Retrograde labeling of MS-projecting SuM neurons.

a, Diagram of retroAAV-EGFP injection into MS and retroAAV-tdTomato injection into hippocampus. b,c, Post-hoc histological images showing the expression of EGFP at the injection site of MS (b), and the expression of tdTomato at the injection site of hippocampus (c). d, Representative serial sections showing the eGFP- and tdTomato-labeled cell bodies in SuM. e, Summary of co-labelling proportion of SuM-Hip and SuM-MS projecting neurons, n = 6 mice. SuM, supramammillary nucleus; MS, medial septum; Hip, hippocampus.

Ca2+ recording of lateral SuM-MS projecting neurons across sleep-wakefulness cycle.

a, Experimental design for virus injection in MS and fiber implantation in lateral SuM. b, Post-hoc histological image showing the expression of GCaMP6m at injection site in MS (insert) and in SuM cell bodies, and optical fiber location in lateral SuM. c, Ca2+ activities in lateral SuM-MS projecting neurons across sleep-wakefulness cycles. Color map indicates power spectrum (μV2) of EEG, Freq., frequency. d, Summary of the area under the curve per second during wake, NREM sleep, and REM sleep. n = 6 mice, RMs 1-way ANOVA with LSD post-hoc comparison. e–h, Ca2+ activities during brain state transitions: wake-to-NREM (e), NREM-to-wake (f), NREM-to-REM (g), and REM-to-wake (h). n = 6 mice, paired t test. n.s. P > 0.05, *P < 0.05, **P < 0.01. Data are reported as mean ± SEM.

activity patterns of single lateral SuM-MS projecting neurons across sleep-wakefulness cycle.

a, Experimental design for virus injection and optrode implantation in the lateral SuM-MS projection. b, Representative image showing virus expression and locations of fiber and tetrode tips in lateral SuM. c, Histogram of stimulus time for spikes in a representative lateral SuM-MS projecting neuron (Cell#B153-TT4-C2). d, Waveforms of spontaneous (grey) and light-induced (red) spikes from the unit in (c). e, Success rate versus temporal jitter of the first light-induced spikes for all recorded lateral SuM-MS projecting neurons, n = 29 neurons from 15 mice. f,g, Distributions of latencies of the first light-induced spikes (f), and correlation coefficients between light-induced spikes and spontaneous spikes (g) for all recorded lateral SuM-MS projecting neurons, n = 29 neurons from 15 mice. h, Firing rate of the neuron in (c) across sleep-wakefulness cycles. Freq., frequency. QW, quiet wakefulness. i, Firing rates of 29 tagged lateral SuM-MS projecting neurons during three different states. Friedman test with Bonferroni post-hoc comparison. j, Firing rate modulation of lateral SuM-MS projecting neurons (red, n = 29 neurons from 15 mice) and non-MS-projecting lateral SuM neurons (black, n = 130 neurons from 15 mice). k, Distribution of REM-active, NREM-active and QW-active neurons of MS-projecting and non-MS-projecting lateral SuM neurons. n.s. P > 0.05, ***P < 0.001. Data are reported as mean ± SEM.

Optogenetic silencing of lateral SuM-MS projecting neurons during REM sleep impairs both social and contextual fear memory.

a–d, Effects of silencing lateral SuM-MS projecting neurons on CA1 local field potential (LFP) activity during REM sleep. a, Experimental design for retrograde labelling of lateral SuM-MS projecting neurons by GtACR1-mCherry, fiber implantation above SuM, and LFP recording in CA1. b, Representative EEG showing the effects of silencing lateral SuM-MS projecting neurons on CA1 LFP during REM sleep. Freq., frequency. c,d, CA1 LFP power spectral analysis pre-, during, and post-photoinhibition. RMs 1-way ANOVA test, n = 6. e, Protocol for three-chamber social memory task. E, empty; M, mouse; N, novel; F, familiar. f, Representative heatmaps of distribution of time in three-chamber task. g,h, Summary of social preference indexes in training (g) and testing (h) of the social novelty task. 1-way ANOVA with LSD post hoc comparison, GtACR1 group n = 8, GtACR1 control n = 7, mCherry control n = 8. i, Protocol for optogenetic silencing and contextual fear memory. j,k, Freezing levels during training (j) and testing (k) phases. RMs 2-way ANOVA with Sidak post-hoc comparison, GtACR1 group n = 7, GtACR1 control n = 8, mCherry control n = 8. n.s. P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001. Data are reported as mean ± SEM.

Optogenetic silencing of MS-CA2 projecting neurons during REM sleep impairs social but not contextual fear memory.

a, Experimental design for retrograde labelling of MS-CA2 projecting neurons by GtACR1-mCherry, and fiber implantation above MS. b, Representative heatmaps of distribution of time in three-chamber task. c,d, Summary of social preference indexes in training (c) and testing (d) of the social novelty task. 1-way ANOVA with Bonferroni post hoc comparison; GtACR1 group n = 10, GtACR1 control n = 8, mCherry control n = 11. e,f, Freezing levels during training (e) and testing (f) phases. RMs 2-way ANOVA test; GtACR1 group n = 8, GtACR1 control group n = 10, mCherry control n = 11. n.s. P > 0.05, ***P < 0.001. Data are reported as mean ± SEM.