results were observed in all mice independently analysed. f Density measurements of labelled mCherry- (left) and c-Fos-positive (middle) neurons in the medial AON, and the corresponding reactivation rate (right) [each group, n = 6 mice; mCherry densities: one-way ANOVA, F(2, 15) = 0.7637, ns, P = 0.4832; c-Fos densities: one-way ANOVA, F(2, 15) = 0.492, ns, P = 0.6209; engram reactivation: one-way ANOVA, F(2, 15) = 81.43, ****P < 0.0001]. g Correlation between the tagged AON neuron reactivation rate and investigation time [Pearson correlation coefficient, **P = 0.0025, r = 0.6675]. Data are means ± SEM. Source data are provided as a Source Data file.

Fig. 2. Artificially expressing unique odour engrams.

a Method for activity-dependent labelling of AON neurons with photoreceptive ChETA-eYFP for subsequent cellular activation. b Representative coronal section of the AON expressing ChETA-eYFP (green) and c-Fos (red); DAPI (blue). Scale bars represents 500 µm (black) or 50 µm (white). Medial (M); Lateral (L); Dorsal/Ventral (DV). Similar results were observed in all mice independently analysed. c Protocol for tagging and examining the behavioural expression of two unique odour engrams. Chocolate (Ch); Limonene (Li). d Injections of 4-OHT induces AAV-mediated gene expression of ChETA-eYFP [each group, n = 3 mice, independent-samples t-test, t(4) = 4.184, *P = 0.0139]. e Left: percentage of ChETA-eYFP or eYFP-only labelled neurons which express c-Fos following light illumination [each group, n = 3 mice, one-way ANOVA, F(2, 6) = 79.67, ****P < 0.0001]. Right: representative confocal image of ChETA-eYFP neurons after stimulation with 473 nm light. Scale bar represents 50 µm. f Measurements of time spent digging 7 days (left) or 60 days (right) post-tagging [each group, n = 10 mice; 7 days (left): two-way ANOVA, main effect of group F(2, 27) = 0.6025, ns, P = 0.5547; main effect of Time F(19, 513) = 4.901, ****P < 0.0001; main effect of subjects (matching) F(27, 513) = 2.870, ****P < 0.0001; interaction F(38, 513) = 2.352, ****P < 0.0001; 60 days: two-way ANOVA, main effect of group F(2, 27) = 5.3, *P = 0.0114; main effect of Time F(9, 243) = 5.432, ****P < 0.0001; main effect of subjects (matching) F(27, 243) = 2.929, ****P < 0.0001; interaction F(18, 243) = 2.342, **P = 0.002]. g Left: real-time place preference scores determined by the time spent in a 473-nm light-paired or unpaired chamber [each group, n = 10 mice, two-way ANOVA, main effect of group F(2, 54) = 3.727, *P = 0.0305; main effect of light F(1, 54) = 1.492, ns, P = 0.2272; interaction F(2, 54) = 1.523, ns, P = 0.2272]. Right: representative track plots of an individual mouse’s position from eYFP- (top), limonene- (middle), or chocolate-tagged (bottom) groups. Data are means ± SEM. Source data are provided as a Source Data file.

Fig. 3. Long-term reactivation of odour engrams.

a Quantification of eYFP-labelling across groups [each group, n = 10 mice, one-way ANOVA, F(2, 27)=0.04997, ns, P = 0.9513]. b Medial AON activation as measured using c-Fos density as a proxy [each group, n = 10 mice, two-tailed independent-samples t-test, t(18) = 0.1807, ns, P = 0.8586]. c Reactivation rate of limonene- or chocolate-tagged engrams following exposure to limonene or chocolate odours in their associated contexts, 69 days after initial tagging [each group, n = 5 mice, one-way ANOVA, F(3, 16) = 6.968, **P = 0.0033, significance for multiple comparisons *P < 0.05]. d Representative confocal sections of each experimental condition. Scale bar represents 50 µm. Data are means ± SEM. Source data are provided as a Source Data file.

Fig. 4. Odour engrams are stored in the AON.

A model diagram depicting the formation, storage and retrieval of unique, contextually-relevant odour engrams within the AON.