Do dreams participate in memory processing? There is good evidence that both REM and NREM forms of sleep participate in various forms of memory processing but what about the mentation that accompanies these sleep states? Are dream images reflections of memory processing? Do dream images play a causal role in ensuring consolidation of some memories over others? Before presenting new evidence from a recent paper that dreams do indeed participate in memory processing it is important to note that even if that is the case it cannot be that all dream content participates in memory. After all most dreams are not mere collections of random memory images. Instead dream images can be novel without reference to the past and are organized into narratives that exhibit a logic unrelated to memory processing. In addition there are themes that are constant across individuals and cultures. Unless we believe that everyone is consolidating the same memories in the same way then it must be the case that dreams contain more than memory images.
In any case memory images are thought to be processed during sleep into two basic states corresponding to the two basic forms of sleep—non-REM slow wave sleep (SWS) and REM sleep. First recent memories from the past day (what Freud called the day residue) are processed during SWS, with some sort of selection process occurring during SWS. Essential or adaptive memory images are retained for further processing and non-essential images are discarded.
The second step takes place during REM sleep, in which adaptive memories are translated into a format suitable for long term storage and integration into existing memories. People who study nightmares have also suggested that part of this translation process in REM also involves a kind of decoupling of intense emotions or at least a decoupling of somatic arousal levels from the images that carry those emotions. Once the decoupling occurs the de-fanged images can then be stored in long term memory. When this decoupling process breaks down you get nightmares during sleep and you get a constant irritant in the memory system that constitutes a core for traumatic memory. The more we know therefore about sleep and dream-related memory processing the more we can help people with nightmares or traumatic memories.
In a recent paper published in the Neurobiology of Learning and Memory Van Rijn and colleagues asked 44 participants keep a daily log for 10 days, reporting major daily activities (MDAs), personally significant events (PSEs), and major concerns (MCs). After keeping the daily logs all participants underwent an ‘awakening from sleep protocol’ in order to collect dreams from specific sleep states (REM and NREM). Twenty participants were assigned to a home awakening condition and the rest of the participants were assigned to the sleep Lab awakening condition. Participants in the sleep Lab were awakened during the night when they went into SWS and REM sleep and asked for dream reports. The participants in the home underwent awakenings via the nightcap sleep monitoring device in the REM state. But due to equipment or other problems data was obtained from only 15 participants in the home condition and 12 subjects awakened during SWS. After the awakenings protocol all participants then kept a daily log again for the next 10 days which thus acted as a controlled stimulus with two salience levels, high (sleep lab) and low (home awakenings).
The researchers wanted to see if any dream images obtained at home or in the sleep Lab were related to or came from any of the MDAs, PSEs or MCs. The researchers asked for 10 days of daily experiences so that they could test whether they could detect a dream lag effect in the dreams of participants. This dream lag effect refers to the fact that some dream images appear to come from events experienced by the dreamer some 5-7 days previous to the dream. van Rijn et al report that the dream-lag effect was found for the incorporation of PSEs into REM dreams collected at home, but not for MDAs or MCs. No dream-lag effect was found for SWS dreams, or for REM dreams collected in the lab after SWS awakenings earlier in the night. The lack of dream lag effect in the Lab was explained by the previous awakening from SWS. If SWS is needed to expedite the memory consolidation process then interrupting SWS via the awakening prevented it from doing its job so no tagged content was sent to the REM stage from SWS so no lag effect could be detected or so the authors appear to argue. But if this were the case would not the same logic apply to REM awakenings.
Would not REM awakenings prevent REM from playing its role in memory processing and thus prevent detection of a dream lag effect? But dream lag effect has been detected in other awakening REM studies. With regard to the dreams recorded after the awakening protocol the dream-lag effect was found for references to the experience of being in the sleep laboratory, but only for participants who had reported concerns beforehand about being in the sleep laboratory. In short, evidence for incorporation of daily experiences into dreams over the course of several days was found only for personally significant events including an event like being awakened
in the sleep laboratory for those who had been concerned about that impending experience. It is always easy to criticize complex experiments like this one. You have multiple conditions with multiple analyses involving an ever decreasing statistical power to detect effects given equipment failures and given individual drop out and missing data and so on. Nevertheless the clear signal that emerges from the data is that the dream images reflecting dream lag effect are selective for personally significant experiences rather than for other equally complex experiences like MDAs (major daily activities) or MCs (major life concerns).
It is not surprising that memory elements are chosen for processing and consolidation dependent on their personal and perhaps their emotional significance. Why bother to remember insignificant events? What is more interesting is that the results of this selection process are reflected in dream content as evidence from van Rijn and the dream lag effect more generally attests. Interestingly although dreams are filled with emotions the emotions are very often not attached to episodic memories—that is emotional images in dreams are very often de-contextualized. Decoupling emotions from their original context may help to defuse their intensity, thus making them more amendable for integration into long term semantic memories and preventing formation of traumatic memories. This decoupling effect probably occurs in late in the sleep period in REM sleep. It may be that emotions are decoupled from context within a single sleep cycle and then emotionally significant elements (characters, themes, actions, emotions etc) are processed over the next few days and given temporal priority in the processing cycle over emotionally neutral objects and items.