Extended Data Fig. 4: mPFC Pyr neuron activity promotes rapid EMs during REM sleep.
(a) Schematic illustrating pupil recordings using an infrared (IR) camera in head-fixed animals. (b) Example of REM sleep with phasic θ events in the EEG and rapid EMs. Two phasic θ events are shown in red in the raw EEG (top) and indicated by dashed lines in the EEG spectrogram (middle). Rapid EMs (blue ticks) are identified as sudden accelerations in the pupil speed (bottom). Inset, consecutive movie frames including a rapid EM. Red circles, current pupil position; blue circles, pupil position in previous video frame; scale bars; 200 µV, 1 s; px/s, pixels per second. (c) Inter-EM distribution calculated for all REM sleep periods in recordings without laser stimulation (n = 12 mice). Red line, threshold for defining EM bursts (250 ms). (d) Frequency of all EMs and EM bursts during phasic θ events and randomly selected epochs of equal duration within REM sleep in baseline recordings without laser stimulation. All EMs, paired t-test, P = 0.0000; EM bursts, paired t-test, P = 0.0000. n = 12 mice. Box plots; dots, individual mice. (e) Duration of REM sleep episodes with (ON) or without (OFF) laser activation of ChR2-expressing mPFC Pyr neurons and REM sleep duration in baseline recordings without laser stimulation (Bsl) in the same animals. One-way RM ANOVA with laser as within-subjects factor (P = 0.0000); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.2029; Bsl versus ON: P = 0.0016; OFF versus ON: P = 0.0005. n = 6 mice. Box plots; lines, individual mice. (f) Frequency of phasic θ events during laser-on and laser-off REM sleep episodes and episodes in baseline recordings without laser in ChR2 mice. One-way RM ANOVA (P = 0.0012); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.1183; Bsl versus ON: P = 0.0427; OFF versus ON: P = 0.0173. n = 6 mice. Box plots; lines, individual mice. (g) Frequency of EMs and EM bursts during laser-on, laser-off and baseline REM sleep episodes in ChR2 mice. Note that REM sleep episodes with duration < 20 s were excluded from this analysis, as EMs only occur with a delay of about 20 s after the onset of spontaneous REM sleep24. All EMs, one-way RM ANOVA (P = 0.0140); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.0483; Bsl versus ON: P = 0.0483; OFF versus ON: P = 0.0483. EM bursts, one-way rm ANOVA (P = 0.0030); t-tests; Bsl versus OFF: P = 0.0493; Bsl versus ON: P = 0.0493; OFF versus ON: P = 0.0493. n = 6 mice. Box plots; lines, individual mice. (h) Latency between the onset of REM sleep and the first EM for laser-on and laser-off episodes and episodes in baseline recordings without laser in ChR2 mice. One-way RM ANOVA with laser as within-subjects factor (P = 0.0682). n = 6 mice. Bars, averages across mice; dots, individual mice; error bars, 95% CIs. (i) Duration of laser-off or laser-on REM sleep episodes and baseline duration in recordings without laser in mice expressing iC++ in mPFC Pyr neurons. One-way RM ANOVA (P = 0.0051); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.1219; Bsl versus ON: P = 0.0395; OFF versus ON: P = 0.0198. n = 6 mice. Box plots; lines, individual mice. (j) Frequency of phasic θ events during laser-on, laser-off, and baseline REM sleep episodes in iC++ mice. One-way RM ANOVA (P = 0.0031); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.6853; Bsl versus ON: P = 0.0272; OFF versus ON: P = 0.0495. n = 6 mice. Box plots; lines, individual mice. (k) Frequency of EMs and EM bursts during laser-on, laser-off, and baseline REM sleep episodes in iC++ mice. Note that REM sleep episodes with duration < 20 s were excluded from this analysis. All EMs, one-way RM ANOVA (P = 0.0000); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.0356; Bsl versus ON: P = 0.0042; OFF versus ON: P = 0.0005. EM bursts, one-way rm ANOVA (P = 0.0000); t-tests with Holm-Bonferroni correction; Bsl versus OFF: P = 0.0353; Bsl versus ON: P = 0.0048; OFF versus ON: P = 0.0003. n = 6 mice. Box plots; lines, individual mice. (l) Latency between the onset of REM sleep and the first EM for laser-on, laser-off, and baseline episodes in iC++ mice. One-way RM ANOVA (P = 0.8167). n = 6 mice. Bars, averages across mice; dots, individual mice; error bars, 95% CIs. (m) Effect of mPFC Pyr neuron activation on the frequency of all EMs and EM bursts for different durations of baseline and laser-on episodes. All EMs, two-way ANOVA with laser (Bsl or ON, P = 0.0000) and REM bout duration as between-subjects factors (interaction, P = 0.0036); t-tests with Holm-Bonferroni correction; 0–40 s, P = 0.0023; 40–80 s, P = 0.0012; 80–120 s, P = 0.3496; > 120 s, P = 0.9968. EM bursts, two-way ANOVA (laser, P = 0.0001; interaction, P = 0.0130); t-tests; 0–40 s, P = 0.0212; 40–80 s, P = 0.0197; 80–120 s, P = 0.8791; > 120 s, P = 0.8791. Bsl, n = 434; ON, n = 240 episodes. Bars, averages across mice; dots, individual mice; error bars, 95% CIs. (n) Effect of mPFC Pyr neuron inhibition on the frequency of all EMs and EM bursts for different durations of baseline and laser-on episodes. All EMs, two-way ANOVA (laser, P = 0.0003; interaction, P = 0.0052); t-tests with Holm-Bonferroni correction; 0–40 s, P = 0.1656; 40–80 s, P = 0.0724; 80–120 s, P = 0.0362; > 120 s, P = 0.0088. EM bursts, two-way ANOVA (P = 0.0004; interaction, P = 0.0017); t-tests; 0–40 s, P = 0.2412; 40–80 s, P = 0.1671; 80–120 s, P = 0.0096; > 120 s, P = 0.0096. Bsl, n = 501; ON, n = 216 episodes. Bars, averages across mice; dots, individual mice; error bars, 95% CIs. See Supplementary Table 1 for detailed statistical information. *P < 0.05, **P < 0.01, ***P < 0.001.
