Why does serotonergic activity drastically decrease during REM sleep?

Here, I postulate two hypotheses that can explain the missing link between sleep and the serotonergic system in terms of spine homeostasis and memory consolidation. As dendritic spines contain many kinds of serotonin receptors, and the activation of serotonin receptors generally increases the number of spines in the cortex and hippocampus, I postulate that serotonin neurons are down-regulated during sleep to decrease spine number, which consequently maintains the total spine number at a constant level. Furthermore, since synaptic consolidation during REM sleep needs long-term potentiation (LTP), and serotonin is reported to inhibit LTP in the cortex, I postulate that serotonergic activity must drastically decrease during REM sleep to induce LTP and do memory consolidation. Until now, why serotonergic neurons show these dramatic changes in the sleep-wake cycle remains unexplained; however, making these hypotheses, I can confer physiological meanings on these dramatic changes of serotonergic neurons in terms of spine homeostasis and memory consolidation.