Profs Probe the Science of Sleep
Part two: What dreams are made of

Part two of a four-part series.
We spend a third of our lives sleeping, and 20 percent of that time we’re dreaming. Subimal Datta, a professor of psychiatry at the School of Medicine, wants to know why — and has spent his career as a sleep researcher finding out how it all works.
Our brain is almost as active during sleep as it is during wakefulness, Datta notes, and there’s at least one good reason: that’s when we’re consolidating memories. “Whatever we learned during the wake state, which is in the short-term memory, is being processed,” he says. It’s like a library card catalog system: your brain records what the memory is and where it’s shelved, so the next time it needs to retrieve a memory, it knows where to go — emotional memories in the amygdala, spatial memories in the hippocampus, for instance.
Some sleep basics: after we drift off, we enter what’s called non-REM sleep, which experts divide into four stages with different brain wave patterns. Then, a little less than 90 minutes later, we’re pushed by an internal mechanism into REM sleep, and the pattern cycles back and forth throughout the night, with REM sleep periods gradually lengthening as the night progresses.
REM sleep is clearly important, since that’s when memory is processed, despite the oddity of accompanying dreams. But, like all mammals, our bodies are paralyzed during REM sleep, and for most animals, which are usually some other animal’s prey, that’s dangerous. So much so that certain animals, like mice, have frequent and brief REM states; they may need to run on short notice. Still, at various points, their bodies are paralyzed. And like other mammals, we humans are not just paralyzed; our autonomic functions are out of our control. “The temperature outside could go up or down, yet nothing is controlling body temperature. Respiration can go up or way down,” Datta says. Blood pressure can spike up or down, and that’s one reason why, as morning comes around and we are in longer REM sleep states, more heart attacks occur then than at any other time of day.
So what could be worth putting ourselves at this risk? It’s the memory consolidation that occurs in REM sleep, Datta says. “Our survival depends on our learning. When we stop learning, we’re dead, if you think of it philosophically.” To process memories, the brain needs a lot of energy, so it shuts down most other systems. “It’s like us,” he says. “We can do multitasking, but when we have something very important, we need to focus, and this is what the brain does.”
Datta and his colleagues in the MED Sleep Research Laboratory, which he directs, have identified the cells that are critical for REM sleep memory processing, and they are now working at the molecular level to understand these mechanisms. This basic research, he notes, has potentially big payoffs for pharmacological interventions to treat some sleep disorders, like narcolepsy, which causes victims to suddenly fall asleep during the day. If he finds the exact receptors that turn on and off different sleep signals, drugs could be designed to target those receptors alone.
And the dreams themselves? From the ancients to Freud and on to this day, dreams have been seen as having meaning, either as omens or as signals of a hidden past. Datta will have none of that. Dreaming is simply a noisy by-product of the memory consolidation that’s happening during REM sleep, he says. While newer memories are being laid down in the long-term memory areas of the brain, they run up against random other memories lodged nearby, activating them. The brain just tries to make some sense of it all by creating a narrative, even if the narrative makes little or no logical sense. It’s random noise in the signal, he says, nothing more.
Datta’s conclusion, that REM sleep is essential because its memory consolidation facilitates learning, isn’t the only possible interpretation of the dream state or of sleep in general. Patrick McNamara, a MED assistant professor of neurology, who’s been looking at this question from the point of view of evolutionary biology, is developing several theories of his own.
One is that REM sleep “undoes something that occurs in non-REM sleep,” he says. “REM and non-REM are controlled by different sets of genes, which are in some ways in conflict with each other. In non-REM you might have rising levels of certain hormones that in REM get opposed by rising levels of another hormone. So you have a sort of arms race between these two sleep states.” For example, in developing organisms, non-REM’s slow-wave sleep is associated with very high productions of growth hormone. But REM sleep releases somatostatins, which inhibit the release of growth hormones. “So REM sleep modulates the rate of growth,” McNamara says.
Another theory suggests that morning moods, which have been shown to be a reflection of emotions produced during REM sleep, might be a signal used for evolutionary purposes.
Still, McNamara says, there’s no clear answer about the purpose of REM sleep, in part because historically it hasn’t been studied much. “I think REM sleep was just such a difficult problem — it’s not like other biologic functions, that you can say this chemical does that and you know why. With REM sleep, it’s got all these paradoxical properties, so it’s harder to figure out what it could possibly be doing.”
The purpose of non-REM sleep is better understood. It apparently has a restorative function, helping repair tissue and realign biological rhythms, says McNamara, although exactly how this happens still isn’t known. Non-REM sleep also produces a number of antibodies, adds Datta. “If you have a cold or infection, to fight against those germs, you need antibodies. During that part of sleep, mostly in slow-wave sleep, those antibodies are synthesized,” he says. “That’s why when you have a cold or infection, you get more sleepy.”
Check BU Today tomorrow to read part three of “Profs Probe the Science of Sleep.” Click here to read "Part one: Cracking the circadian code."
This article originally ran in the Winter 2006–2007 edition of Bostonia.
Taylor McNeil can be reached at tmcneil@bu.edu.