Infraslow calcium oscillation of granule cells during NREM sleep.

A. Representative recording session showing infraslow oscillation (ISO) during NREM sleep. From top to bottom: brain states, EEG power spectrogram (0-25 Hz), EMG amplitude, photometric signal. B. Left, Schematic representation of the recording setup. Right, a fluorescence image showing the expression of GCaMP6s (green) in the granule cell layer and the optic fiber placement (dashed line) in a Dock10-Cre mouse. Scale bar, 500 µm. C. Quantification of calcium activity in granule cells (13 recording sessions in 5 Dock10-Cre mice, n.s. – no significance, ** - P<0.01, paired t-test) in wake (W), NREM sleep (N), and REM sleep (R). D. Left: oscillation peak frequency during NREM sleep and wake based on Fourier transformation of the photometry signal. Right: Quantification of calcium oscillations in GCs (13 sessions in 5 mice). E. A representative example showing the coincidence of calcium troughs with MAs. F. Percentage of state transition outcome from each calcium dip. G. Peri-stimulus time histogram (PSTH) in one recording session showing calcium signal aligned with the onset of MAs. Bottom left: parameters of the calcium signal used for quantification. H. Quantification of the latency (t) and magnitude of the calcium trough (Drop) during MAs (13 sessions from 5 Dock10-Cre mice).

Infraslow calcium oscillation of mossy cells during NREM sleep.

A. Representative recording session showing infraslow calcium oscillation during NREM sleep in a Drd2-Cre+/- mouse injected with AAV-FLEX-GCaMP6s. From top to bottom: brain states, EEG power spectrogram (0-25 Hz), EMG amplitude, photometric signal. B. Fluorescence images showing the expression of GCaMP6s (green) in the mossy cells in the area marked with the white rectangle. Blue, DAPI. Scale bars, 500 µm and 50 µm. C. Quantification of calcium activity in mossy cells during different brain states (12 recording sessions in 4 Drd2-Cre+/- mice; n.s. – no significance, ** - P<0.01, paired t-test). D. Left: oscillation peak frequency during NREM sleep based on Fourier transformation of the photometry signal. Right: quantification of calcium oscillations during NREM sleep. E. A representative example showing the coincidence of calcium troughs with MAs (represented with vertical green lines). F. Percentage of state transition outcome from each calcium dip. G. PSTH from one recording session showing calcium signals (red) and EMG signals (light green) aligned with the onset of MAs. H. Quantification of the latency (t) and the magnitude of calcium troughs (Drop) during the MAs (12 sessions from 4 Drd2-Cre mice).

Two-photon calcium imaging of DG activity in sleeping mice.

A. Left, schematic layout of two-photon imaging and EEG recording setup. Right, representative images showing putative granule cells (pGCs) and putative mossy cells (pMCs). Scale bars, 50 µm. B-C Representative two-photon recording sessions of field of views containing putative pGCs pMCs. From top to bottom: brain states (gray – awake, orange – NREM), EEG spectrogram (0-25 Hz), EMG, velocity (vel.), and calcium traces in individual cells. Red traces - upregulated cells, blue traces –downregulated cells, gray traces – non-significant cells. D-E. Left, Percentage of Up-, downregulated and non-significant cells from the entire recorded cell populations. Middle, quantification of calcium activity in different brain states in up-, down-regulated cells (pGCs: P<0.05 in up-regulated cells between NREM and MA, P=0.77 in down-regulated cells between NREM and MA; pMCs: P<0.001 in up-regulated cells between NREM and MA, P=0.63 in down-regulated cells between NREM and MA, paired t-test). Right, Averaged activity in up-, down-regulated pGCs and pMCs during microarousals (MA). Time 0 indicates the MA onset. Putative GCs: 369 cells from 3 C57BL/6J mice; putative MCs: 269 cells from 2 C57BL/6J mice.

Phasic release of 5-HT in the DG during NREM sleep.

A. Left, schematic of experimental design. Right, expression of 5-HT sensor in the hippocampus. B. A representative example of 5-HT signals during different brain states. From top to bottom: brain states, EEG power spectrogram (0-25 Hz), EMG signal, photometric signal. The dashed-box enlarged below in panel B. Note the coincidence of 5-HT release with MAs during NREM sleep (black arrows and vertical green lines). C. Percentage of state transition outcome from each 5-HT event (averaged data from 6 mice). D. Quantification of 5-HT signals in the DG during different brain states (14 sessions from 6 C57BL/6J mice, ** - P<0.01, *** - P<0.001, paired t-test). Data were normalized to Z scores in each recording session. E. Quantification of oscillatory cycles of 5-HT signals in the DG (14 recording sessions from 6 C57BL/6J mice).

Correlation between DG oscillation and activity of 5-HT neurons during NREM sleep.

A. Left, Schematic representation of the 2-site photometry experimental design. Right, Expression of CaMKII-GCaMP6s and fiber placement in the DG and raphe nuclei. B. A representative example of concurrent recording of DG and raphe 5-HT neurons in a Sert-Cre+/- mouse during sleep. From top to bottom: brain states, EEG power spectrogram (0-25 Hz), EMG amplitude, photometric calcium signals (CaMKII-G6s) in DG and in dorsal raphe. C. Correlation analysis of calcium activity between DG and raphe 5-HT neurons during NREM sleep and wakefulness in one recording session. D. Quantification of correlation coefficient between DG activity and raphe activity during different brain states (11 sessions from 3 Sert-Cre+/- mice, *** - P<0.001, paired t-test).

Genetic knockdown of 5-HT1a receptors in DG impairs ISO and memory performance.

A. Schematic representation of the experimental design. A mix of AAV9-CamKII-GCaMP6s and AAV1-hSyn-Cre was injected into the DG of 5-HT1aflox/flox mice. B. Representative example showing photometry and EEG recordings in the DG of a control mouse injected with AAV9-CaMKII-GCaMP6s alone. Right, Fourier transformation of calcium activity during wake (blue) NREM sleep (red). C. A representative example showing photometry and EEG recordings in the DG of a mouse injected with AAV9-CaMKII-GCaMP6s and AAV1-hSyn-Cre. Right, Fourier transformation of calcium activity during wake (blue) NREM sleep (red). D. Left, Quantification of the relative power of the calcium oscillation in the range of 1-2 cycles/min in the Cre and control groups (16 sessions from 5 mice for Cre, 16 sessions from 6 mice for control). Right, Quantification of calcium oscillation amplitudes in the Cre and control groups. Calcium signals in each mouse were normalized to Z scores. ** - P<0.01, *** - P<0.001, unpaired t-test. E, Schematic representation of the CFC experimental design. F, Left, Contextual fear recall tests showing percentage of freezing in one minute time bins for 5-HT1aflox+/+ mice bilaterally injected with AAV9-CaMKII-Cre-GFP (Cre) or AAV9-CaMKII-GFP (GFP). Right, Quantification of freezing behavior over 5-minute interval during contextual recall tests in Cre and GFP groups (N=11 for Cre, N=12 for GFP, *, P<0.05, ***, P<0.001, un-paired t-test).

Characterization of infraslow oscillation of granule cells (GC) during NREM sleep.

A, A representative recording session showing GC activity during wake/sleep cycles. Blue lines (middle) indicate the microarousals. B, Quantification of GC infraslow oscillation power (left) and amplitude (right) at different stages of NREM sleep (16 sessions from 5 Dock10-Cre mice, *** P<0.001, n.s. no significance, paired t-test). T0 refers to the first minute of the first NREM epochs, defined as defined as those with the prior wakefulness longer than 5 minutes. C, A representative example showing EEG spectrogram (0-15Hz), EEG sigma power, and GC calcium signals. D, Correlation analysis between GC calcium activity and EEG sigma power during NREM sleep in one recording session. E, Quantification of correlation coefficient between GC activity and EEG sigma power during different brain states (13 sessions from 5 Dock10-Cre mice, ** P<0.01, *** P<0.001, paired t-test).

Two-photon calcium imaging in sleeping mice.

A, Representative EEG recording session showing sleep states in a head-fixed mouse under 2-photon microscope. B, Quantification of brain states in putative GCs (pGCs, green circles) and putative MCs (pMCs, red circles) during imaging sessions used in Fig 3. C, A representative example of per-stimulus time histogram (PSTH) showing neuronal activity of up-regulated (left) and down-regulated (middle) GC cells during all 153 microarousal events in a recording session. The averaged activity across all up-regulated or down-regulated cells were shown in each row in PSTH. Right, aligned EMG amplitude during microarousal events. D, Quantification of neuronal calcium activity in different brain states in up-regulated (N=183 cells), down-regulated (N=153 cells), and unchanged (N=33 cells) putative GCs. Gray: individual cells, red: Mean ± SEM. Data was collected from 3 imaging sessions in 3 C57BL/6J mice. Up-regulated, not down-regulated GCs displayed decreased activity during MAs, compared to that in NREM sleep (*, P<0.05, n.s., no significance, paired t-test). E, Quantification of neuronal activity in different brain states in up-regulated (N=38 cells), down-regulated (N=87 cells), and unchanged (N=10 cells) putative MC. Data was collected from 2 imaging sessions in 2 C57BL/6J mice. Up-regulated, not down-regulated putative MCs displayed decreased activity during MAs, compared to that in NREM sleep (***, P<0.001, n.s., no significance, paired t-test).

Phasic release of 5-HT in the raphe nuclei during NREM sleep.

A. Representative recording session showing 5-HT signals during different brain states. From top to bottom: brain states, EEG power spectrogram (0-25 Hz), EMG, photometric signal. The dashed-box enlarged below in panel B. Note that the 5-HT release during NREM sleep coincided with the microarousal (MA) episodes (black arrows). C, Percentage of outcome of brain states following 5-HT release during NREM sleep (average data from 11 sessions in 5 C57BL/6J mice). D, Left, A representative example showing 5-HT release during MAs. Right, aligned EMG burst during MAs. PSTHs were aligned to the onset of MAs. E, Quantification of latency to the MAs and number of 5-HT release (Events, defined as those signals with a peak above the baseline) per minute during NREM sleep (11 sections from 5 C57BL/6J mice). F, Quantification of 5-HT signals in the raphe nuclei during different brain states (11 sections from 5 C57BL/6J mice, **, P<0.01, paired t-test). Data were normalized to Z scores in each recording session.

CaMKII-labeled cells in the DG display oscillatory activity during NREM sleep.

A, Representative recording session showing calcium activity in a wildtype mouse injected with AAV9-CaMKII-GCaMP6s in the DG. From top to bottom: brain states, EEG power spectrogram (0-25 Hz), EMG, photometric signal. B, A fluorescent image showing the constrained expression of GCaMP6s (green) in the granule cell layer of the DG. Scale bar, 500 µm. C, Quantification of calcium activity (16 recording sessions in 6 C57BL/6J mice, n.s., no significance, **, P<0.01, paired t-test) in wake (W), NREM sleep (N), and REM sleep (R). D. Left: oscillation peak frequency during NREM sleep and wake based on Fourier transformation of the photometry signal. Right: Quantification of calcium oscillations in CaMKII-labeled cells (16 sessions in 6 mice).

Fluorescent in situ hybridization (FISH) of 5-HT1a receptors in the DG.

A, Left, Schematic of the experimental design. AAV9-CaMKII-GFP-Cre was injected unilaterally into the dorsal DG of 5-HT1aflox/flox mice. B, representative examples of FISH signal of Htr1a (red) and GFP (green) in a brain section of a 5-HT1aflox/flox mouse (3 mice in total). C, D, Enlarged view of Htr1a and GFP expression in the DG corresponding to the yellow boxes in B. Blue, DAPI. Scale bars, 200 μm in B and 20 μm in C and D.

Sleep and photometry analysis in 5-HT1a mice.

A, Quantification of the total duration of wake, NREM sleep, REM sleep, and the number of microarousals (MA) in wildtype (WT, N=5) and 5-HT1a knockout mice (N=5) injected with AAV-hSyn-Cre in the DG. Data were analyzed from intermittent EEG/photometry recordings across light and dark cycles. No significant difference was observed between WT and 5-HT1a mice (unpaired t-test). B, Representative examples of PSTH in a wildtype (WT) and 5-HT1a knockout mouse (injected with AAV-hSyn-Cre in the DG) aligned with the onset of MAs. C, Quantification of the latency and the magnitude of calcium drops during the MAs in WT (16 sessions from 6 mice) and 5-HT1a mice (16 sessions from 5 mice). * P<0.05, ** P<0.01, unpaired t-test.