Fig. 2. Effect of deletions of serotonergic neurons on spontaneous sleep–wake (a–f) and CO₂ arousal (g–i).

Deletion of serotonin neurons in the DR (Sert-Cre DR-DTA, n = 9) or MR (Sert-Cre MR-DTA, n = 4) compared with the wild-type (WT, n = 6) had no effect on the percentage of time spent (mean ± SEM) in wake (a), non-REM sleep (b) or REM sleep (c) during both the 12 h of light and dark phases. The hour-wise comparison of the wake (d) NREM (e) and REM sleep (f) for the three groups are virtually superimposable. By contrast, during 30 sec exposure to 10% CO₂, the animals with DR deletions (Sert-Cre DR-DTA, n = 9) had a 2.2-fold longer mean latency to arousal and the animals with MR deletions (Sert-Cre MR-DTA, n = 4) a 1.6-fold longer latency (g). Panel h shows the survival curve for the percentage of mice (mean ± SEM) still asleep at various time points after onset of the CO₂ stimulus. Panel i illustrates the percentage of animals (two-way ANOVA for treatments and time, with multiple comparisons to the WT) still asleep at 15 (WT vs. DR-DTA, < P = 0.001, vs. MR-DTA, P = 0.002), 30 (WT vs. DR-DTA, P < 0.001) and 40 sec (WT vs. DR-DTA, P = 0.002) after CO₂ onset. Almost all control animals are awake by 30 sec, but almost 46% of animals with DR^Sert deletions and almost 20% with MR^Sert deletions failed to arouse by this time point. The error bars in all the graphs (a–i) represent the SEM. **P < 0.01; ***P < 0.001; based on either one-way (g) or two-way ANOVA (i, h).