Review of the Antonio Zadra and Robert Stickgold (2021) article Theater of the Mind in the January 2021 issue of Psychology Today

Zadra and Stickgold’s (2021) excellent article Theater of the Mind in the January 2021 issue of Psychology Today, https://www.psychologytoday.com/gb/articles/202101/theater-the-mind,  summarises ideas from their new book When Brains Dream: Exploring the Science and Mystery of Sleep. That book hasn’t arrived yet in the UK and so this piece reviews the Psychology Today article, which, I hope, gets wide readership as it addresses in great detail much of the science of dreaming and the arguments as to whether dreaming has a function. Zadra and Stickgold start their article by noting that dreams can be very wide-ranging and bizarre, but that dreams also have some common themes. Common here means that a large proportion of people will report having had such a theme at least once, such as of falling, or being chased. Given this, they ask the question, what function do dreams serve? Their answer is to propose that dreams explore new associations, often weak associations between memories, so that the dreamer / brain, when awake, understands possibilities for the future. To expand on this they describe empirical findings on the contents of dreams. For example, most dreams have a narrative, and often there is some type of problem that is responded to. There are also some common aspects of content, for example, misfortunes, which are found to be seven times more frequent in dreams than are good fortunes.
When addressing they question ‘what are our dreams for?’, Zadra and Stickgold state that ‘Whatever the function of dreams may be, it cannot depend on remembering them once we awaken.’ They distinguish such a biological adaptive function for all dreams, whether remembered or unremembered, from the uses that we may choose to make of dreams that we do remember. They review then various theories of dream function that encompass remembered and unremembered dreams. For example, Freud’s conclusion that dreams allow the partial expression of repressed wishes, and the views that we may find solutions to personal problems in dreams, or practice overcoming threats, or practice social interactions, or extinguish our fears. Importantly, they also refer to the proposal by Ernest Hartmann that dreams weave emotional memories from the day into existing memories, making broader and looser connections during sleep than the more obvious and strong connections that are prioritized during wakefulness. (See Hartmann, 1996, for more on this.) 
They state that each of these theories may be partly true, and allow also for Hobson’s emphasis on randomness in dreams. But they then add a further, memory function for dreams. In this they review the work of Erin Wamsley, conducted with Stickgold, on task learning before sleep and the relationship of subsequent performance to sleep and to dream content. Wamsley and colleagues found that performance was improved across sleep if dreaming of the task occurred, but that the dreams were usually only indirectly related to the task that was learned. For example, in a maze learning task, dreaming of caves was classed as a relevant albeit indirect association. The proposal is thus that in dreams there is a search for weak, novel associations, and so the article gives the example of the experience of being creative during the day resulting in a dream of discovering a new room in one’s childhood home. Zadra and Stickgold conclude that dreams are thus exploring possibilities, which results in gaining new understanding about ourselves and the world.
Zadra and Stickgold are undoubtedly correct to propose that in dreams we explore rarer or newer or novel associations and that this can result in novel possibilities being understood, even if only unconsciously, when we are awake. However, this review gives a caution in deeming that a memory function has been shown in current experiments on dreaming and learning. Wamsley’s work, and her recent (2019) replication with Stickgold of that work, are amenable to an alternative interpretation in that dreaming of the learning task was not only associated with performance improvement across sleep, but also with poor performance prior to sleep! It may thus be that dreaming of the task was related to personal concerns and even embarrassment about poor performance, rather than related to memory consolidation during sleep. But the idea of exploring weak or novel associations during dreams does seem very plausible. Zadra and Hartmann credit Hartmann on this, who, in 1996, wrote that in dreaming we make connections more broadly than when we are awake. Hartmann held that these broad connections are not made at random, but are guided by our emotions. Thus, dreams contextualise our dominant emotions or emotional concerns, thus producing metaphors for the emotional state of the dreamer. According to Hartman this spreading out of excitation in the dream, finding novel connections for our new memories, ‘is probably functional.’ He contrasts our goal-directed waking thinking with the wider connections made in dreams. For Hartmann, this explains his empirical findings that we rarely dream of reading or writing or typing in dreams.
This article by Zadra and Stickgold, and the previous work by Hartmann, rely on the claim that by the exploration of weak associations, dreams leads to improvements in the storing of our memories, as part of the elaborations of memory that occur when memories are processed and evolve during sleep. We must remember though that Hartman cautions that this process is ‘probably functional.’ Hartmann is correct here for two reasons:

1. It is not yet been shown that functional brain processes are related to dream content. For example, there are complex memory processing functions occurring during slow-wave sleep but often without accompanying dream recall. Dreaming might thus not be a component or consequence of complex memory processing during sleep.
2. Even if the contents of dreams are related to the original, often emotional material that is processed during sleep, such as a concern or a traumatic event, the dream may act on that material with separate processes than the functional brain processes that are also occurring. The idea here is that the brain could be doing two things at once!

If we put aside the current debate about whether REM sleep has a memory consolidation function at all, the caveats raised by Domhoff (2018, pp.268-275) persist, namely that the incorporation of waking events into dreams is very rare, and that dreams depict more long-standing concerns and conceptualisations of the dreamer. He gives the example of a person driving a car and daydreaming at the same time, quoting Antrobus that processing during sleep might not be represented in the form of imagery.
I would not go as far as Domhoff has done here in distancing dream imagery from functional brain processes, because there may be complex social cognition brain processes that are evidenced in dream imagery, and Zadra and Stickgold may be correct that sleep and dreaming enable the exploration of novel connections and possibilities in such social cognitive processing. The point here, though, is that Hartmann may well have been correct to say that the dream processes are only ‘probably functional.’
Aside from that caution about the Zadra and Stickgold article, the other concluding comment here is to take issue with their claim that a function for dreaming must hold for unremembered dreams. Our work on dream sharing and empathy (Blagrove et al., 2019) does lead to the speculative possibility that the sharing of dreams across human history has led to the evolutionary selection of fictional and salient dream content, on a timescale similar to the development of story-telling in humans. It may thus be that it is only remembered dreams that are functional and adaptive!
 
Blagrove, M., Hale, S., Lockheart, J., Carr, M., Jones, A., & Valli, K. (2019). Testing the Empathy Theory of Dreaming: The Relationships Between Dream Sharing and Trait and State Empathy. Frontiers in Psychology, 10, 1351. https://doi.org/10.3389/fpsyg.2019.01351,
Domhoff, G.W. (2018). The emergence of dreaming: Mind-wandering, embodied simulation, and the default network. New York: Oxford University Press.
Hartmann, E. (1996). Outline for a theory on the nature and functions of dreaming. Dreaming, 6, 147–170.
Wamsley, E.J., & Stickgold, R. (2019). Dreaming of a learning task is associated with enhanced memory consolidation: Replication in an overnight sleep study. Journal of Sleep Research, 28, 312749. https://doi.org/10.1111/jsr.12749.
Wamsley, E.J., Tucker, M., Payne, J.D., Benavides, J.A., & Stickgold, R. (2010). Dreaming of a learning task is associated with enhanced sleep-dependent memory consolidation. Current Biology, 20, 850-5. https://doi.org/10.1016/j.cub.2010.03.027.
Zadra, A., & Stickgold, R. (2021). When brains dream: Exploring the science and mystery of sleep. NY: W.W.Norton.