The Multifunctionality of Dreaming and the Oblivious Avatar a Commentary on Revonsuo & Colleagues

The Multifunctionality of Dreaming and the Oblivious Avatar A Commentary on Revonsuo & Colleagues

Martin Dresler

Sleep and dreaming do not serve a single biological function, but are multifunc- Commentator tional. Their functions include memory consolidation and integration, emotion regulation, creativity and problem solving, and preparation for waking life. One prom- Martin Dresler ising level of description is that of dreaming as a virtual reality: The dreamer in- martin.dresler@ donders.ru.nl teracts with a simulated environment including other simulated avatars. While dreaming can be considered a multifunctional general reality simulator, the threat Radboud Universiteit Medical Center simulation and social simulation functions of dreaming are unique among other Nijmegen, Netherlands dream functions in their ability to explain a striking feature of dream phenomenology: obliviousness towards the true state of mind. Target Authors

Katja Valli katval@ utu.fi Turun yliopisto, Turku, Finland Högskolan i Skövde, Skövde, Sweden

Editors

Thomas Metzinger metzinger @uni-mainz.de Johannes Gutenberg-Universität Mainz, Germany

Jennifer M. Windt [email protected] Monash University Melbourne, Australia

1 Introduction

Sleep is an almost ubiquitous phenomenon higher and many lower species. The specific within the animal kingdom, existing in all function of sleep, however, is still an enigma:

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 1 | 18 www.open-mind.net sleep helps an organism to save energy through SST is not the conclusive explanation of the extended periods of inactivity, yet at the same function of dreaming—which I consider a multi- time leaves it in a potentially dangerous state of functional state—, but that they are the only non-responsiveness. While several possible func- candidates among the variety of dream functions of sleep have been discussed in recent tions that are capable of explaining a striking years (Frank 2006; Vassalli & Dijk 2009), the feature of most dreams: obliviousness towards function of dreaming might be seen as an even the current state of mind. bigger mystery: the hyper-realistic imagery experienced during dreaming does not inform the 2 Sleep physiology and the function of organism about its current environment, and dreaming the virtual motor activity processed in interaction with these hallucinations is not executed to When speculating about the function of dream- affect the external world—or even worse, in ing, some clarifications about the level of ex- pathological conditions like REM sleep behavior planation are necessary. By definition (e.g., disorder it is, thereby threatening the health of Windt 2010), dreaming is a phenomenon occur- the dreamer and his bed partner. After awaken- ring during sleep. In an account of biological ing from a dream, the often emotionally-toned realism (Revonsuo 2006), the function of dream- preoccupation with the dream narrative can ing cannot be discussed independently from the confuse the dreamer and distract his from po- neurophysiology of sleep. Even if the phenomen- tentially dangerous conditions in the real world. ology of dreaming serves a function that can be An increasingly widespread idea is that conceptually (and maybe evolutionarily) differ- the function of dreaming consists in the simula- entiated from the original function realized by tion of waking life. In a variation of their threat its physiological correlates, this function is not simulation theory (TST; Revonsuo 1995, 2000), independent from the neurophysiology of sleep Revonsuo et al. (this collection) now propose a and its specific functions: if the neurophysiolo- social simulation theory of dreaming (SST), ac- gical functions change their mechanisms, this cording to which dream function could best be would also affect the phenomenological aspects characterized as simulating social reality. Con- of dreaming—philosophically speaking, phenom- sidering the social nature of most of our enal properties of dreaming supervene on neuro- dreams, SST is an intuitively plausible ap- physiological properties of sleep. However, proach, and Revonsuo et al. review a number of neither can the function of dreaming be equated studies that provide support for SST. Neverthe- with the function of sleep, since there are func- less, several questions remain to be clarified: is tions of sleep for which it is rather unlikely that the prime function of dreaming threat simula- any phenomenological aspects play a role, e.g., tion or social simulation—or something com- myelin sheath proliferation (Bellesi et al. 2013); pletely different? What is the relationship synaptic downscaling (Tononi & Cirelli 2006); between the various proposed functions of sleep metabolite clearance (Xie et al. 2013); or gen- and dreaming, including TST and SST? If the eral metabolic (Morselli et al. 2012) and im- TST and SST turn out not to be the sole or munological functions (Besedovsky et al. 2012). even prime functions of dreaming, do they nev- There are also functions of sleep that might be ertheless provide unique insights into the func- described conceptually without referring to phe- tion of dreaming? nomenal aspects, but in fact happen to be bio- In this commentary, I shall review several logically associated with dream mentation, e.g., widely propagated functions of sleep and physiological microprocesses underlying memory dreaming. I shall then compare these functions consolidation (see below). And in these cases, with the social and threat simulation functions one can differentiate dream phenomenology and of dreaming, and finally discuss why and in sleep physiology on a conceptual, but not biolo- which regard these two functions might be spe- gical level—unless one adopts a radically dual- cial. I shall argue that the merit of TST and istic approach, that is. Hence, speaking of the

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 2 | 18 www.open-mind.net function of dreaming —in contrast to the func- early deep sleep benefits declarative memories, tion of sleep more generally—always implies while late REM-rich sleep supports procedural both phenomenological and physiological as- skills (Plihal & Born 1997). Further support for pects. the role of REM sleep in procedural memory When considering the neurophysiology of consolidation came from studies showing that dreaming, coarse sleep stages as defined by clas- REM sleep intensity (total number of REMs sical polysomnography have been the prime tar- and REM densities) increased following proced- gets of investigation. Among these, REM sleep ural-task acquisition (Smith et al. 2004) and im- harbors the most prototypical dreams, with a provements in procedural memory performance story-like dream narrative including interactive after a night of sleep were proportional to time visuomotor hallucinations and often intense spent in REM sleep (Fischer et al. 2002). emotions. In addition, REM sleep dreams can Moreover, brain areas activated during a pro- be most elegantly related to their neuro- cedural learning task were more active during physiological correlates (Hobson & Pace-Schott REM sleep in subjects who were trained at the 2002). Nevertheless, dream-like mentation can task (Maquet et al. 2000; Peigneux et al. 2003). be found in in all sleep stages (Nielsen 2000), More recent studies, however, speak and hence also the neurophysiology of other against a prominent role of REM sleep in the sleep stages has to be taken into account when consolidation of procedural motor skills or other investigating the function of dreaming. In con- forms of non-emotional memories, and instead clusion, when speculating about the function of emphasize non-REM sleep processes (Genzel et dreaming, all those REM and NREM sleep al. 2014). On the neurophysiological level, it has functions have to be considered that can reason- been suggested that dreaming represents the ably be expected to be associated with phenom- phenomenological reflection of a neural replay of enal aspects. In the following, I will highlight activation patterns associated with recent learn- four clusters of such sleep functions. ing experiences (Wilson & McNaughton 1994; Wamsley & Stickgold 2011; Wamsley 2014). Al- 3 Dream function 1: Memory though memory reactivations have been ob- consolidation and integration served in REM sleep as well (Louie & Wilson 2001), the most advanced models of sleep-re- In recent years, the most widely discussed func- lated memory consolidation propose that neural tion of sleep and dreaming concerns the consol- replay is orchestrated by an interaction of non- idation of declarative memory, including se- REM sleep microprocesses, including slows os- mantic, episodic, and autobiographical informa- cillations and sleep spindles (Genzel et al. tion; and procedural memory including percep- 2014). tual and motor skills (Rasch & Born 2013). In Events and episodes from waking life are particular the role of REM sleep in memory sometimes incorporated into dreams, either as consolidation has been studied for several dec- classical day-residues the following night or ades. While many studies from the 1970s have after a “dream lag” of about 5–7 days (Nielsen been criticized for being heavily confounded by & Powell 1989; Nielsen et al. 2004). Supporting too stressful REM sleep deprivation procedures the idea that such dream incorporations reflect (Horne & McGrath 1984), research in the 1990s processes of memory consolidation, items that raised interest in the role of REM sleep for were incorporated into dreams have been ob- memory consolidation: Karni (1994) demon- served to lead to better memory retention (de strated that a basic visual discrimination task Koninck et al. 1990; Cipolli et al. 2004). While improved after a normal night’s sleep, but not an actual episodic replay of waking events was after selective REM sleep deprivation. Following found in no more than 1–2% of the dream re- this, a leading research aim in the field has been ports (Fosse et al. 2003), with NREM-sleep to identify which memory systems benefit from dreams appearing to include more identifiable which sleep stages: it was demonstrated that episodic memory sources than REM-sleep

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 3 | 18 www.open-mind.net dreams (Baylor & Cavallero 2001), it has been several theoretical considerations and empirical suggested that particularly engaging learning findings are inconsistent with the idea of mne- experiences have a more robust influence on monic encoding strategies acting during dream- dream content relative to more passive experi- ing (Dresler & Konrad 2013). ences (Wamsley 2014). To sum up, a first important function of In contrast to recent episodes, incorpora- sleep and dreaming is memory consolidation tions of autobiographical memory features could and integration, including the rehearsal of pro- be identified in the majority of dreams (Malin- cedural motor skills, replay of episodic and se- owski & Horton 2014). This suggests that mantic memories, and integration of memory dreaming might serve to assimilate recent episodes into autobiographical memory schemas. memory fragments into autobiographical memory schemas and thus supports autobio- 4 Dream function 2: Emotion regulation graphical self-model maintenance (Metzinger 2013). For semantic memories, evidence of a re- Converging evidence suggests that the regula- lationship between dreaming and neural tion of emotional processes is an important memory reactivations stems from studies of de- function of sleep and dreaming. Early content clarative memory that present memory cues analyses of REM sleep dreams showed that during sleep: these cues, when associated with many dreams are highly emotional, with un- the pre-sleep learning session, induce associated pleasant emotions prevailing (Hall & Van de dream imagery (Schredl et al. 2014) and en- Castle 1966; Snyder 1970). This is in line with hance post-sleep memory retrieval (Rasch et al. neuroimaging studies of REM sleep, demon- 2007). For procedural memories, learning of an strating that neural areas involved in emotion engaging visuomotor task led to integration of regulation like the amygdala, medial prefrontal task-related imagery into dream-like activity cortex, and anterior cingulate cortex are highly during non-REM sleep (Wamsley et al. 2010a), activated during REM sleep (Nir & Tononi and such dream-incorporations of recent learn- 2010). Several REM-sleep characteristics differ ing experiences were associated with later between healthy subjects scoring low in depres- memory performance (Wamsley et al. 2010b). sion scales and those with higher but still sub- This memory-enhancing re-experience reminds clinical depression scores (Cartwright et al. us of motor imagery training during wakeful- 1998). After highly emotional life events, REM ness, which has been repeatedly demonstrated sleep changes can be observed in those subjects to improve motor skills (Driskell et al. 1994; that react with symptoms of depression Schuster et al. 2011). (Cartwright 1983), and dreams of depressed Recently it has been suggested that in- subjects differ from patients in remission stead of consolidating memories, REM sleep (Cartwright et al. 2006). Likewise, in depressed serves as a state of elaborative (re-)encoding, patients the distribution of rapid eye move- during which the hippocampus integrates recent ments in REM sleep differs in nights after which episodic memory fragments into remote episodic mood is estimated better than in the preceding memories (Llewellyn 2013). It has been pro- evening compared to nights after which mood is posed that this process relies upon principles unchanged (Indursky & Rotenberg 1998). It was that also underlie the mnemonic encoding therefore proposed that REM sleep dreaming strategies of ancient orators, such as vivid, com- serves as a mood regulation system and that a plex and often bizarre associative imagery, nar- disturbance of this process might play a role in ratives with embodiment of oneself, and associ- the development of affective disorders ations with known locations, later serving as re- (Cartwright 2011). Changes in REM sleep are trieval cues. Subjectively, this process would be symptomatic of affective disorders and the experienced as the typical dream mentation sleep-memory relationship is altered in these with its hyper-associative and bizarre imagery. diseases (Dresler et al. 2014). In healthy sub- However, despite being intuitively appealing, jects, the consolidation of emotional texts

(Wagner et al. 2001) or pictures (Hu et al. 2006; 5 Dream function 3: Creativity and Nishida et al. 2009) is enhanced through REM problem solving sleep, an effect that has been shown to last for several years (Wagner et al. 2006). Anecdotal reports on scientific discovery, invent- While at first sight it might look as if ive originality, and artistic productivity suggest REM sleep unequivocally strengthens emotional that creativity can be triggered or enhanced by memory processes, some studies suggest a more sleeping and dreaming. Several studies confirm complex picture: referring to the fact that emo- these anecdotes, showing that sleep promotes tional experiences are remembered better than creative problem-solving compared to wakeful- neutral ones, however their emotional tone dur- ness. For example, when subjects performed a ing retrieval decreases with time, it was pro- cognitive task that could be solved much faster posed that REM sleep serves an emotional de- through applying a hidden rule, after a night of coupling function: we sleep to remember emo- sleep more than twice as many subjects gained tionally-tagged information yet at the same insight into the hidden rule as in a control time to forget the associated emotional tone group staying awake (Wagner et al. 2004). Sim- (Walker & van der Helm 2009). While some ilarly, subjects benefited in a creativity task studies support this model (Hu et al. 2006; from an afternoon nap but not from staying Nishida et al. 2009), others suggest that the af- awake (Cai et al. 2009; Beijamini et al. 2014), fective tone of emotional memories is preserved and the likelihood of solving a problem en- rather than reduced during REM sleep (Groch countered before sleep can be increased by cued et al. 2013). reactivations during sleep (Ritter et al. 2012). Besides negative emotions, sleep and According to the classical stage model of dreaming have also been associated with pos- creativity, creative insights may be described by itive affects. Recent dream report analyses a process consisting of several stages, of which suggest that positive emotions in dreams have the incubation phase appears to be most intim- been underestimated in previous studies and ately associated with sleep and dreaming might be even more common than negative (Dresler 2011, 2012; Ritter & Dijksterhuis emotions (Malcolm-Smith et al. 2012; Sikka et 2014). The most common psychological ap- al. 2014). In addition, the processing of re- proaches support this view: psychoanalytical ward has been associated with REM sleep and models of creativity emphasize the primary pro- dreaming. For example, the expectancy of a cess concept, which denotes free-associative and reward enhances memory consolidation pro- dream-like thinking, compared to the more ra- cesses during sleep (Fischer & Born 2009), tional and analytical secondary-process thinking and reactivations of neural activity related to (Kris 1952). Cognitive models propose that a a reward-searching task have been observed in state of defocused attention facilitates creativity reward-related brain regions such as the vent- (Mendelsohn 1976)—creative individuals seem ral striatum during sleep (Pennartz et al. to have less narrowly-focused attention than un- 2004). Instead of a simulation of purely avers- creative ones, which leads to unorthodox con- ive content such as threats, according to this nections of remote ideas that might eventually account sleep favors the activation of repres- lead to creative cognitions. In a similar vein, entations of high emotional and motivational creative individuals are thought to have relat- relevance in general (Perogamvros & Schwartz ively flat association hierarchies (i.e., more, yet 2012, 2014). weaker associations between cognitive ele- In summary, a second important function ments), which accounts for the ability to make of sleep and dreaming is the regulation of emo- remote associations; whereas uncreative indi- tions, including both an enhancement of emo- viduals are thought to have relatively steep as- tionally-tagged information and a decoupling of sociation hierarchies (Mednick 1962). Physiolo- this information from its associated emotional gical models emphasize the level of cortical tone. arousal as an important variable influencing cre-

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 5 | 18 www.open-mind.net ativity: both a lower level of cortical arousal— innervations by the brainstem, leading to particularly in the prefrontal cortex—and a bizarre sequences of loosely associated dream higher variability in cortical arousal levels are topics that might eventually activate particular expected in creative compared to uncreative in- problem-relevant cognitions or creative cogni- dividuals, depending on specific phases of the tions in general (Hobson & Wohl 2005). In addi- creative process (Martindale 1999). In addition, tion, in REM sleep, which is characterized by low levels of norepinephrine are thought to facil- low levels of norepinephrine, visual cortices, the itate creativity, shifting the brain toward in- hippocampus, and the anterior cingulate cortex trinsic neuronal activation with an increase in have all been shown to be strongly activated, the size of distributed concept representations potentially facilitating insight events. In conclu- and co-activation across modular networks sion, the phenomenological and neural correl- (Heilman et al. 2003). The prefrontal cortex ates of sleeping and dreaming provide ideal con- seems to be of particular importance for creat- ditions for the genesis of creative ideas and in- ive processes; however there is evidence that sights. both prefrontal activation and prefrontal deac- In summary, a third important function of tivation facilitate creativity—maybe depending sleep and dreaming is the association of remote on the specific phase of the creative process. cognitive elements in order to facilitate creativ- Brain areas showing selective activation for in- ity and problem solving. sight events are—besides the prefrontal cortex— the visual cortices, the hippocampus, and in 6 Dream function 4: Preparation and particular the anterior cingulated cortex, which simulation of waking life is thought to be involved in breaking the im- passe that marks the critical step of insight into Consolidation, integration, regulation, and re- a problem (Dietrich & Kanso 2010). evaluation of acquired information during sleep Both theoretical models and empirical prepare the organism for its waking life. How- neuroscience of creativity suggest that sleep and ever, such processes do not necessarily need to dreaming provide an ideal environment for cre- be purely reactive, depending solely on the ex- ative incubation: primary-process thinking is experiences of the preceding day: several authors plicitly conceptualized as dream-like, and the propose that a major function of sleep and hyper-associative nature of dreams can be con- dreaming might include primarily preparational sidered a prime example of a flat associative mechanisms. Since REM sleep dominates sleep hierarchy. Defocused attention is a phenomenal more during early developmental periods in feature of most dreams, physiologically probably comparison to later in life, some researchers caused by prefrontal cortex deactivation. And have argued that REM sleep plays a role in daydreaming has the potential to increase cre- early brain maturation (Roffwarg et al. 1966; ativity (Lewin 1989), while the level of engage- Marks et al. 1995; Mirmiran 1995); however, ment in such mind-wandering in contrast to ex- also a life-long preparational function of REM plicitly directed thoughts is associated with cre- sleep has been proposed. One of the first ap- ative performance (Baird et al. 2012). The sleep proaches in this direction was offered by Jouvet cycle provides the brain with highly alternating (1979), who combined the brain maturation hy- arousal levels, and the chaotic activation of the pothesis with a metaphor offered by Dewan cortex in REM sleep through brain stem regions (1970), in which he claims that the brain is a in absence of external sense data leads to a computer that is programmed during REM much more radical renunciation of unsuccessful sleep—suggesting that innate behaviors are re- problem solving attempts, leading to co-activa- hearsed during REM-sleep dreaming in order to tions of cognitive data that are highly remote in prepare the organism for their application in waking life (Kahn et al. 2002a). These co-activ- waking life. Jouvet later revised his approach, ations, woven into a dream narrative in a self- assuming that REM sleep constitutes an iterat- organizing manner, repeatedly receive further ive genetic programming that helps to maintain

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 6 | 18 www.open-mind.net the process of psychological individuation are supported by a number of studies, but face (Jouvet 1998). In a similar vein, Hobson (2009) inconsistent data (Revonsuo et al. this collec- proposed that REM sleep may constitute a tion). “protoconscious” state, preparing the organism On a neurobiological level, empirical sup- for waking conscious experiences. The develop- port for simulation theories of dreaming comes ment of consciousness during ontogenetic devel- from a recent study demonstrating that the opment in this view is a gradual and lifelong ventromedial prefrontal cortex subserves the process, building on the more primitive innate simulation and evaluation of possible future ex- virtual reality generator, which is phenomenally periences, integrating arbitrary combinations of experienced as dreaming. With the recent integ- knowledge structures to simulate the emergent ration of Friston’s (2010) predictive coding ap- affective quality that a possible future episode proach into this theory, the brain is thought to may hold (Benoit et al. 2014). As the ventro- run a virtual world model (see also Revonsuo medial prefrontal cortex is known to be activ- 1995, 2006; Metzinger 2003) that is continu- ated in REM sleep (Nir & Tononi 2010), this ously updated by processing prediction errors mechanism might also underlie episodes of real- during wakefulness. Freed from external sensory ity simulation during dreaming. Further neuro- constraints, processing of prediction errors in biological support for the preparational role of the dreaming brain actively refines intermediate sleep comes from recent research demonstrating hierarchy levels of the virtual world model. a neural “preplay” of future learning-related Dreaming thereby minimizes internal model place-cell sequences in the hippocampus (Dragoi complexity in order to generate more efficient & Tonegawa 2011, 2013). In contrast to the in- predictions during subsequent wakefulness tuitive view that such activation patterns are (Hobson & Friston 2012; Hobson et al. 2014). established for the first time during a novel ex- One of the first and today the most widely perience, according to these findings the specific discussed preparational approach is based on temporal firing sequence during learning seems the observation that during dreaming particu- rather to be selected from a larger repertoire of larly threatening experiences are overrepresen- preexisting activation patterns, thus suggesting ted: the Threat Simulation Theory (TST) pro- that sleep plays a role not only in the sub- poses that one function of sleep is to simulate sequent consolidation, but also in the preceding threatening events, and to rehearse threat per- preparation for new experiences. It has been ception and threat avoidance (Revonsuo 1995, demonstrated that sleep preceding the learning 2000). Such a mechanism of simulating the experience indeed influences memory acquisition threats of waking life over and over again in during the following day (van der Werf et al. various combinations would be valuable for the 2009). Interestingly, support for the hypothesis development and maintenance of threat-avoid- that sleep mentation constitutes a virtual real- ance skills. Several empirical studies support ity model preparing for waking life comes also TST (Revonsuo 2006; Valli & Revonsuo 2009), from research outside of sleep neuroscience: ap- however some inconstant findings have been re- proaches probing artificial intelligence demon- ported (Zadra et al. 2006; Malcolm-Smith et al. strate that robots perform better in naviga- 2008, 2012). In a variation of TST, Revonsuo et tional tasks if they create and update models of al. (this collection) propose the Social Simula- their own structure and actions during a state tion Theory (SST), according to which the func- of motoric inactivity (Bongard et al. 2006). Not tion of dreaming consists in the simulating of surprisingly, this process of evaluation and sim- “the social skills, bonds, interactions and net- ulation of prior and future actions was inter- works that we engage in during our waking preted as dream-like (Adami 2006). lives”. The SST aims to predict and explain the In summary, a fourth important function simulations of social interaction of dream of sleep and dreaming is preparation for waking avatars that happen outside threatening events life. This includes proposals of REM sleep as an in dreams. Like the TST, predictions of the SST iterative genetic programming system, dreaming

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 7 | 18 www.open-mind.net as a state of protoconsciousness and virtual Both principles, dubbed natural and sexual se- world model optimization, and dreaming as a lection respectively, eventually determine repro- simulation of threats (TST) and social interac- ductive success as the ultimate decision points tions (SST). for selection. In contemporary accounts, sexual selection was generalized to the concept of so- 7 The multifunctionality of dreaming cial selection, of which the former is considered a subtype (Lyon & Montgomerie 2012; West- Numerous suggestions for solving the mystery of Eberhard 2014). The concept of runaway selec- sleep and dream function can be found in the tion, famously illustrated by the evolution of literature. In the previous sections I have re- the peacock’s tail, was thought to also be ap- viewed four clusters of proposed functions of plicable to the evolution of social skills in higher sleep and dreaming: 1) consolidation of recently animals, eventually leading to the development acquired memories, including procedural motor of theory of mind, language, dance, or artistic skill rehearsal, replay of recently acquired creativity in humans (Flinn & Alexander 2007). memories, and integration of memory episodes This process of an arms race of social skills into autobiographical memory schemas; 2) emo- would require increasing cognitive capacity— tion regulation, including both an enhancement and in fact, at least in primates, relative brain of emotionally-tagged information and a de- size has been related to social group size (Dun- coupling of this information from its associated bar 1992; Dunbar & Shultz 2007). emotional tone; 3) creativity and problem solv- It is tempting to associate natural and so- ing; and (4) preparation and simulation of wak- cial selection as the main principles of evolution ing life, including iterative genetic program- with TST and SST, respectively. This interpret- ming, virtual world model optimization, the ation would strongly support TST and SST, as simulation of threats (TST), and the simulation it would equate the function of dreaming with of social interactions (SST). The question thus two main principles of evolution in general. In remains what the real or primary function of this broad sense, however, certain attributes like sleep and dreaming is—and what the relation- learning capacity or motor skills increase fitness ship between the different candidates might be. in terms of natural selection, but do not neces- SST aims to independently cover the social sim- sarily serve to help us avoid direct threats. Like- ulations that fall outside the scope of TST, wise, certain attributes such as emotion regula- thereby describing an “original evolutionary tion or artistic creativity increase fitness in function of dreams alongside with the threat terms of social selection, but are not necessarily simulation function of dreaming” (Revonsuo et themselves social in a strict sense. Ultimately, of al. this collection). course, all these functions serve reproductive The concept of evolutionary function has success—however, if any skill ultimately helping been one of the main topics in the philosophy of us to acquire sexual partners is interpreted as biology (Mahner & Bunge 2000) and philosophy social and any possible obstacle to reproduction of mind (Millikan 1984; Neander 1991). Several is interpreted as a threat, then TST and SST notions of biological functions exist (Wouters would be trivial, as a biological function is by 2003); however a general idea is that the biolo- definition one that supports reproductive suc- gical function of a trait is determined by its cess. In contrast, if TST and SST are inter- contribution to evolutionary fitness (Walsh & preted in a more narrow, non-trivial way, there Ariew 1996). Darwin (1871) differentiated is ample space in dreams for further functions: between selection occurring as a consequence of consolidation of navigational information ac- ecological factors that directly threaten the or- quired during exploration; rehearsal of a re- ganism’s survival, such as predators or other cently learned motor sequence; facilitation of a potentially life-threatening dangers of nature, behavior recently rewarded with food; incub- and interactions with members of the same spe- ated creative insight into the solution of a re- cies in order to compete for mating partners. cent unsuccessful attempt to build a helpful

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 8 | 18 www.open-mind.net tool; refinement of the discriminative skills re- social interaction would be the two main drivers garding recently perceived pattern, etc.—all of evolution, however to the cost that the an- these potential benefits of sleep and dreaming swer to the question of the function of dreaming increase inclusive fitness of the individual, but becomes a trivial “to support reproductive suc- do not directly refer to the simulation of threats cess”. Of note is that also the other discussed or social interactions. functions might be interpreted within a simula- This problem can further be illustrated by tion framework: e.g., simulation visuomotor Revonsuo’s (1995, 2006) approach, where he activity after learning a respective task in the considers any phenomenal experience as a vir- memory function, simulating affective experi- tual world model: what is the function of wak- ences in the emotion regulation function, and ing consciousness, threat avoidance, or social in- simulating problem solving attempts in the cre- teraction? Both threat avoidance and social in- ativity function. These different functions are teraction, of course—and many others. That neither mutually exclusive nor strictly inde- this rather uninformative answer can also be pendent from each other. In particular the emo- transferred back to the function of dreaming tion-processing function largely overlaps with might be illustrated with another ubiquitous ex- both TST an SST—all threats and at least the ample of simulation: in child’s play, simulation most important social interactions induce strong of real life and the practice of skills needed emotions, and successful coping with these emo- therein is considered one of the main functions tions would be of considerable help when facing —play allows children to simulate coping with threats or social situations. Also other functions threats in a safe environment, and to develop of dreaming overlap with TST and SST: consol- the social skills needed later in life (Mellou idation of threat-related information or social 1994; Pellegrini & Bjorklund 2004). However, gossip improves threat avoidance or social skills, these aspects, while important, are not the only as does creative incubation on threat-related or functions of play—it also offers the rehearsal of social problems. On a more abstract level, all motor and sensory skills, training in predatory these simulations serve the integration of re- behavior, and general intellectual development. cently experienced information into the behavi- Hence, child’s play can be considered multifunc- oral repertoire in order to adapt it to the cur- tional, as can waking or dreaming conscious- rent waking environment (Hobson et al. 2014). ness. Identifying the original function of a given Segmentation of reality (including dream trait has proven to be a notoriously difficult is- reality) is possible along numerous lines. In a sue in the philosophy of biology (Wouters 2013). sense, TST and SST could be interpreted as ex- Dreaming might have originally developed as an pressing two orthogonal dimensions of dream epiphenomenon of rather basal neurophysiolo- space: a security dimension with the directions gical sleep functions, and this phenomenological threat vs. safety, and a sociality dimension with level might eventually have acquired additional the directions social vs. individual. Dreamed ac- functions. Such exaptations (Gould & Vrba cidents or natural disasters would be character- 1982) might have been further adapted and in ized by low security and sociality, dreamed ex- turn developed further neurophysiological experience of bullying by high sociality and low aptations without phenomenological correlates, security, and dreamed bonding by high sociality etc. The original function of dreaming might be and security, etc. Threat and social interactions unimportant today compared to subsequently in a narrow sense are important aspects both of evolved functions. Instead of singling out one or waking and dreaming life, however they are not two functions of dreaming as original, dreaming the only aspects. Other segmentations are also might be best seen as a multifunctional general possible, e.g., by a dimension of motor activity reality simulator, including the simulation of vs. inactivity, or emotional vs. neutral dream motor skills, emotional processing, problem content, or a novelty dimension. In the broad solving attempts, threats, and social interac- sense of natural and social selection, threat and tions. To follow specific research questions, of

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 9 | 18 www.open-mind.net course certain functions still could be high- over the other of its several functions, depend- lighted and followed as research heuristics with ing on the current requirements and constraints a given purpose. All functions of sleep and of the environment. dreaming serve reproductive success ultimately, even though some might be more important 8 The oblivious avatar than others from a selection point of view. For all dream functions discussed in this chapter, Even though it is likely that no ‘original’ func- there are convincing supporting but also incon- tion of dreaming can be acknowledged, but sistent data. The fact that dreaming is not an rather a multiplicity of functions depending on unselective simulation of the waking world as, specific research questions and segmentations of e.g., the continuity hypothesis suggests (Schredl the dream space, one aspect of dreaming might & Hofmann 2003), is a sign that some simula- distinguish TST and SST from other functions tion functions might be more important than of sleep and dreaming, including other simula- others. We should note, however, that quantit- tion functions: obliviousness of the avatar about ative overrepresentation of a specific function being in a dream. Impaired insight into the own does not necessarily prove the primacy of this state of mind is a hallmark of normal dreaming, function: different functions might rely on dif- (Dresler et al. 2015a). The well-known excep- ferent processes with different timescales, with a tion of this symptom of most dreams is the case highly important function potentially requiring of lucid dreaming (Dresler et al. 2015), which in only seconds to be processed, while an unim- turn can be used to test whether state oblivi- portant function might take hours. In times of ousness is indeed a characterizing feature of sufficient sleep, dream content related to the re- TST and SST when compared with other dream latively unimportant function might thus be functions. overrepresented. The relative importance of one There is no obvious reason why oblivious- function over another might be tested in cases ness about the dream state would be necessary of scarcity of sleep, e.g., under sleep depriva- for the memory function of sleep and dreaming. tion, when different functions would have to For procedural memory consolidation, lucid compete for restricted simulation time. Also of dreaming has even been suggested as a state interest in this regard is a comparative ap- that allows for a hyper-realistic mental training proach: it has been demonstrated that sleep of recently learned motor skills (Erlacher & propensity, and particularly REM sleep, negat- Chapin 2010). Several studies support this idea: ively correlates with predatory risk across spe- lucidly dreamed training of coin tossing (Er- cies (Lima et al. 2005), which would rather lacher & Schredl 2010) or a finger tapping task speak against TST. Concerning SST, the tend- (Stumbrys et al. 2015) has been demonstrated ency to sleep in groups has been reported to to be effective, and a considerable number of negatively correlate with sleep time, which, professional athletes use lucid dreams to prac- however, has been interpreted either in terms of tice sports skills, with most of them having the social sleep being more efficient due to reduced impression that their performance is thereby predatory risk, or as more social species sacrifi- improved (Erlacher et al. 2011). For the creativ- cing sleep to service social relationships during ity and insight function of sleep and dreaming, wakefulness (Capellini et al. 2008). Against this obliviousness regarding the current state of background, sleep and dreaming pose an optim- mind is no prerequisite, and lucid dreaming has ization problem: how much time is best spent explicitly been suggested and shown to be used asleep, spent in specific sleep stages, and spent as a tool to increase creative processes (Stum- engaging in specific dream mentation in order brys & Daniels 2010; Schädlich & Erlacher to optimize the interplay between the different 2012; Stumbrys et al. 2014). As with non-lucid functions of sleep and dreaming? Dreaming as a dreaming, lucid dreaming is associated with de- general reality simulator might dynamically focussed attention and flat association hierarch- change its functional priorities, favoring one ies—lucid dreams have been reported to include

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 10 | 18 www.open-mind.net even more uncommon and bizarre elements ized for recurrent nightmares, where lucid than non-lucid dreams (McCarley & Hoffman dreaming has been demonstrated to be of thera- 1981). At the same time, regained reflective peutic value (Spoormaker et al. 2003, 2006; capabilities enable the creative dreamer to eval- Dresler et al. 2015; Rak et al. in press). Thus, uate new associations and ideas, a step in the for the threat simulation function of dreaming, phase model of creativity that for non-lucid obliviousness regarding the current state of dreams is reserved for subsequent wakefulness. mind is essential. This mechanism is illustrated by two interesting For SST, several lines of evidence indicate case studies: Barrett (2001) describes the case that obliviousness regarding the current state of of a painter who in his lucid dreams visited gal- mind is a prerequisite for social simulation to be leries, and then searching for interesting motifs effective. During normal dreams, non-self dream to be painted soon after awakening from the lu- characters are attributed with feelings and cid dream. A comparable strategy was used by thoughts just like in waking life (Kahn & Hob- one of our own study participants (Dresler et al. son 2005). Being oblivious about the true 2011, 2012), a music composer: when he aimed nature of these dream characters might ensure to compose a new piece of music, he turned on that non-perfect social simulations are also a radio in his lucid dreams and changed radio taken as autonomous agents instead of mere stations until he heard a composition that he puppets controlled by the dreamer: dream char- considered interesting. He then woke himself up acters are often implausible compared to their and wrote the new composition down. In line real-life waking counterparts (Kahn & Hobson with these data, questionnaire studies reported 2003) , however, are nevertheless recognized and that frequent lucid dreamers might be more cre- identified without major puzzlement (Kahn et ative than less-frequent lucid dreamers al. 2000, 2002b). During a lucid dream, im- (Blagrove & Hartnell 2000). plausible dream characters might be treated less For the emotion regulation function of seriously by the dreamer, rendering the social sleep and dreaming the situation is less clear, simulation much less effective. This is illus- however here there is also some evidence indic- trated by a recent study demonstrating that be- ating that obliviousness is not generally neces- ing tickled by an intentionally-controlled non- sary: for the case of positive affects, subjects of- self dream character during a lucid dream was ten report that lucid dreams are associated with comparably ineffective as self-tickling during particularly positive emotions. And for negative wakefulness, whereas being unexpectedly tickled affects, the successful use of lucid dreaming as a by another dream character felt more ticklish therapeutic tool in affective disorders indicates (Windt et al. 2014). Non-self dream characters that dream lucidity does not interfere with the lead to different predictions depending on their emotion regulation function of dreaming perceived autonomy, and their respective simu- (Holzinger 2014). lation thus serves different functions. Lucid In contrast, for those cases where a general dreaming frequency correlates with the amount emotion regulation function of dreaming over- of control over the dream (Wolpin et al. 1992; laps with the TST, the necessity of staying ig- Stumbrys et al. 2014), implying that frequent norant about the true state of consciousness be- lucid dreamers would conceive dream characters comes obvious: to successfully serve as an au- as less autonomous than less frequent lucid thentic simulation of a threat, the dreamer has dreamers. Thus, although non-self dream char- to take the threat as real and thus be oblivious acters appear to have quasi-independent mental towards his true state of mind. The cognitive lives during lucid dreams (Tholey 1989), convin- insight that everything encountered consists cing training of social skills would require the only of hallucinated dream imagery and thus dreamer to be oblivious to the fact that dream cannot harm the dreamer in reality immediately characters are not real, but hallucinated. takes the sting out of the threatening experi- In summary, in contrast to other functions ence. This mechanism has been successfully util- of sleep and dreaming, TST and SST essentially

Dresler, M. (2015). The Multifunctionality of Dreaming and the Oblivious Avatar - A Commentary on Revonsuo & Colleagues. In T. Metzinger & J. M. Windt (Eds). Open MIND: 32(C). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570672 11 | 18 www.open-mind.net depend on state obliviousness of the dreamer. sent in dreaming—dream lucidity is a rare phe- State obliviousness in dreaming might therefore nomenon (Schredl & Erlacher 2011), and even be seen as a prime example of an epistemic con- during lucid dreams, lucidity lapses are common straint of phenomenal experience that leads to (Barrett 1992)—, suggests that state oblivious- new and beneficial functional properties (Met- ness during dreaming has an important func- zinger 2003). While both TST and SST (and tion. As demonstrated here, among the different other functions of sleep) might be applicable to candidates for explaining the function of dream- humans and other social animals alike, state ob- ing, TST and SST are the only ones that are liviousness might be a function that specifically capable of elucidating this specific function: developed in humans: it is unlikely that animals state obliviousness is necessary for the effective without sophistic language skills possess the simulation of threats and social interactions. ability to reflect on their current state of mind Even though recent neurobiological re- and compare it to alternative mind-states. In search has begun to reveal the neural correlates turn, such animals do not need a differential of state reflectiveness and, by contrast, of state mechanism switching state reflectiveness on and obliviousness (Voss et al. 2009, 2014; Dresler et off depending on the current vigilance state. Of al. 2012), the specific neural mechanisms pre- note, neural correlates of state reflectiveness, i.e. venting the dreaming brain from realizing its lucid dreaming, strikingly mirror brain differ- full repertoire of cognitive capabilities are still ences seen in humans vs. non-human primates largely unclear. Further research into these (Dresler et al. 2013). mechanisms might enable exciting opportunities for sleep and dream research by revealing simple 9 Conclusion methods of dream-lucidity induction. However, if such ways to induce a simulated reality under Sleep and dreaming do not serve a single biolo- full control of its user become available too eas- gical function, but are multifunctional states. ily and broadly, this might also lead to unfore- Their functions include memory consolidation seen problems, as at least two important func- and integration, emotion regulation, creativity tions of dreaming—simulation of threats and so- and problem solving, and preparation for wak- cial interactions—probably cannot be processed ing life. One promising description level is that without state obliviousness. This proposed ne- of dreaming as a general reality simulator. TST cessity generates a testable hypothesis: individu- and SST describe two important purposes of als with very frequent lucid dreams can be ex- simulation, namely successful coping with pected to differ from the majority of infrequent threats and social interactions. The merit of lucid dreamers in their threat-avoidance and so- TST and SST is not so much that they conclus- cial skills. ively explain the function of dreaming—although they represent the two classical principles of evolution, natural and social selection, there are also several other sleep and dream functions. TST and SST might be the only candidates among the multiple functions of sleep and dreaming that explain a particularly striking feature of dream phenomenology: dreaming is a remarkably realistic simulation of waking life, with the exception of a complete failure to successfully reflect on the current state of consciousness. Veridical insight into the dream state is biologically possible, as the phenomenon of lucid dreaming demonstrates. The fact that state reflectiveness is nevertheless generally ab-

References Capellini, I., Barton, R. A., McNamara, P., Preston, B. T. & Nunn, C. L. (2008). Phylogenetic analysis of the Adami, C. (2006). What do robots dream of? Science, ecology and evolution of mammalian sleep. Evolution, 314 (5802), 1093-1094. 10.1126/science.1135929 62 (7), 1764-1776. 10.1111/j.1558-5646.2008.00392.x Baird, B., Smallwood, J., Mrazek, M. D., Kam, J. W., Cartwright, R. D. (1983). Rapid eye movement sleep Franklin, M. S. & Schooler, J. W. (2012). Inspired by characteristics during and after mood-disturbing distraction: Mind wandering facilitates creative incuba- events. Archives of General Psychiatry, 40 (2), 197-201. tion. Psychological Science, 23 (10), 1117-1122. 10.1001/archpsyc.1983.01790020095009 10.1177/0956797612446024 (2011). Dreaming as a mood regulation system. In Barrett, D. (1992). Just how lucid are lucid dreams? M. H. Kryger, T. Roth & W. C. Dement (Eds.) Prin- Dreaming, 2 (4), 221-228. ciples and Practice of Sleep Medicine (pp. 620-627). St. (2001). The committee of sleep. Norwalk, CT: Louis, MO: Saunders. Crown House Publishing. Cartwright, R., Luten, A., Young, M., Mercer, P. & Baylor, G. W. & Cavallero, C. (2001). Memory sources Bears, M. (1998). Role of REM sleep and dream affect associated with REM and NREM dream reports in overnight mood regulation: A study of normal volun- throughout the night: A new look at the data. Sleep, teers. Psychiatry Research, 81 (1), 1-8. 24 (2), 165-170. 10.1016/S0165-1781(98)00089-4 Beijamini, F., Pereira, S. I., Cini, F. A. & Louzada, F. M. Cartwright, R., Agargun, M. Y., Kirkby, J. & Friedman, (2014). After being challenged by a video game prob- J. K. (2006). Relation of dreams to waking concerns. lem, sleep increases the chance to solve it. PLoS One, Psychiatry Research, 141 (3), 261-270. 9 (1), e84342-e84342. 10.1371/journal.pone.0084342 10.1016/j.psychres.2005.05.013 Bellesi, M., Pfister-Genskow, M., Maret, S., Keles, S., Cipolli, C., Fagioli, I., Mazzetti, M. & Tuozzi, G. (2004). Tononi, G. & Cirelli, C. (2013). Effects of sleep and Incorporation of presleep stimuli into dream contents: wake on oligodendrocytes and their precursors. The Evidence for a consolidation effect on declarative know- Journal of Neuroscience, 33 (36), 14288-300. ledge during REM sleep? Journal of Sleep Research, 13 10.1523/JNEUROSCI.5102-12.2013 (4), 317-326. Benoit, R. G., Szpunar, K. K. & Schacter, D. L. (2014). Darwin, C. (1871). Sexual selection and the Descent of Ventromedial prefrontal cortex supports affective fu- Man. London, UK: Murray. ture simulation by integrating distributed knowledge. de Koninck, J., Christ, G., Hébert, G. & Rinfret, N. Proceedings of the National Academy of Sciences of the (1990). Language learning efficiency, dreams and REM U.S.A., 111 (46), 16550-16555. sleep. Psychiatric Journal of the University of Ottawa, 10.1073/pnas.1419274111 15 (2), 91-92. Besedovsky, L., Lange, T. & Born, J. (2012). Sleep and im- Dewan, E. M. (1970). The programing (P) hypothesis for mune function. Pflügers Archiv European Journal of REM sleep. International Psychiatry Clinics, 7 (2), Physiology, 463 (1), 121-137. 10.1007/s00424-011-1044-0 295-307. Blagrove, M. & Hartnell, S. J. (2000). Lucid dreaming: Dietrich, A. & Kanso, R. (2010). A Review of EEG, ERP, Associations with internal locus of control, need for and Neuroimaging Studies of Creativity and Insight. cognition and creativity. Personality and Individual Psychological Bulletin, 136 (5), 822-848. Differences, 28 (1), 41-47. 10.1037/a0019749 10.1016/S0191-8869(99)00078-1 Dragoi, G. & Tonegawa, S. (2011). Preplay of future place Bongard, J., Zykov, V. & Lipson, H. (2006). Resilient ma- cell sequences by hippocampal cellular assemblies. Nature, chines through continuous self-modeling. Science, 314 469 (7330), 397-401. 10.1038/nature09633 (5802), 1118-1121. 10.1126/science.1133687 (2013). Distinct preplay of multiple novel spatial Cai, D. J., Mednick, S. A., Harrison, E. M., Kanady, J. experiences in the rat. Proceedings of the National C. & Mednick, S. C. (2009). REM, not incubation, im- Academy of Sciences of the U.S.A., 110 (22), 9100- proves creativity by priming associative networks. Pro- 9105. 10.1073/pnas.1306031110 ceedings of the National Academy of Sciences of the Dresler, M. (2011). Kreativität, Schlaf und Traum – U.S.A., 106 (25), 10130-10134. Neurobiologische Zusammenhänge. In K. Hermann 10.1073/pnas.0900271106 (Ed.) Neuroästhetik (pp. 32-44). Kassel.

(2012). Sleep and creativity. Theoretical models Erlacher, D., Stumbrys, T. & Schredl, M. (2011). Fre- and neural basis. In D. Barrett & P. McNamara (Eds.) quency of lucid dreams and lucid dream practice in Encyclopedia of Sleep and Dreams Vol II.. Santa Bar- German athletes. Imagination, Cognition and Person- bara. ality, 31, 237-246. 10.2190/IC.31.3.f Dresler, M., Koch, S., Wehrle, R., Spoormaker, V. I., Erlacher, D. & Schredl, M. (2010). Practicing a motor Holsboer, F., Steiger, A., Sämann, P. G., Obrig, H. & task in alucid dream enhances subsequent performance: Czisch, M. (2011). Dreamed movement elicits activa- A pilotstudy. The Sport Psychologist, 24 (2), 157-167. tion in the sensorimotor cortex. Current Biology, 21 Fischer, S. & Born, J. (2009). Anticipated reward en- (21), 1833-1837. 10.1016/j.cub.2011.09.029 hances offline learning during sleep. Journal of Experi- Dresler, M., Wehrle, R., Spoormaker, V. I., Holsboer, F., mental Psychology. Learning, Memory and Cognition, Steiger, A., Koch, S., Obrig, H., Sämann, P. G. & 35 (6), 1586-1593. 10.1037/a0017256 Czisch, M. (2012). Neural correlates of dream lucidity Fischer, S., Hallschmid, M., Elsner, A. L. & Born, J. obtained from contrasting lucid versus non-lucid REM (2002). Sleep forms memory for finger skills. Proceed- sleep: A combined EEG/fMRI case study. Sleep, 35 (7), ings of the National Academy of Sciences of the 1017-1020. 10.5665/sleep.1974 U.S.A., 99 (18), 11987-11991. 10.1073/pnas.182178199 Dresler, M., Eibl, L., Fischer, C., Wehrle, R., Spoor- Flinn, M. V. & Alexander, R. D. (2007). Runaway social maker, V. I., Steiger, A., Czisch, M. & Pawlowski, M. selection in human evolution. The Evolution of Mind (2013). Volitional components of consciousness during (pp. 249-255). New York, NY: Guilford Press. wakefulness, dreaming and lucid dreaming. Frontiers in Fosse, M. J., Fosse, R., Hobson, J. A. & Stickgold, R. J. Psychology, 4, 987-987. 10.3389/fpsyg.2013.00987 (2003). Dreaming and episodic memory: A functional Dresler, M., Spoormaker, V. I., Beitinger, P. A., Czisch, dissociation? Journal of Cognitive Neuroscience, 15 (1), M., Kimura, M., Steiger, A. & Holsboer, F. (2014). 1-9. 10.1162/089892903321107774 Neuroscience-driven discovery and development of sleep Frank, M. G. (2006). The mystery of sleep function: Current therapeutics. Pharmacology & Therapeutics, 141 (3), perspectives and future directions. Reviews in the Neur- 300-334. 10.1016/j.pharmthera.2013.10.012 osciences, 17 (1), 375-92. 10.1515/revneuro.2006.17.4.375 Dresler, M., Erlacher, D., Czisch, M. & Spoormaker, V. I. Friston, K. (2010). The free-energy principle: A unified (2015). Lucid dreaming. In M. Kryger, T. Roth & W. brain theory? Nature Reviews Neuroscience, 11, 127- Dement (Eds.) Principles and Practice of Sleep Medi- 138. 10.1038/nrn2787 cine. Amsterdam, NL: Elsevier. Genzel, L., Kroes, M. C., Dresler, M. & Battaglia, F. P. Dresler, M., Wehrle, R., Spoormaker, V. I., Holsboer, F., (2014). Light sleep versus slow wave sleep in memory Steiger, A., Czisch, M. & Hobson, J. A. (2015a). consolidation: A question of global versus local pro- Neural correlates of insight in dreaming and psychosis. cesses? Trends in Neurosciences, 37 (1), 10-19. Sleep Medicine Reviews 20, 92-99. 10.1016/j.tins.2013.10.002 Dresler, M. & Konrad, B. N. (2013). Mnemonic expertise Gould, S. J. & Vrba, E. S. (1982). Exaptation – a missing during wakefulness and sleep. Behavioral and Brain term in the science of form. Paleobiology, 8 (1), 4-15. Sciences, 36 (6), 616-617. 10.1017/S0140525X13001301 Groch, S., Wilhelm, I., Diekelmann, S. & Born, J. (2013). Driskell, J. E., Copper, C. & Moran, A. (1994). Does mental The role of REM sleep in the processing of emotional practice enhance performance? Journal of Applied Psy- memories: Evidence from behavior and event-related chology, 79 (4), 481-492. 10.1037/0021-9010.79.4.481 potentials. Neurobiology of Learning and Memory, 99, Dunbar, R. I. (1992). Neocortex size as a constraint on 1-9. 10.1016/j.nlm.2012.10.006 group size in primates. Journal of Human Evolution, Hall, S. C. & van de Castle, R. I. (1966). The content 22 (6), 469-493. 10.1016/0047-2484(92)90081-J analysis of dreams. New York, NY: Appleton-Century- Dunbar, R. I. & Shultz, S. (2007). Evolution in the social Crofts. brain. Science, 317 (5843), 1344-1347. Heilman, K. M., Nadeau, S. E. & Beversdorf, D. O. (2003). 10.1126/science.1145463 Creative innovation: Possible brain mechanisms. Erlacher, D. & Chapin, H. (2010). Lucid dreaming: Neurocase, 9 (5), 369-379. 10.1076/neur.9.5.369.16553 Neural virtual reality as a mechanism for performance Hobson, J. A. (2009). REM sleep and dreaming: Towards enhancement. International Journal of Dream Re- a theory of protoconsciousness. Nature Reviews Neur- search, 3 (1), 7-10. 10.11588/ijodr.2010.1.588 oscience, 10 (11), 803-813. 10.1038/nrn2716

Hobson, J. A., Hong, C. C. & Friston, K. J. (2014). Vir- Kahn, D. & Hobson, A. (2003). State dependence of char- tual reality and consciousness inference in dreaming. acter perception. Journal of Consciousness Studies, 10 Frontiers in Psychology, 5, 1133-1133. (3), 57-68. 10.3389/fpsyg.2014.01133 (2005). Theory of mind in dreaming: Awareness of Hobson, J. A. & Friston, K. J. (2012). Waking and feelings and thoughts of others in dreams. Dreaming, dreaming consciousness: Neurobiological and functional 15 (1), 48-57. 10.1037/1053-0797.15.1.48 considerations. Progress in Neurobiology, 98 (1), 82-98. Karni, A., Tanne, D., Rubenstein, B. S., Askenasy, J. J. 10.1016/j.pneurobio.2012.05.003 & Sagi, D. (1994). Dependence on REM sleep of Hobson, J. A. & Pace-Schott, E. F. (2002). The cognitive overnight improvement of a perceptual skill. Science, neuroscience of sleep: Neuronal systems, consciousness 265 (5172), 679-682. 10.1126/science.8036518 and learning. Nature Reviews Neuroscience, 3 (9), 679- Kris, E. (1952). Psychoanalytic explorations in art. New 693. 10.1038/nrn915 York, NY: International Universities Press. Hobson, J. A. & Wohl, H. (2005). From angels to neur- Lewin, I. (1989). The effect of ‘waking-dreaming‘ on creativity ons. Art and the new science of dreaming. Fidenza, I: and rote memory. Cognition and Perception, 9 (3), 225-236. Mattioli. Lima, S. L., Rattenborg, N. C., Lesku, J. A. & Amlaner, J. C. Holzinger, B. (2014). Lucid dreaming in Psychotherapy. (2005). Sleeping under the risk of predation. Animal Beha- In R. Hurd & K. Bulkeley (Eds.) Lucid Dreaming: New viour, 70 (4), 723-736. 10.1016/j.anbehav.2005.01.008 Perspectives on Consciousness in Sleep (pp. 37-62). Llewellyn, S. (2013). Such stuff as dreams are made on? Horne, J. A. & McGrath, M. J. (1984). The consolidation Elaborative encoding, the ancient art of memory, and hypothesis for REM sleep function: Stress and other the hippocampus. Behavioral and Brain Sciences, 36 confounding factors--a review. Biological Psychology, (6), 589-607. 10.1017/S0140525X12003135 18 (3), 165-184. 10.1016/0301-0511(84)90001-2 Louie, K. & Wilson, M. A. (2001). Temporally structured Hu, P., Stylos-Allan, M. & Walker, M. P. (2006). Sleep replay of awake hippocampal ensemble activity during facilitates consolidation of emotional declarative rapid eye movement sleep. Neuron, 29 (1), 145-156. memory. Psychological Science, 17 (10), 891-898. 10.1016/S0896-6273(01)00186-6 10.1111/j.1467-9280.2006.01799.x Lyon, B. E. & Montgomerie, R. (2012). Sexual selection is a Indursky, P. & Rotenberg, V. (1998). Change of mood form of social selection. Philosophical Transactions of the during sleep and REM sleep variables. International Royal Society of London. Series B, Biological Sciences, Journal of Psychiatry in Clinical Practice, 2 (1), 47-51. 367 (1600), 2266-2273. 10.1098/rstb.2012.0012 10.3109/13651509809115114 Mahner, M. & Bunge, M. (2000). Function and function- Jouvet, M. (1979). What does a cat dream about? Trends alism: A synthetic perspective. Philosophy of Science, in Neurosciences, 2, 280-282. 68 (1), 75-94. 10.1016/0166-2236(79)90110-3 Malcolm-Smith, S., Solms, M., Turnbull, O. & Tredoux, (1998). Paradoxical sleep as a programming sys- C. (2008). Threat in dreams: An adaptation? Con- tem. Journal of Sleep Research, 7 (Suppl 1), 1-5. sciousness and Cognition, 17 (4), 1281-1291. 10.1046/j.1365-2869.7.s1.1.x 10.1016/j.concog.2007.07.002 Kahn, D., Stickgold, R., Pace-Schott, E. F. & Hobson, J. Malcolm-Smith, S., Koopowitz, S., Pantelis, E. & Solms, M. A. (2000). Dreaming and waking consciousness: A char- (2012). Approach/avoidance in dreams. Consciousness and acter recognition study. Journal of Sleep Research, 9 Cognition, 21 (1), 408-412. 10.1016/j.concog.2011.11.004 (4), 317-325. 10.1046/j.1365-2869.2000.00213.x Malinowski, J. E. & Horton, C. L. (2014). Memory Kahn, D., Combs, A. & Krippner, S. (2002a). Dreaming sources of dreams: The incorporation of autobiograph- as a function of chaos-like stochastic processes in the ical rather than episodic experiences. Journal of Sleep self-organizing brain. Nonlinear Dynamics, Psychology, Research, 23 (4), 441-447. 10.1111/jsr.12134 and Life Sciences, 6 (4), 311-322. Maquet, P., Laureys, S., Peigneux, P., Fuchs, S., Petiau, C., 10.1023/A:1019758527338 Phillips, C., Aerts, J., Del Fiore, G., Degueldre, C., Meule- Kahn, D., Pace-Schott, E. & Hobson, J. A. (2002b). Emo- mans, T., Luxen, A., Franck, G., van der Linden, M., tion and cognition: Feeling and character identification Smith, C. & Cleeremans, A. (2000). Experience-dependent in dreaming. Consciousness and Cognition, 11 (1), 34- changes in cerebral activation during human REM sleep. 50. 10.1006/ccog.2001.0537 Nature Neuroscience, 3 (8), 831-836. 10.1038/77744

Marks, G. A., Shaffery, J. P., Oksenberg, A., Speciale, S. Nir, Y. & Tononi, G. (2010). Dreaming and the brain: G. & Roffwarg, H. P. (1995). A functional role for From phenomenology to neurophysiology. Trends in REM sleep in brain maturation. Behavioural Brain Re- Cognitive Sciences, 14 (2), 88-100. search, 69 (1-2), 1-11. 10.1016/0166-4328(95)00018-O 10.1016/j.tics.2009.12.001 Martindale, C. (1999). Biological bases of creativity. Nishida, M., Pearsall, J., Buckner, R. L. & Walker, M. P. Handbook of Creativity (pp. 137-152). Cambridge, UK: (2009). REM sleep, prefrontal theta, and the consolida- Cambridge University Press. tion of human emotional memory. Cerebral Cortex, 19 McCarley, R. W. & Hoffman, E. (1981). REM sleep (5), 1158-1166. dreams and the activation-synthesis hypothesis. Amer- Peigneux, P., Laureys, S., Fuchs, S., Destrebecqz, A., Col- ican Journal of Psychiatry, 138 (7), 904-912. lette, F., Delbeuck, X., Phillips, C., Aerts, J., Del, F. Mednick, S. A. (1962). The associative basis of the creat- iore, Degueldre, C., Luxen, A., Cleeremans, A. & ive process. Psychological Review, 69 (3), 220-232. Maquet, P. (2003). Learned material content and ac- 10.1037/h0048850 quisition level modulate cerebral reactivation during Mellou, E. (1994). Play theories: A contemporary review. posttraining rapid-eye-movements sleep. NeuroImage, Early Child Development and Care, 102 (1), 91-100. 20 (1), 125-134. 10.1080/0300443941020107 Pellegrini, A. D. & Bjorklund, D. F. (2004). The onto- Mendelsohn, G. A. (1976). Associative and attentional geny and phylogeny of children’s object and fantasy processes in creative performance. Journal of Per- play. Human Nature, 15 (1), 23-43. sonality, 44 (2), 341-369. 10.1007/s12110-004-1002-z 10.1111/j.1467-6494.1976.tb00127.x Pennartz, C. M., Lee, E., Verheul, J., Lipa, P., Barnes, C. Metzinger, T. (2003). Being no one: The self-model the- A. & McNaughton, B. L. (2004). The ventral striatum ory of subjectivity. Cambridge, MA: MIT Press. in off-line processing: Ensemble reactivation during (2013). The myth of cognitive agency: Subpersonal sleep and modulation by hippocampal ripples. Journal thinking as a cyclically recurring loss of mental of Neuroscience, 24 (29), 6446-6456. 10.1523/JNEUR- autonomy. Frontiers in Psychology, 4, 931-931. OSCI.0575-04.2004 10.3389/fpsyg.2013.00931 Perogamvros, L. & Schwartz, S. (2012). The roles of the Millikan, R. (1984). Language, thought and other biolo- reward system in sleep and dreaming. Neuroscience & gical categories. Cambridge, MA: MIT Press. Biobehavioral Reviews, 36 (8), 1934-1951. Mirmiran, M. (1995). The function of fetal/neonatal 10.1016/j.neubiorev.2012.05.010 rapid eye movement sleep. Behavioral Brain Research, (2014). Sleep and emotional functions. Current 69 (1-2), 13-22. 10.1016/0166-4328(95)00019-P Topics in Behavorial Neurosciences. Morselli, L. L., Guyon, A. & Spiegel, K. (2012). Sleep and 10.1007/7854_2013_271 metabolic function. Pflügers Archiv, 463 (1), 139-160. Plihal, W. & Born, J. (1997). Effects of early and late 10.1007/s00424-011-1053-z nocturnal sleep on declarative and procedural memory. Neander, K. (1991). Functions as selected effects. Philo- Journal of Cognitive Neuroscience, 9 (4), 534-547. sophy of Science, 58 (2), 168-184. 10.1162/jocn.1997.9.4.534 Nielsen, T. A. (2000). A review of mentation in REM and Rak, M., Beitinger, P. A., Steiger, A., Schredl, M. & NREM sleep: “covert” REM sleep as a possible recon- Dresler, M. (in press). Increased lucid dreaming fre- ciliation of two opposing models. Behavioral and Brain quency in narcolepsy. Sleep. Sciences, 23 (6), 851-866. Rasch, B. & Born, J. (2013). About sleep’s role in Nielsen, T. A., Kuiken, D., Alain, G., Stenstrom, P. & memory. Physiological Review, 93 (2), 681-766. Powell, R. A. (2004). Immediate and delayed incorpor- 10.1152/physrev.00032.2012 ations of events into dreams: Further replication and Rasch, B., Büchel, C., Gais, S. & Born, J. (2007). Odor implications for dream function. Journal of Sleep Re- cues during slow-wave sleep prompt declarative search, 13 (4), 327-336. memory consolidation. Science, 315 (5817), 1426-1429. Nielsen, T. A. & Powell, R. A. (1989). The ‘dream-lag’ ef- 10.1126/science.1138581 fect: A 6-day temporal delay in dream content incor- Revonsuo, A. (1995). Consciousness, dreams and virtual poration. Psychiatry Journal of the University of Ot- realities. Philosophical Psychology, 8 (1), 35-58. towa, 14 (4), 561-565. 10.1080/09515089508573144

(2000). The reinterpretation of dreams: An evolu- Smith, C. T., Nixon, M. R. & Nader, R. S. (2004). tionary hypothesis of the function of dreaming. Behavi- Posttraining increases in REM sleep intensity implicate oral and Brain Sciences, 23 (6), 877-901. REM sleep in memory processing and provide a biolo- (2006). Inner presence. Boston, MA: MIT Press. gical marker of learning potential. Learning & Memory, Revonsuo, A., Tuominen, J. & Valli, K. (2015). The 11 (6), 714-719. 10.1101/lm.74904 avatars in the machine: Dreaming as a simulation of Snyder, F. (1970). The phenomenology of dreaming. The social reality. In T. Metzinger & J. M. Windt (Eds.) Psychodynamic Implications of The Physiological Stud- Open MIND. Frankfurt a. M., GER: MIND Group. ies on Dreams (pp. 124-151). Springfield, IL: Charles Ritter, S. M. & Dijksterhuis, A. (2014). Creativity-the C. Thomas. unconscious foundations of the incubation period. Spoormaker, V. I., van den Bout, J. & Meijer, E. J. G. Frontiers in Human Neuroscience, 8, 215-215. (2003). Lucid dreaming treatment for nightmares: A 10.3389/fnhum.2014.00215 series of cases. Dreaming, 13 (3), 181-186. Ritter, S. M., Strick, M., Bos, M. W., van Baaren, R. B. 10.1023/A:1025325529560 & Dijksterhuis, A. (2012). Good morning creativity: Spoormaker, V. I. & van den Bout, J. (2006). Lucid Task reactivation during sleep enhances beneficial ef- dreaming treatment for nightmares: A pilot study. Psy- fect of sleep on creative performance. Journal of Sleep chotherapy and Psychosomatics, 75 (6), 389-394. Research, 21 (6), 643-647. 10.1159/000095446 10.1111/j.1365-2869.2012.01006.x Stumbrys, T. & Daniels, M. (2010). An exploratory study Roffwarg, H. P., Muzio, J. N. & Dement, W. C. (1966). of creative problem solving in lucid dreams: Prelimin- Ontogenetic development of the human sleep-dream ary findings and methodological considerations. Inter- cycle. Science, 152 (3722), 604-619. national Journal of Dream Research, 3 (2), 121-129. 10.1126/science.152.3722.604 10.11588/ijodr.2010.2.6167 Schredl, M. & Erlacher, D. (2011). Frequency of lucid Stumbrys, T., Erlacher, D., Johnson, M. & Schredl, M. dreaming in a representative German sample. Percep- (2014). The phenomenology of lucid dreaming: An on- tual and Motor Skills, 112 (1), 104-108. line survey. American Journal of Psychology, 127 (2), 10.2466/09.PMS.112.1.104-108 191-204. Schredl, M., Hoffmann, L., Sommer, J. U. & Stuck, B. A. Stumbrys, T., Erlacher, D. & Schredl, M. (2015). Effect- (2014). Olfactory stimulation during sleep can reactiv- iveness of motor practice in lucid dreams: A compar- ate odor-associated images. Chemosensory Perception. ison with physical and mental practice. Journal of 10.1007/s12078-014-9173-4 Sports Sciences. 10.1080/02640414.2015.1030342 Schredl, M. & Hofmann, F. (2003). Continuity between Tholey, P. (1989). Consciousness and abilities of dream waking activities and dream activities. Consciousness characters observed during lucid dreaming. Perceptual and Cognition, 12 (2), 298-308. and Motor Skills, 68 (2), 567-578. 10.1016/S1053-8100(02)00072-7 10.2466/pms.1989.68.2.567 Schuster, C., Hilfiker, R., Amft, O., Scheidhauer, A., An- Tononi, G. & Cirelli, C. (2006). Sleep function and syn- drews, B., Butler, J., Kischka, U. & Ettlin, T. (2011). aptic homeostasis. Sleep Medicine Review, 10 (1), 49- Best practice for motor imagery: A systematic literat- 62. ure review on motor imagery training elements in five Valli, K. & Revonsuo, A. (2009). The threat simulation different disciplines. BMC Med, 9 (75). theory in light of recent empirical evidence: A review. 10.1186/1741-7015-9-75 American Journal of Psychology, 122 (1), 17-38. Schädlich, M. & Erlacher, D. (2012). Applications of van der Werf, Y. D., Altena, E., Schoonheim, M. M., lucid dreams: An online study. International Journal Sanz-Arigita, E. J., Vis, J. C., de Rijke, W. & van of Dream Res, 5 (2), 134-134. Someren, E. J. (2009). Sleep benefits subsequent hip- 10.11588/ijodr.2012.2.9505 pocampal functioning. Nature Neuroscience, 12 (2), Sikka, P., Valli, K., Virta, T. & Revonsuo, A. (2014). I 122-123. 10.1038/nn.2253 know how you felt last night, or do I? Self- and ex- Vassalli, A. & Dijk, D. J. (2009). Sleep function: Current ternal ratings of emotions in REM sleep dreams. Con- questions and new approaches. European Journal of sciousness and Cognition, 25, 51-66. Neuroscience, 29 (9), 1830-1841. 10.1016/j.concog.2014.01.011 10.1111/j.1460-9568.2009.06767.x

Voss, U., Holzmann, R., Tuin, I. & Hobson, J. A. (2009). Windt, J. M. (2010). The immersive spatiotemporal hal- Lucid dreaming: A state of consciousness with features lucination model of dreaming. Phenomenology and the of both waking and non-lucid dreaming. Sleep, 32 (9), Cognitive Sciences, 9, 295-316. 1191-1200. 10.1007/s11097-010-9163-1 Voss, U., Holzmann, R., Hobson, A., Paulus, W., Koppe- Windt, J. M., Harkness, D. L. & Lenggenhager, B. hele-Gossel, J., Klimke, A. & Nitsche, M. A. (2014). In- (2014). Tickle me, I think I might be dreaming! Sens- duction of self awareness in dreams through frontal low ory attenuation, self-other distinction, and predictive current stimulation of gamma activity. Nature Neuros- processing in lucid dreams. Frontiers in Human Neur- cience, 17 (6), 810-812. 10.1038/nn.3719 oscience, 8, 717-717. 10.3389/fnhum.2014.00717 Wagner, U., Gais, S. & Born, J. (2001). Emotional Wolpin, M., Marston, A., Randolph, C. & Clothier, A. memory formation is enhanced across sleep intervals (1992). Individual difference correlates of reported lu- with high amounts of rapid eye movement sleep. cid dreaming frequency and control. Journal of Mental Learning and Memory, 8 (2), 112-119. Imagery, 16, 231-236. 10.1101/lm.36801 Wouters, A. G. (2003). Four notions of biological func- Wagner, U., Gais, S., Haider, H., Verleger, R. & Born, J. tion. Studies in History and Philosophy of Biology and (2004). Sleep inspires insight. Nature, 427 (6972), 352– Biomedical Sciences, 34 (4), 633-668. 355-352–355. 10.1038/nature02223 10.1016/j.shpsc.2003.09.006 Wagner, U., Hallschmid, M., Rasch, B. & Born, J. (2006). (2013). Function, Biological. In W. Dubitzki, O. Brief sleep after learning keeps emotional memories Wolkenhauer, H. Yokota & K.-H. Cho (Eds.) Encyclo- alive for years. Biological Psychiatry, 60 (7), 788-790. pedia of Systems Biology. Berlin, GER: Springer. 10.1016/j.biopsych.2006.03.061 Xie, L., Kang, H., Xu, Q., Chen, M. J., Liao, Y., Thiy- Walker, M. P. & van der Helm, E. (2009). Overnight agarajan, M., O’Donnell, J., Christensen, D. J., Nich- therapy? The role of sleep in emotional brain pro- olson, C., Iliff, J. J., Takano, T., Deane, R. & Neder- cessing. Psychological Bulletin, 135 (5), 731-748. gaard, M. (2013). Sleep drives metabolite clearance 10.1037/a0016570 from the adult brain. Science, 342 (6156), 373-377. Walsh, D. M. & Ariew, A. (1996). A taxonomy of func- 10.1126/science.1241224 tions. Canadian Journal of Philosophy, 26 (4), 493-514. Zadra, A., Desjardins, S. & Marcotte, E. (2006). Evolu- Wamsley, E. J. (2014). Dreaming and offline memory con- tionary function of dreams: A test of the threat simula- solidation. Current Neurology and Neuroscience Re- tion theory in recurrent dreams. Consciousness and ports, 14, 433-433. 10.1007/s11910-013-0433-5 Cognition, 15 (2), 450-463. Wamsley, E. J., Perry, K., Djonlagic, I., Reaven, L. B. & Stickgold, R. (2010a). Cognitive replay of visuomotor learning at sleep onset: Temporal dynamics and relationship to task performance. Sleep, 33 (1), 59-68. Wamsley, E. J., Tucker, M., Payne, J. D., Benavides, J. A. & Stickgold, R. (2010b). Dreaming of a learning task is associated with enhanced sleep-dependent memory consolidation. Current Biology, 20 (9), 850- 855. 10.1016/j.cub.2010.03.027 Wamsley, E. J. & Stickgold, R. (2011). Memory, sleep and dreaming: Experiencing consolidation. Sleep Medi- cine Clinics, 6 (1), 97-108. West-Eberhard, M. J. (2014). Darwin’s forgotten idea: The social essence of sexual selection. Neuroscience and Biobehavioral Reviews 46, 501-508. 10.1016/j.neubiorev.2014.06.015 Wilson, M. A. & McNaughton, B. L. (1994). Reactivation of hippocampal ensemble memories during sleep. Sci- ence, 265 (5172), 676-679. 10.1126/science.8036517