The neural substrate for dreaming: Is it a subsystem of the default network?
G. William Domhoff
University of California, Santa Cruz
NOTE: If you use this paper in research, please use the following citation, as this on-line version is simply a reprint of the original article:
Domhoff, G. W. (2011). The neural substrate for dreaming: Is it a subsystem of the default network? Consciousness and Cognition, 20, 1163-1174.
Building on the content, developmental, and neurological evidence that there are numerous parallels between waking cognition and dreaming, this article argues that the likely neural substrate that supports dreaming, which was discovered through converging lesion and neuroimaging studies, may be a subsystem of the waking default network, which is active during mind wandering, daydreaming, and simulation. Support for this hypothesis would strengthen the case for a more general neurocognitive theory of dreaming that starts with established findings and concepts derived from studies of waking cognition and neurocognition. If this theory is correct, then dreaming may be the quintessential cognitive simulation because it is often highly complex, often includes a vivid sensory environment, unfolds over a duration of a few minutes to a half hour, and is usually experienced as real while it is happening.
A wide range of content, developmental, and neurological findings related to dreaming suggest many parallels with waking cognition (e.g., Antrobus, 1978; Domhoff, 1996; Foulkes, 1985; Nir & Tononi, 2010). This perspective contrasts with the emphasis in most twentieth-century theories on the seemingly large differences between dreaming and waking cognition, with the focus on the primary process and repressed wishes in Freudian theory (Freud, 1900) and the expression of archetypal symbols lodged within an inherited collective unconscious in Jungian theory (Jung, 1974). In the activation-synthesis theory developed by Hobson (e.g. 1988, 1997, 2009), dreaming is a form of delirium (an organic brain disease characterized by disorientation, illogical cognition, distracted attention, unstable emotion, and dull intellectual functions) in which the cortex tries to make sense of random firings from the pontine tegmentum during REM sleep.
However, the slowly accumulating content, developmental, and neurological findings of the past fifty-five years lead to the possibility that a new neurocognitive theory of dreams can be developed that draws on concepts and findings based on studies of waking cognition and neurocognition (Domhoff, 2001, 2003, chap. 1). In this article I build on this claim to suggest that a tentative neural substrate for dreaming, uncovered by converging evidence from lesion and neuroimaging studies (e.g., Braun et al., 1997; Dang-Vu, Schabus, Desseilles, Schwartz, & Maquet, 2007; Solms, 1997), may be based in a subsystem of the waking default network, which is active when the mind is wandering, daydreaming, or simulating past or future events (e.g., Buckner, Andrews-Hanna, & Schacter, 2008; Schacter, Addis, & Buckner, 2008; Szpunar, 2010). It is also active at sleep onset (e.g., Horovitz et al., 2009; Kaufmann et al., 2006; Larson-Prior et al., 2009; Laufs, Walker, & Lund, 2007) and partially active during REM (Pace-Schott, 2010).
The possibility that the default network might be involved in dreaming has been raised by earlier authors (e.g., Fosse & Domhoff, 2007; Ioannides, Kostopoulos, Liu, & Fenwick, 2009; Nir & Tononi, 2010), with Pace-Schott (2007, pp. 137-141; 2010, pp. 565-566) providing the most detailed comparison of the neural substrates that underlie the default network and REM, a comparison in which REM serves as a proxy for the neural substrate for dreaming. However, a case for the overlap between these two neural networks never has been made based on evidence from the literature on dreaming and dream content. In addition, there is recent neuroimaging research on mind wandering that may make the equation more compelling from the neurological side (e.g., Andrews-Hanna, Reidler, Huang, & Buckner, 2010; Andrews-Hanna, Reidler, Sepulcre, Poulin, & Buckner, 2010; Christoff, Gordon, Smallwood, Smith, & Schooler, 2009; Mason et al., 2007).
The article begins with an overview of the several different types of parallels that have been found between dream content and waking thought. These parallels are important to establish because they serve to both guide and constrain the search for a neural substrate for dreaming by demonstrating that this substrate has to include enough interacting regions of the brain to account for the significant number of elaborate, detailed, and deeply thoughtful dreams that reflect waking personal concerns about the past, present, or future. It next turns to the studies of neurological patients with dreaming deficits or excesses, which suggest the brain regions that are necessary and not necessary for dreaming. It then shows that neuroimaging studies of differential activation of different regions of the brain during various phases of sleep are reasonably consistent with the inferences concerning the neural substrate for dreaming based on the lesion studies. Evidence is also presented that some of the regions of the brain that are as active during REM dreaming as they are during mind wandering also may be active at times during NREM sleep, with a special emphasis for purposes of this article on Stage 2 of NREM late in the sleep period (e.g., Cavallero, Cicogna, Natale, & Occhionero, 1992; Cicogna, Natale, Occhionero, & Bosinelli, 1998; Foulkes, 1962, 1999; Wamsley et al., 2007).
Finally, the article turns to a consideration of the burgeoning literature on mind wandering, simulation, and future episodic thought (e.g., Buckner et al., 2008; Gusnard & Raichle, 2001; Smallwood & Schooler, 2006). Recent studies of mind wandering that include concurrent neuroimaging results support the hypothesis that the neural substrate for dreaming shares some of the same neural regions that underlie the default network, including the medial prefrontal cortex, the anterior cingulate cortex, and the temporoparietel junction (Andrews-Hanna, Reidler, Huang et al., 2010; Andrews-Hanna, Reidler, Sepulcre et al., 2010; Christoff, et al., 2009; Mason et al., 2007).
Dreaming and waking cognition
The possibility that dreaming may have much in common with waking cognition first arose with the finding that the dream reports from awakenings in sleep laboratories in the 1950s and 1960s were less unusual and seemingly "symbolic" than researchers had assumed they would be. This general finding is best demonstrated by a study that brought together 635 dream reports collected from 58 young adult men and women in two different sleep laboratories in a series of investigations over a period of seven years between 1960 and 1967 (Snyder, 1970, p. 127; Snyder, Karacan, Tharp, & Scott, 1968). The researchers describe a prototypical REM dream report as a "clear, coherent, and detailed account of a realistic situation involving the dreamer and other people caught up in very ordinary activities and preoccupations, and usually talking about them" (Snyder, 1970, p. 148). Overall, they conclude that as many as "90% would have been considered credible descriptions of everyday experience" (Snyder et al., 1968, p. 375), and that "dreaming consciousness" is therefore "a remarkably faithful replica of waking life" (Snyder, 1970, p. 133).
The general expectation that dreams would have a large amount of unusual and bizarre content led another team of investigators to undertake a detailed study of 16 young adult women, each of whom spent two consecutive nights in the lab and answered questions about the familiarity and likelihood of specific dream elements after an average of four REM awakenings per night (Dorus, Dorus, & Rechtschaffen, 1971). Each of the various types of elements that appear in dreams, such as settings, characters, activities, and social interactions, was placed in one of six categories for types of "novelty," ranging from exact replication of reality to the fantastic or improbable. Their conclusions "emphasize the rarity of the bizarre in dreams" because major distortions of actual waking experiences reach a high of only 16.7% of all the activities and social interactions, and of 6.2% and 7.8% for all characters and physical surroundings; when they carried out global ratings of each dream for overall novelty, they found that 25.8% showed large but plausible differences from previous waking experiences and that only 8.9% were highly improbable by waking standards (Dorus et al., 1971, p. 367).
The relative lack of emotion in dream reports from REM awakenings, especially strong emotion, also came as a surprise to early lab investigators, leading to several studies confirming that it did occur less often than expected. In the most comprehensive of these studies, 17 young adults (9 women, 8 men) were quizzed in detail as to the presence of emotions and the appropriateness of the emotion to the content after an average of six REM awakenings per person over two nonconsecutive nights. Based on ratings by both participants and na´ve judges, it was concluded that about 70% of the dream reports had at least some affect (Foulkes, Sullivan, Kerr, & Brown, 1988), although most of it not very strong, a finding that was supported in later studies (Fosse, Stickgold, & Hobson, 2001; Strauch & Meier, 1996). The type of emotion, or lack thereof, was appropriate to the dream situation in 60% of the dreams, but unexpectedly there was no emotion in 17% of the cases where there would have been some in a similar waking situation, as well as a few instances (3.2%) where emotion was present when there would have been none in waking life (Foulkes et al., 1988).
It is also striking that speech in dreams is as complex and nuanced as it is in waking. Sentences uttered in dream reports are as grammatically correct as in waking life, and they are appropriate to the situation in the dream (Meier, 1993). Then, too, thinking of a goal-directed nature is demonstrated by the use of such terms as "contemplate," "decide," "realize," "ponder," and "think about:" 13.8% of the 500 men's dreams and 21.0% of the 500 women's dreams in the Hall and Van de Castle (1966) normative sample of classroom-collected dream reports contain at least one such thinking element. There is further evidence for the presence of wake-like thinking within dreams in another study of dreams collected in the classroom setting; it concluded that "cognition within a dream scenario was similar to that of wake-state cognition," although it also went onto say that "thinking about the scenario itself was deficient and very different than wake-state thinking" (Kahn & Hobson, 2005, p. 437). In still other studies, dreamers sometimes recognize that something is not right when they are in unusual or impossible situations through the expression of confusion or doubt, or else they report other forms of very sophisticated thinking, such as puzzlement (Domhoff, 2007; Kozmova & Wolman, 2006; Wolman & Kozmova, 2007).
Conversely, studies of relaxed waking thought in laboratory settings reveal that it can be as fragmented or unusual as dreams. In a laboratory study in which participants reclined in a moderately lighted room with their wakefulness monitored by EEG and EMG recordings, Foulkes and Scott (1973) discovered that 24% of sampled thoughts from 16 female college students were described as visual, dramatic, and dream-like. In a replication study using 10 men and 10 women, who were signaled to report their thoughts at 12 random times in sessions of 45 to 60 min, Foulkes and Fleisher (1975, p. 69) reported a very similar figure for dream-like episodes, 19%, along with the fact that another 20% contained "what might be called mindwandering (the subject is not controlling his thoughts, but he is aware he is in the laboratory, and his mentation is nonhallucinatory)" and yet another 22% contained "what might be called 'lost in thought' (the subject may or may not be controlling his thoughts, but he is not aware he is in the laboratory, and his mentation is nonhallucinatory)." A study comparing REM reports to waking streams of thought from the same participants sitting in a darkened room found that there were more abrupt topic changes in the waking sample than in their REM reports (Reinsel, Antrobus, &Wollman, 1992). All of these older findings from drifting waking thought are consistent with the possibility that a subsystem of the default network, which is highly active during resting states, is the basis for dreaming (e.g., Andrews-Hanna, Reidler, Huang et al., 2010; Christoff et al., 2009).
In field studies in which participants were paged at random times during the day to report their thoughts, about one-third of all thoughts were judged by participants as "spontaneous," meaning that they just popped into their minds (Klinger, 1999). Demonstrating that spontaneous thoughts can be highly unusual, 21% of the reports analyzed in these studies have aspects that are physically impossible, and many thoughts were judged as disconnected. Importantly, there were also wide individual differences in how much thinking is said to be deliberate or spontaneous. For two-thirds of the participants, the majority of their thoughts were deliberate and intentional, but for the other one-third the majority of their thoughts were spontaneous (Klinger, 1999, 2009). Based on these overall findings, it seems likely that the best basis for eventually ascertaining the relative bizarreness of waking thought and dreaming would be comparisons of dream reports and waking thought samples from the same participants.
Studies of dream content suggest that a majority of dreams are based on concerns about the past, present, or future; these concerns are often dramatized as highly negative worst-case scenarios. For example, two-fifths of men and women's dream reports in the Hall and Van de Castle (1966) normative study contain hostile thoughts or aggressive actions, and both men and women are the victims in three-fifths or more of these aggressive interactions; similarly one-third for the dreams of both genders contain some type of misfortune, defined as negative events not caused by others. More generally, 73.6% of the men's dream reports and 68.2% of the women's dream reports contained at least one aggression, misfortune, or failure, compared to 51.2% for the men and 48.7% for the women that contained at least one friendly interaction, good fortune, or success (Schneider & Domhoff, 1995, drawing on a spreadsheet based on the original Hall/Van de Castle normative codings). In a similar fashion, studies of daydreams find that they revolve around current concerns, regrets, unfinished intentions, and hopes for the future (e.g., Gold & Reilly, 1985/1986; Klinger, 1990). However, it also needs to be added that daydreams are generally more pleasant, more satisfactory in outcome, and more transparent in meaning than dreams are, and people usually feel more in control of them as well (Klinger, 1990, pp. 16, 64).
Developmental findings on dreaming
Large and detailed longitudinal and cross-sectional studies of the dream reports of dozens of children between 3 and 16, who were awakened several thousand times by the same investigator, and also given an extensive battery of personality and cognitive tests outside the laboratory setting by other members of the research team, support the characterization of dreaming and dream content overviewed in the previous section. They reveal that the ability to produce a coherent and animated dream narrative parallels the development of waking cognitive capabilities, with the ability to generate mental imagery as the most important correlate (Foulkes, 1982, 1999; Foulkes, Hollifield, Sullivan, Bradley, & Terry, 1990). To begin with, children between 3 and 5 reported dreams from only 27% of REM awakenings, and for the most part the content of these dreams was static, bland, and underdeveloped. Dream reports collected from children in the sleep laboratory became more "dreamlike" in the 5-to-7 year-olds in terms of characters, themes, and the ability to introduce action and their own selves into the process, but it was not until the children were 11 to 13 that their dreams began to resemble those of adult laboratory participants in frequency, length, emotions, and overall structure, or to show any relationship to personality (Foulkes, 1982, p. 217). The most important conclusion that can be drawn from these developmental studies is that dreaming is a gradual cognitive achievement that relates to the development of waking cognitive abilities.
Brain lesions, waking thought, and dreaming
The final parallel that can be drawn between dreaming and waking cognition is based on studies of brain lesions. Such studies began with several case histories in the neurological literature going back to the 1880s, which document that lesions to specific areas in the medial occipito-temporal region have similar impacts on the ability to generate visual imagery in both waking and dreaming, such as complete loss of visual imagery, the inability to visualize motion, the loss of facial imagery, or the loss of color vision (Solms, 1997, chaps. 2-3). However, these studies did not take on theoretical importance until more detailed analyses of imagery-impaired patients in the sleep laboratory verified the lack of any visual imagery in both waking and dreaming in the case of a patient with Turner's Syndrome (Kerr, Foulkes, & Jurkovic, 1978) and for a patient with static visual imagery in both waking and dreaming (Kerr & Foulkes, 1981). These parallels between visual imagery impairments in dreaming and waking are important to the argument presented in this article because they provide the most specific example to date for the idea that the neural substrate for dreaming may share regions with the neural substrate that subserves waking mental imagery, mind wandering, and simulation. At least for the time being, most of the remaining lesion evidence to be presented is less localized.
The findings from the handful of case studies concerning impairments in visual imagery in dreaming were replicated and extended in neuropsychological work in the late 1980s by Solms (1997) in which he questioned 361 newly hospitalized neurological patients about any possible changes in the frequency and nature of their dreaming since their injury or illness. The findings with questionnaires were compared with findings from neurological tests and CAT scans, with a primary emphasis on patients with focal brain lesions so that hypotheses about the role of specific regions of the brain in dreaming could be formulated.
Based on his comparison of questionnaire and neurological results, as well as follow-up discussions and observations with some patients, Solms concluded that there are two different types of dreaming "deficits" -- loss of visual dreaming, which is related to lesions in the same regions identified in earlier case studies, and complete loss ("cessation") of dreaming, which is associated with either lesions on or near the parietal-temporal-occipital junction or with bifrontal lesions in the ventromesial region (Solms, 1997, chaps. 4 and 16). The results concerning the effects of lesions in the parietal-temporaloccipital region have been refined by recent studies, which suggest that global cessation of dreaming can occur with tempero- occipital lesions, with no parietal involvement (Bischof & Basset, 2004; Poza & Masso, 2006; Yu, 2006). As for the findings on the loss of dreaming with bifrontal lesions in the ventromesial region, they are noteworthy because the lesions are in almost exactly the same region that was surgically severed from the frontal cortex in thousands of leucotomy patients in the United States, Canada, and Europe from the 1930s to the 1950s, a majority of whom reported a loss of dreaming as well as a lack of spark and imagination in waking life (Solms, 1997, chap. 5). In addition, a REM awakening study comparing leucotomized and non-leuctomized schizophrenic patients in a Montreal hospital supports the earlier leucotomized patients' self-reports of loss of dreaming (Jus et al., 1973).
Solms (1997) also found that there are at least two types of dreaming "excesses," the first of which, characterized by increased frequency and vividness of dreaming, along with the intrusion of dreaming into waking thought, is associated with injuries in the medial prefrontal cortex, anterior cingulate cortex, and basal forebrain. Moreover, some of these patients said that their waking thoughts quickly turned into pictures or realistic events; in several instances, observations by members of the hospital staff supported the notion that these patients suffered from a confusion between dreaming and waking thought (Solms, 1997, pp. 198-199). Whitty and Lewin (1957) reported several similar cases, and Damasio, Graff-Radford, Eslinger, Damasio, and Kassell (1985, p. 269) wrote that their patients with similar lesions were suffering from "waking dreams." I think these findings are of great significance because, as will be shown, the medial prefrontal cortex and anterior cingulate cortex are more active in REM than NREM and during mind wandering than during directed thinking. However, the differences from case to case and the vagueness of the descriptions of neural damage suggest that these lesion findings are not yet well enough localized.
The second dreaming excess, an increased frequency of nightmares, was linked to lesions in the temporal lobe, with 5 of the 9 patients with this syndrome showing symptoms of epilepsy. These patients sometimes suffered from daytime hallucinations as well (Solms, 2000, p. 847 for a summary and references on epilepsy and nightmares). In addition, studies that use stereotaxic electrodes to locate the sites causing seizures in epileptics reveal that the "dreamy state" sometimes experienced as part of the diagnostic process is related to the temporal-limbic region. In one such study, the amygdala, anterior hippocampus, and temporal cortex were involved in every spontaneous occurrence of this state during the procedure (Bancaud, Brunet-Bourgin, Chauvel, & Halgren, 1994). Thus, the possibility arises that the seizures may be activating the neural substrate for dreaming (Solms, 1997, p. 243). These findings take on greater importance when the results of imaging studies of the default network are discussed in a later section.
There is one final finding from Solm's (1997) study that is relevant here: 200 of the 332 patients with brain lesions reported no changes in dreaming, which is potentially useful information because it may reveal those regions in the brain that are not necessary for dreaming. They turn out to be regions that are essential to sensation, locomotion, and higher-order executive functions: the sensorimotor cortices and the dorsolateral prefrontal cortex. When these discoveries concerning the brain regions that may not be necessary for dreaming are combined with the regions that create four different types of changes in dreaming, all of which have parallels in waking cognitive defects, the result is a relatively circumscribed hypothetical neural substrate for dreaming that can be compared to the results of neuroimaging studies of sleep on the one hand and to studies of the waking default network on the other. It is to these issues that the article turns in the next two sections, but with the proviso that any neuroimaging findings concerning sleep states eventually have to be linked to reports of dreaming, just as neuroimagining studies of waking cognition have to be linked to behavioral measures and/or reports on subjective experience.
Neuroimaging studies during sleep
The contours of the neural substrate for dreaming that can be inferred from the lesion studies have been reinforced by several neuroimaging studies of the sleeping brain. Most of these studies concentrate their attention on the REM stage because early sleep lab studies showed that awakenings from it lead to dream reports 80% to 90% of the time with adult participants (e.g. Dement & Kleitman, 1957; Kamiya, 1961; Snyder, 1970). However, it is an imperfect proxy for the putative neural substrate for dreaming because there is evidence that dreaming also can occur in NREM stages of sleep, especially late in the sleep period, a point that is discussed more fully a little later in this section (e.g., Antrobus, Kondo, & Reinsel, 1995; Cicogna et al., 1998; Fosse, Stickgold, & Hobson, 2004; Nielsen, 2000; Occhionero, 2004). Just as Christoff et al. (2009) correctly predicted, based on behavioral studies, that part of the executive network is recruited during mind wandering, I predict on the basis of the dreamlike properties of many NREM reports late in the sleep period, and especially those in a spontaneous morning awakening study by Cicogna et al. (1998), that the neural substrate that provides the platform for dreaming is active in Stage 2 late in the sleep period.
Several early neuroimaging studies using PET scans discovered that the medial prefrontal cortex, anterior cingulate cortex, limbic region, and basal forebrain are more active during REM than NREM, and in some cases as active as they are in presleep or early-morning waking comparisons (e.g., Braun et al., 1997; Maquet et al., 1996; Nofzinger, Mintun, Wiseman, Kupfer, & Moore, 1997). It is noteworthy for my purposes that all of these regions were identified as part of the neural substrate for dreaming by the lesion studies discussed in the previous section. In addition, the neuroimaging results are consistent with the inclusion of the temporal lobe in the neural substrate for dreaming because they show that the occipital-temporal region is more active during REM than NREM (Braun et al., 1997; Maquet et al., 1996) Then, too, the visual association cortex -- and the auditory association cortex as well -- are active during REM (Braun et al., 1998; Maquet et al., 2000). Finally, the pontine tegmentum is more active in REM than NREM, with one study suggesting it is even more active in REM than in waking (Braun et al., 1997). This pontine activation appears to spread to the thalamus through cholinergic pathways, and then to the basal ganglia, basal forebrain, and limbic/paralimbic regions mentioned earlier in this paragraph. Pace-Schott (2007, 2010) and Dang-Vu et al. (2007) provide detailed overviews of these and related findings.
Conversely, the early PET studies that looked at all stages of sleep suggested that in comparison with pre-sleep or earlymorning waking records, the dorsolateral prefrontal cortex, sensorimotor cortex, and motor cortex are less active during both REM and NREM, along with the opercular cortex and posterior cingulate cortex, which fits with evidence that lesions in these areas do not affect dreaming (Braun et al., 1997, 1998; Maquet et al., 1996, 1997).
However, later studies suggest some slight alterations in this general picture, which underscores the fact that the issue of what regions are most active in REM is not settled. Maquet et al. (2005) found that some areas within the dorsolateral prefrontal cortex are more active in REM than NREM, contrary to earlier neuroimaging studies. Similarly, Ioannides et al. (2009), using magnetoencephalography (MEG) with four participants, each of whom had a full night's sleep in the laboratory, also discovered higher activity in regions within the dorsolateral prefrontal cortex than in NREM. These more recent findings related to the higher activity in at least some regions of the dorsolateral prefrontal cortex become relevant in the later discussion of the default network because one fMRI study demonstrates that regions in the dorsolateral prefrontal cortex are more active during mind wandering than in directed thinking, as is a region that is always active in REM, the dorsal medial prefrontal cortex (Christoff et al., 2009).
Different investigators highlight slightly different brain regions in discussing their findings. Maquet et al., 2000, p. 222) conjectures that the amygdala may be the central structure in the modulation of cortical activity in REM, as shown by the fact that it is closely connected to the anterior cingulate cortex and the inferior parietal lobule, but has few connections to the dorsolateral prefrontal cortex and parietal lobes. Nofzinger et al., 2001, 1997 emphasize the potential importance of the anterior cingulate cortex, which plays a role in attentional states, performance monitoring, and error detection in waking thought. Ioannides et al. (2009) suggest that the dorsal medial prefrontal cortex may be the central point from which further activation spreads. Braun et al. (1997, p. 1190) provide the most general statement, one that is consistent with the argument advanced in this article: "REM sleep may constitute a state of generalized brain activity with the specific exclusion of executive systems that normally participate in the highest order analysis and integration of neural information."
In terms of a cognitive approach to locating the neural substrate for dreaming, the important point is that these slightly different emphases nonetheless support my hypothesis that specific regions within the association cortices, along with some paralimbic and limbic structures, such as the amygdala, underlie the process of dreaming. They seem to form an operative subsystem that is cut off from both the primary sensory cortices and the regions within the prefrontal cortices that integrate incoming sensory information with memory and emotion in the process of decision-making (c.f. Braun et al., 1998, p. 94). This neural subsystem seems to contain enough cognitive processing areas, such as the medial frontal cortex and anterior cingulate cortex, to produce the complexity of dreaming that is required by the many findings on dream content discussed earlier in this article. It thereby qualifies as a plausible "neural substrate for dreaming."
However, as already stated, the outlines of the neural substrate for dreaming cannot be asserted with confidence until they can be shown to subserve the very similar dreaming that sometimes occurs in NREM, especially Stage 2, which is most frequent later in the sleep period (.i.e., in the early mornings for everyone except those who work at night). In addition, there is dreamlike mental activity, albeit of a more fragmented, disjointed, and fleeting nature, for several seconds to a minute or more at sleep onset, which might provide some clues as to the substrate that enables imagistic thinking in dreaming (Cicogna et al., 1998; Fosse & Domhoff, 2007; Vogel, 1991). In fact, the sudden switch to dreamlike thinking, whether at sleep onset, during the drowsiness of a slow morning awakening, or in brief episodes of relaxed waking thought, suggest that the transition to dreaming can be very rapid (Bertini, Lewis, & Witkin, 1964; Foulkes & Fleisher, 1975; Foulkes & Vogel, 1965). Such quick transitions are consistent with the possibility that dreaming is generated by a subsystem of the default network because there is evidence that the default network is still active at sleep onset and into Stage 2 of NREM before Stages 3 and 4 occur (Horovitz et al., 2009; Laufs, 2008, for a review and critique of several studies).
The fact of dreaming in NREM was first established in detail in work by Kamiya (1961) and Foulkes (1962), but it tended to be explained away at first by REM-oriented investigators (including the author of this article) because some of it was more "thought-like," or was conjectured to be based on memories from dreams during earlier REM periods (Domhoff, 2004). Nevertheless, work in pioneer sleep researcher Alan Rechtschaffen's laboratory (Foulkes & Rechtschaffen, 1964; Rechtschaffen, Verdone, & Wheaton, 1963) demonstrated otherwise, as did a study carried out in two different sleep labs, which included careful controls for the expectations and biases of both the investigators and the participants (Herman, Ellman, & Roffwarg, 1978). Herman et al.'s (1978) study showed that raising the expectancy of dream recall in either the experimenters or the participants could increase the frequency of dream reports from NREM awakenings, but the findings with a fourth experimental group revealed a large amount of NREM dreaming that could not be accounted for by expectancy biases. The investigators concluded that their results were "in line with those investigators who found substantial amounts of NREM mentation and a relatively small REM-NREM discriminability" (Herman et al., 1978, p. 90).
Although one later study found evidence for at least some degree of dreaming in all stages of sleep (Cavallero et al., 1992), most NREM studies since the 1980s have concentrated on awakenings from NREM late in the sleep period, when sleepers are most likely to be in Stage 2. They conclude that the reports from Stage 2 are more similar to REM reports on a variety of rating scales (Antrobus et al., 1995; Fosse et al., 2004). This point is also made in an internal pilot study at the Neurophysiology Laboratory at Harvard Medical School: "none of the ten expert judges was able to correctly sort a mixed sample of twenty early-night REM and late-night NREM dreams above a chance level, using any criteria they wanted" (Fosse & Domhoff, 2007, p. 61). Two different studies of dream content using separate representative samples, which were drawn randomly from the same dream reports used by Fosse et al. (2004), concluded that the REM and late-night NREM reports were similar for most of the Hall and Van de Castle (1966) content categories that were employed, with the exception of categories relating to aggression (McNamara, McLaren, & Durso, 2007; McNamara, McLaren, Smith, Brown, & Stickgold, 2005).
Most important on this issue, a unique large-scale laboratory study analyzed 72 dream reports from spontaneous morning awakenings for 36 young adults (20 female, 16 male), who spent at least two and sometimes more nights in the laboratory so they could each contribute two reports on the few occasions when there was no waking recall on one of the first two nights (Cicogna et al., 1998). Seventy-four% of the spontaneous awakenings were from NREM, usually in Stage 2, and 36% from REM, which is consistent with earlier studies of the percentage of morning awakenings from REM and NREM; recall rates were 95% from REM awakenings and 91% from Stage 2 awakenings (Cicogna et al., 1998, p. 466). Significantly, there were virtually no differences between the NREM and REM reports on a variety of rating scales that were applied by blind coders to a single randomized portfolio that included both sets of dream reports. The one difference, not expected or readily understandable to the investigators, revealed the NREM reports to be more "bizarre" in terms of "spatio-temporal units," such as "the fusing of different places," or "impossible or incongruous spaces and times," even though the two sets of dream reports were similar on a global measure of bizarreness (Cicogna et al., 1998, p. 467). Based on this study and the several other studies showing few or no differences between dream content from REM and NREM awakenings, it should be possible to find indications that the neural substrate for dreaming is active in whole or part during Stage 2 late in the sleep period.
The neural substrate for dreaming in stage 2
Three studies have discussed the possible neural basis for dreaming in Stage 2 (Fosse & Domhoff, 2007; Ioannides et al., 2009; Wamsley et al., 2007). According to theorizing by Fosse and Domhoff (2007, p. 60), the greater activation in the midbrain reticular formation and thalamus in Stage 2 may "set the stage" for dreaming. However, it seems likely that more than the activation of these two basic components of arousal systems would be necessary to provide a platform for complex mental activity. It is therefore significant that the anterior regions of the thalamus are linked to limbic regions that show heightened activity in REM, which include the hippocampus and posterior hypothalamus.Widening this potential neural substrate for dreaming during Stage 2 one step further, findings with intracranial stereotaxic electrodes in epileptic patients suggest that the amygdala and orbitofrontal cortex may be active enough in Stage 2 to sustain dreaming (Fosse & Domhoff, 2007, p. 60).
Ioannides et al. (2009, p. 455) discovered evidence that some of the same regions that are active in REM are also differentially active in Stage 2, especially the dorsal medial prefrontal cortex and the precuneus. They speculate that the dorsal medial prefrontal cortex may be the "geographic center" for dreaming in both REM and Stage 2 late in the sleep period, an idea that may gain in importance based on the recent studies of the default network that are discussed in the next section (Ioannides et al., 2009, p. 465). Their findings also have parallels in the studies of the default network in Stage 2 shortly after sleep onset (e.g., Horovitz et al., 2009; Larson-Prior et al., 2009; Laufs, 2008).
The possibility that the brain in general slowly returns to waking levels of activation late in the sleep period, as indexed by rising brain temperature from a low point in the middle of the night, led Wamsley et al. (2007) to argue for a "dual rhythm model." In this model the gradual activation in the second half of the sleep period interacts with the activation in Stage 2 and REM to make dreaming more likely towards morning whatever the sleep stage. They provide evidence for this point by showing that both Stage 2 and REM reports became longer, more dreamlike, and more bizarre when compared to reports from the likely low point for brain activation in their 20 participants (five male, 15 female). They theorize that the differences in global and regional brain activation between REM and Stage 2 remain steady over the night, but they also note that "no neuroimaging studies have examined the global or regional activation in REM and NREM sleep during the late morning" (Wamsley et al., 2007, pp. 352-353).
While the evidence I have presented to support the idea that the neural substrate for dreaming may be active in Stage 2 late in the sleep period is suggestive, there is obviously a need for further neuroimaging studies that focus in detail on Stage 2 episodes from which dreams are subsequently reported. It may be that studies using fMRI, MEG and/or high-density electroencephalography (hd-EEG) can provide the temporal and spatial resolution necessary, as Nir and Tononi (2010) suggest, especially if used in conjunction with spontaneous NREM morning awakenings similar to those studied by Cicogna et al. (1998).
Dreams and the default network
Recent research on mind wandering, daydreaming, and simulation provides a new avenue for those dream researchers who begin with the premise that systematic studies of waking cognitive processes and their neural correlates are the best starting point for developing a neurocognitive theory of dreaming and dream content. Although there are a large number of studies that provide indirect evidence that links mind wandering to the set of neural subsystems called the default network (e.g., Buckner et al., 2008; Gusnard & Raichle, 2001; Smallwood & Schooler, 2006, for summaries), I will focus on the handful of recent studies that included neuroimaging scans at the same time that behavioral or subjective evidence of mind wandering was being collected.
Cognitively, both the putative neural substrate for dreaming and the default network subserve stimulus-independent, non-directed thinking (e.g., Andrews-Hanna, Reidler, Huang et al., 2010; Foulkes, 1985). Both of them often make use of "simulation," defined as the "imaginative construction of hypothetical events or sequences;" in addition, simulations are usually generated "with a view toward addressing a current or future problem" (Schacter et al., 2008, p. 42; Smallwood, Nind, & O'Connor, 2009). Schacter et al. (2008, p. 42) then write about simulations in the default network in a way that could just as easily be applied to dreams: "We view simulation as a particular kind or subset of thinking that involves imaginatively placing oneself in a hypothetical scenario and exploring possible outcomes."
Although there is often simulation when the mind is wandering, thinking during this state also can shift suddenly and include unusual topics, leading Smallwood and Schooler (2006) to speak of "the restless mind" in their review of systematic research on mind wandering. Slipping into the default network leads to inattention and impairments in executive system abilities that seem to me to have important parallels with dreaming. For example, Smallwood and Schooler (2006, p. 956) describe mind wandering as "a state of decoupled information processing, which occurs because of a shift of attention from the immediate environment," a statement that might be made about dreaming as well. Put another way, both mind wandering and dreaming involve thinking that is turned inward to personal concerns and sometimes jumps around.
As suggested in the Introduction, the neural substrate for dreaming and the default network seem to share many brain regions in common. Although circumstantial evidence for this point can be found in many places, the best sources for my purposes are recent fMRI studies that include on-line tasks, simulation instructions, and/or subjective probes. Mason et al. (2007, p. 393) explicitly set out to determine whether the default network is implicated in mind wandering by comparing brain activity when participants were performing either highly practiced -- and thus easily carried out -- workingmemory tasks, or else doing new working-memory tasks of the same complexity, which require more focus and showed less evidence of mind wandering in earlier, non-neuroimaging studies. As predicted, they found that several regions in the default network exhibited greater activation during highly practiced tasks that allow for mind wandering, including the medial prefrontal cortex, the anterior cingulate cortex, the posterior cingulate cortex, and the precuneus, among several (Mason et al., 2007, p. 394). They then propose that mind wandering is "a psychological baseline that emerges when the brain is otherwise unoccupied, supported by activity in a default network of cortical regions" (Mason et al., 2007, p. 394). This conclusion seems to fit dreaming equally well because the brain is "otherwise unoccupied" at the times when it becomes more active during sleep.
Christoff et al. (2009) took a further step in linking mind wandering directly to a brain state by collecting mind-wandering reports as well as determining performance errors on a test measuring sustained attention to a response task while the participants' brains were being scanned using fMRI. For my purposes here, the crucial finding is that the most prominent regions of the default network, such as the medial prefrontal cortex, ventral anterior cingulate cortex, precuneus and temporoparietal area, were most active during mind wandering, along with "the two main executive regions" in the brain, the dorsal anterior cingulate cortex and the dorsolateral prefrontal cortex (Christoff et al., 2009, p. 8720). This is a combination of neural regions anchored in the default network that seems ideal for dreaming, although it remains to be determined just which regions within the dorsolateral prefrontal cortex might be recruited by the default network during dreaming. Similarly, Spreng, Stevens, Chamberlain, Gilmore, and Schacter (2010) demonstrate that the default network and the frontoparietal control network are both active when people think about personal planning concerning the future, while externally directed attention networks remain relatively inactive.
It is also noteworthy in terms of any attempt to establish parallels between mind wandering and dreaming that neural recruitment in the Christoff et al. (2009, p. 8719) study "was strongest when subjects were unaware of their own mind wandering, suggesting that mind wandering is most pronounced when it lacks meta-awareness." A lack of "meta-awareness" is reminiscent of the "single-mindedness" of dreams, with dreamers rarely aware that they are dreaming (Rechtschaffen, 1978, 1997). The parallels between the lack of meta-awareness during dreaming and the failure to encode external events during mind wandering may provide an opening to a cognitive explanation, in terms of a lack of focused attention when the mind is involved in simulation, for why both dreams and drifting waking thoughts are usually soon forgotten (Chapman & Underwood, 2000; Foulkes, 1999; Smallwood, Baracaia, Lowe, & Obsonsawin, 2003).
The work of Andrews-Hanna, Reidler, Sepulcre et al., 2010 revealing subsystems within the default network provides an important bridge for arguing that the neural substrate for dreaming is a subsystem of the default network, or at least the center of a subsystem that also includes areas within the dorsolateral prefrontal cortex (Maquet et al., 2005). Their findings are based on three separate fMRI experiments involving 129 participants (82 female, 47 male) who took part in one research step or another. Using correlational, graph-analytic, and clustering analyses of fluctuations in oxygen blood flows, they first discovered that the default network consists of two distinct subsystems. The two subsystems are connection by two "hubs," or central connecting points, the dorsal medial prefrontal cortex and the posterior cingulate cortex. The first subsystem, which they call the "dorsal medial prefrontal cortex system," includes the dorsal medial prefrontal cortex, the temperoparietal junction, the lateral temporal cortex, and the temporal pole of the temporal lobe. The second subsystem, called the medial temporal lobe system, includes the ventral medial prefrontal cortex, posterior inferior parietal lobule, retrosplenial cortex, parahippocampal cortex, and hippocampal formation (Andrews-Hanna, Reidler, Sepulcre et al., 2010, p. 559).
Using instructions to imagine oneself in a present or future situation, Andrews-Hanna, Reidler, Sepulcre et al. (2010, p. 554) next demonstrated that "the subsystems and core act as functionally coherent units during task performance, exhibiting both correlation and independence as predicted by the analyses in experiment one." More specifically, the dorsal medial prefrontal cortex system is activated by instructions to think about the person's present situation or present mental state ("present self"), whereas the medial temporal lobe system is called into action by thinking about personal situations and decisions in the future ("future self"). The authors then note "both subsystems are activated during passive states, when participants engage in spontaneous cognition" (Andrews-Hanna, Reidler, Sepulcre et al., 2010, p. 559).
In a related set of studies Andrews-Hanna and a slightly different group of colleagues (Andrews-Hanna, Reidler, Huang et al., 2010) took the further step of exploring the substance of this spontaneous cognition. Through a variety of approaches that included probes while 199 young adult participants (97 men, 102 women) were doing different types of response tasks, and later filling out a post-fMRI questionnaire, they demonstrate that most of the thoughts were self-relevant and emotionally engaging when spontaneous cognition occurred during periods of heightened activity in the default network. Most of these thoughts concerned the recent past and the near future, and they "correlate with functional interactions between the medial temporal lobe [which has a role in autobiographical recall and imagination] and cortical regions within the default network" (Andrews-Hanna, Reidler, Huang et al., 2010, p. 330).
Taken together, the two publications by Andrews-Hanna and her colleagues lead to the conclusion that "when left alone undisturbed, people tend to engage in self-relevant internal cognition processes predominantly about significant past and future events" (Andrews-Hanna, Reidler, Sepulcre et al., 2010, p. 559). This conclusion is similar to the emphasis on dreams as the expression of conceptions and personal concerns in the cognitive theory of dreaming (Domhoff, 2003; Foulkes, 1985; Hall, 1953).
Based primarily on earlier studies of the neural correlates of mind wandering, several previous researchers have commented on the parallels between the default network and the neural substrate for dreaming, but they are not in complete agreement on the usefulness of the comparison (Fosse & Domhoff, 2007; Ioannides, Kostopoulos, Liu, and Fenwick, 2009; Nir & Tononi, 2010; Pace-Schott, 2007, 2010). Fosse and Domhoff's theorizing (2007, p. 68), which draws on an early synthesis of various studies by Gusnard and Raichle (2001), notes that the default network "has characteristic brain activation patterns that suggest it lies between alert executive waking cognition and REM sleep," which fits well with Christoff et al., 2009 findings revealing that any "antagonisms" between externally oriented executive networks and internally oriented default networks during directed thinking are overridden during mind wandering. Fosse and Domhoff (2007, p. 68) note that the "activation of mediofrontal and medioparietal regions that belong to the executive cognitive system, in addition to ventromedial prefrontal regions and the temporoparietal junction on the right side of the brain," is "consistent with the notion of the baseline state [default network] as a condition of reduced cognitive control combined with enhanced thought flow and imagery".
As stated in the Introduction, Pace-Schott (2007, 2010) provides the most detailed comparison of the two brain networks to date. He stresses that it is the anterior regions of the default network, such as the anterior cingulate cortex and areas in the orbitofrontal cortex, that are more active in REM than NREM, whereas the posterior aspects of the default network are less active. He is clearly of the mind that the most active areas of the default network might support dreaming, although he also discusses several problems that need to be resolved before the similarities can be pushed too far (e.g., Pace-Schott, 2007, p. 140). In particular, he points to the lower levels of activity in the posterior cingulate cortex and other posterior areas of the default network, as found in work by Nofzinger and his colleagues (Nofzinger et al., 1997, 2004). As Pace-Schott (2010, p. 566) puts it, "reactivation of default network structures during REM, as measured by PET, is only partial," but he then concludes, in an analysis that anticipates the one in this article, that "dreaming might simulate lifelike events in a manner unconstrained by past realities."
Ioannides et al. (2009) also suggest parallels between the neural substrate for dreaming and the default network. Based on their work and an assessment of the default literature, they believe that the dorsal medial prefrontal cortex and the dorsolateral prefrontal cortex, along with the precuneus, are more active in REM, Stage 2, and the waking default network than in the deeper stages of NREM. Nir and Tononi (2010), in their state-of-the-art synthesis of work on the relationship between waking consciousness and dreaming, draw on several studies in pointing out that there are similarities between the default network and the neural substrate for dreaming. However, the lower levels of activation in the posterior cingulate cortex and other posterior regions during REM, along with questions at the time they wrote about whether the thinking emanating from the default network is self-referential enough to support dreaming, or is simply aimless mind wandering, led them to conclude that the default network may not be a key factor in dreaming (Nir & Tononi, 2010, p. 96). Since their analysis was completed, however, the two studies by Andrews-Hanna, Reidler, Huang et al., 2010; Andrews-Hanna, Reidler, Sepulcre et al., 2010 have shown that the default network does include considerable self-reflection. But that still leaves the fact that some posterior parts of the default network are inactive during REM.
The low levels of activation in the posterior cingulate cortex compared to various anterior default regions was not viewed as a major problem by Pace-Schott (2007) or Fosse and Domhoff (2007) because that structure is generally found to be involved in monitoring events in the environment, especially potential threats. Furthermore, as Pace-Schott (2007, p. 139) and Schacter et al. (2008, p. 44) point out, the posterior cingulate cortex also seems to be involved in retrieving past episodic memories, which may not be a central issue as far as dreaming if studies showing little episodic memory in dreams are correct (Baylor & Cavallero, 2001; Fosse, Hobson, & Stickgold, 2003). Nor did Ioannides et al. (2009) see the relative inactivity of the posterior cingulate cortex as a major stumbling block to looking at subsystems of the default network as the most likely brain basis for dreaming.
Above and beyond these arguments that the posterior cingulate cortex might not be essential to dreaming, there is a possible methodological issue that needs to be resolved. Given that this brain region is very active during mind wandering, it may be that a more relevant comparison of its level of activation during dreaming would be with an active, externally driven task during waking, when it is less active. Put another way, the posterior cingulate cortex may be somewhat more active during dreaming as compared to alert waking than REM neuroimaging studies have shown due to their use of relaxed waking states just before falling asleep or upon awakening in the morning as their baseline of comparison.
Based on these several different arguments and considerations concerning the relevance and relative activation of various posterior areas during dreaming, it seems premature to reject the possibility that a subsystem of the default network may be the neural substrate for dreaming. The results of the Andrews-Hanna, Reidler, Sepulcre et al., 2010 study also can be recalled as evidence for this point because they demonstrated that the default network can indeed be "fractionated" into subsystems. This raises the possibility that the neural substrate for dreaming is a slightly different subsystem than either of the two they discovered, one that can function without the posterior areas because there is no need for environmental monitoring or for episodic memories. Moreover, as already mentioned in this article, this subsystem may recruit areas within the dorsolateral prefrontal cortex.
The several different analyses and findings presented in the preceding paragraphs therefore raise the possibility that future fMRI, high-density EEG, and/or MEG studies of Stage 2 at the end of the sleep period can provide the information necessary to determine the degree of overlap between the neural substrate for dreaming and the default network. This might be especially the case if some studies included dream reports from spontaneous awakenings from Stage 2, thereby making it possible to replicate the Cicogna et al. (1998) findings with greater neurological specificity than standard EEG techniques provide. Moreover, if waking neuroimaging studies can more exactly pinpoint the regions in the default network that lead to highly imaginative simulations, which might be possible because waking participants can respond readily to initiating cues and follow-up questions, then perhaps the neural substrate for dreaming can be further specified by seeing if those same specific regions are more active during REM and Stage 2 than in Stages 3 and 4 of NREM.
If this strategy proved fruitful, then it might be useful to build on the several studies of the transition to sleep and the first hour of sleep, which show that the default network remains at least somewhat active during the transition to sleep and in Stage 2 on the way to Stages 3 and 4 (e.g., Horovitz et al., 2009; Kaufmann et al., 2006; Larson-Prior et al., 2009). Although these studies did not involve awakenings, the activity they found in the default network fits with the findings of dreamlike reports at sleep onset and early in the sleep period in earlier sleep laboratory studies before anything was known about the default network (e.g., Cicogna et al., 1998; Foulkes & Vogel, 1965; Vogel, 1991).
If the systematic findings on the parallels between dreaming and waking cognition discussed in the early sections of the article are judged to be convincing, and if the argument linking dreaming to a perhaps-unique subsystem of the same neural substrate that enables mind wandering seems to be plausible, then dreams can be seen as a unique and more fully developed form of mind wandering, and therefore as the quintessential cognitive simulation. They are the quintessential cognitive simulation not only because they have elaborate story lines that are often enacted with exquisite sensory, motor, and cognitive involvement, with some dreams unfolding over a period of several minutes to half an hour or more. There is also the striking fact that they are usually experienced as real while they are happening.
If the step-by-step argument in this article were to be supported with systematic empirical evidence, it would reinforce the general assumption that dreaming and dream content must be understood in terms of concepts developed through rigorous studies of waking cognition. At the cognitive level, the understanding of dreams would begin with the idea that they are based on many of the same conceptions and concerns that shape thinking and behavior in waking life (e.g., Hall, 1953; Klinger, 1971, 2009). At the neurocognitive level, dreaming would be understood as the product of a subsystem of the brain's default network, perhaps augmented by the recruitment of parts of the dorsolateral prefrontal cortex. It is likely that this subsystem is operative whenever there is (1) an intact and fully mature neural substrate for dreaming, a qualification that allows for the impact of lesions on the functioning of this substrate and for the lack of dreaming in young children; (2) an adequate level of cortical activation, which can be provided by the REM mechanism and/or generally higher brain activation at sleep onset and late in the sleep period; (3) an occlusion of external stimuli, most likely through gates in the thalamus; and (4) the loss of conscious self-control, i.e., a shutting down of the prefrontal executive systems that connect us to the external world by integrating the massive amounts of external and internal information they are constantly receiving. Within this context, the new challenge would be to determine what changes within the brain cause drifting waking thoughts to turn into dreamlike scenarios. It is on this issue that the dreamlike mentation at sleep onset -- and during the drifting between waking consciousness and dreaming after a relaxed morning awakening -- may be of more than passing interest.
If the neural substrate for dreaming is a subsystem of the default network, there also may be implications for the hitherto intractable debate over the possible adaptive function(s) of dreaming, all of which have been rejected as implausible and contrary to the evidence by cognitive dream researchers, who see dreaming as a by-product of adaptive waking cognitive processes, such as the ability to generate visuospatial imagery and develop an autobiographical self (Antrobus, 1993; Blagrove, 2000; Domhoff, 2003; Flanagan, 1995; Foulkes, 1993). If the default network could be added to the list of waking cognitive systems that have adaptive value due to the new associations and ideas it provides via mind wandering and daydreaming, then it might be argued that dreaming may have similar functions as a residual by-product of the activation of a subsystem of the default network during sleep. At the same time, however, it should be stressed that it is by no means established that mind wandering has adaptive value. As Mason et al. (2007, p. 395) emphasize: "Although the thoughts the mind produces when wandering are at times useful, such instances do not prove that the mind wanders because these thoughts are adaptive; on the contrary the mind may wander simply because it can."
So, too, the mind may dream simply because it can. Dreams contain a considerable degree of psychological meaning in terms of the coherency of most individual dreams, the consistency of dream content over months, years, and decades, and the correspondences of dream content with waking psychological variables (e.g., Domhoff, 2007; Zadra & Domhoff, 2011). They also have been put to use by people in many different times and places as important parts of religious and healing ceremonies, which means that they have an emergent cultural function due to human inventiveness. However, they may or may not have any adaptive value as evolutionary theorists use the term. They may simply be dramatic simulations of our conceptions, concerns, and interests that occur when a specific constellation of neural regions is activated in a context where there is no engagement with the external world. If that proves to be the case, then psychological meaning and cultural usefulness have to be distinguished from each other and from the issue of adaptive function in order to develop an adequate theory of dreams.
My thanks to Jessica Andrews-Hanna, Roar Fosse, Yuval Nir, Edward Pace-Schott, Jerome M. Siegel, Jonathan Smallwood, and Mark Solms for their very helpful suggestions in reaction to earlier drafts of this paper.
Andrews-Hanna, J., Reidler, J., Huang, C., & Buckner, R. (2010). Evidence for the default network's role in spontaneous cognition. Journal of Neurophysiology, 104, 322-335.
Andrews-Hanna, J., Reidler, J., Sepulcre, J., Poulin, R., & Buckner, R. (2010). Functional-anatomic fractionation of the brain's default network. Neuron, 65, 550-562.
Antrobus, J. (1978). Dreaming as cognition. In A. Arkin, J. Antrobus, & S. Ellman (Eds.), The mind in sleep: Psychology and psychophysiology (1st ed., pp. 569-581). Hillsdale, NJ: Erlbaum.
Antrobus, J. (1993). Dreaming: Could we do without it? In A. Moffitt, M. Kramer, & R. Hoffmann (Eds.), The Functions of Dreaming (pp. 549-558). Albany, NY: State University of New York Press.
Antrobus, J., Kondo, T., & Reinsel, R. (1995). Dreaming in the late morning: Summation of REM and diurnal cortical activation. Consciousness & Cognition, 4, 275-299.
Bancaud, J., Brunet-Bourgin, F., Chauvel, P., & Halgren, E. (1994). Anatomical origin of deja vu and vivid 'memories' in human temporal lobe epilepsy. Brain, 117, 71-90.
Baylor, G., & Cavallero, C. (2001). Memory sources associated with REM and NREM dream reports throughout the night: A new look at the data. Sleep, 24, 165-170.
Bertini, M., Lewis, H., & Witkin, H. (1964). Some preliminary observations with an experimental procedure for the study of hypnagogic and related phenomena. Archivo Di Psicologia, Neurologia e Psichiatria, 6, 493-534.
Bischof, M., & Basset, C. (2004). Total dream loss: A distinct neuropsychological dysfunction after bilateral PCA stroke. Annals of Neurology, 56, 583-586.
Blagrove, M. (2000). Dreams have meaning but no function. Behavioral and Brain Sciences, 23, 910.
Braun, A., Balkin, T., Wesensten, N., Carson, R., Varga, M., Baldwin, P., et al (1997). Regional cerebral blood flow throughout the sleep-wake cycle: An (H2O)- O-15 PET study. Brain, 120, 1173-1197.
Braun, A., Balkin, T., Wesensten, N., Gwadry, F., Carson, R., Varga, M., et al (1998). Dissociated pattern of activity in visual cortices and their projections during human rapid eye movement sleep. Science, 279, 91-95.
Buckner, R., Andrews-Hanna, J., & Schacter, D. (2008). The brain's default network: Anatomy, function, and relevance to disease. Annals of the New York Academy of Sciences, 1124, 1-38.
Cavallero, C., Cicogna, P., Natale, V., & Occhionero, M. (1992). Slow wave sleep dreaming. Sleep, 15, 562-566.
Chapman, P., & Underwood, G. (2000). Mental states during dreaming and daydreaming: Some methodological loopholes. Behavioral and Brain Sciences, 23, 917-918.
Christoff, K., Gordon, A., Smallwood, J., Smith, R., & Schooler, J. (2009). Experience sampling during fMRI reveals default network and executive system contributions to mind wandering. Proceedings of the National Academy of Sciences of the United States of America, 106, 8719-8724.
Cicogna, P., Natale, V., Occhionero, M., & Bosinelli, M. (1998). A comparison of mental activity during sleep onset and morning awakening. Sleep, 21(5), 462-470.
Damasio, A., Graff-Radford, N., Eslinger, P., Damasio, H., & Kassell, N. (1985). Amnesia following basal forebrain lesions. Archives of Neurology, 42, 263-271.
Dang-Vu, T., Schabus, M., Desseilles, M., Schwartz, S., & Maquet, P. (2007). Neuroimaging of REM sleep and dreaming. In D. Barrett & P. McNamara (Eds.), The new science of dreaming: Biological aspects (Vol. 1, pp. 95-113). Westport, CT: Praeger/Greenwood.
Dement, W., & Kleitman, N. (1957). The relation of eye movements during sleep to dream activity: An objective method for the study of dreaming. Journal of Experimental Psychology, 53, 339-346.
Domhoff, G. W. (1996). Finding meaning in dreams: A quantitative approach. New York: Plenum.
Domhoff, G. W. (2001). A new neurocognitive theory of dreams. Dreaming, 11, 13-33.
Domhoff, G. W. (2003). The scientific study of dreams: Neural networks, cognitive development, and content analysis. Washington, DC: American Psychological Association.
Domhoff, G. W. (2004). Why did empirical dream researchers reject Freud? A critique of historical claims by Mark Solms. Dreaming, 14, 3-17.
Domhoff, G. W. (2007). Realistic simulation and bizarreness in dream content: Past findings and suggestions for future research. In D. Barrett & P. McNamara (Eds.), The new science of dreaming: Content, recall, and personality correlates (Vol. 2, pp. 1-27). Westport, CT: Praeger.
Dorus, E., Dorus, W., & Rechtschaffen, A. (1971). The incidence of novelty in dreams. Archives of General Psychiatry, 25, 364-368.
Flanagan, O. (1995). Deconstructing dreams: The spandrels of sleep. Journal of Philosophy, 92, 5-27.
Fosse, R., & Domhoff, G. W. (2007). Dreaming as non-executive orienting: A conceptual framework for consciousness during sleep. In D. Barrett & P. McNamara (Eds.), The new science of dreaming: Content, recall, and personality correlates (Vol. 2, pp. 49-78). Westport, CT: Praeger.
Fosse, R., Hobson, J. A., & Stickgold, R. (2003). Dreaming and episodic memory: A functional dissociation? Journal of Cognitive Neuroscience, 15, 1-9.
Fosse, R., Stickgold, R., & Hobson, J. A. (2001). The mind in REM sleep: Reports of emotional experience. Sleep, 24, 947-955.
Fosse, R., Stickgold, R., & Hobson, J. A. (2004). Thinking and hallucinating: Reciprocal changes in sleep. Psychophysiology, 41, 298-305.
Foulkes, D. (1962). Dream reports from different states of sleep. Journal of Abnormal and Social Psychology, 65, 14-25.
Foulkes, D. (1982). Children's dreams. New York: Wiley.
Foulkes, D. (1985). Dreaming: A cognitive-psychological analysis. Hillsdale, NJ: Erlbaum.
Foulkes, D. (1993). Data constraints on theorizing about dream function. In A. Moffitt, M. Kramer, & R. Hoffmann (Eds.), The functions of dreaming (pp. 11-20). Albany, NY: State University of New York Press.
Foulkes, D. (1999). Children's dreaming and the development of consciousness. Cambridge, MA: Harvard University Press.
Foulkes, D., & Fleisher, S. (1975). Mental activity in relaxed wakefulness. Journal of Abnormal Psychology, 84, 66-75.
Foulkes, D., Hollifield, M., Sullivan, B., Bradley, L., & Terry, R. (1990). REM dreaming and cognitive skills at ages 5-8: A cross-sectional study. International Journal of Behavioral Development, 13, 447-465.
Foulkes, D., & Rechtschaffen, A. (1964). Presleep determinants of dream content: Effects of two films. Perceptual and Motor Skills, 19, 983-1005.
Foulkes, D., & Scott, E. (1973). An above-zero baseline for the incidence of momentarily hallucinatory mentation. Sleep Research, 2, 108.
Foulkes, D., Sullivan, B., Kerr, N., & Brown, L. (1988). Appropriateness of dream feelings to dreamed situations. Cognition and Emotion, 2, 29-39.
Foulkes, D., & Vogel, G. (1965). Mental activity at sleep onset. Journal of Abnormal Psychology, 70, 231-243.
Freud, S. (1900). The interpretation of dreams. London: Oxford University Press (J. Crick, Trans.).
Gold, S., & Reilly, J. (1985/1986). Daydreaming, current concerns, and personality. Imagination, Cognition and Personality, 5, 117-125.
Gusnard, D., & Raichle, M. (2001). Searching for the baseline: Functional imagining and the resting human brain. Nature Reviews Neuroscience, 2, 685-694.
Hall, C. (1953). A cognitive theory of dreams. Journal of General Psychology, 49, 273-282.
Hall, C., & Van de Castle, R. (1966). The content analysis of dreams. New York: Appleton-Century-Crofts.
Herman, J., Ellman, S., & Roffwarg, H. (1978). The problem of NREM dream recall reexamined. In A. Arkin, J. Antrobus, & S. Ellman (Eds.), The mind in sleep: Psychology and psychophysiology (pp. 59-62). Hillsdale, NJ: Erlbaum.
Hobson, J. A. (1988). The dreaming brain. New York: Basic Books.
Hobson, J. A. (1997). Dreaming as delirium: A mental status analysis of our nightly madness. Seminars in Neurology, 17, 121-128.
Hobson, J. A. (2009). REM sleep and dreaming: Towards a theory of protoconsciousness. Nature Reviews Neuroscience, 10, 803-813.
Horovitz, S., Braun, A., Carr, W., Picchioni, D., Balkin, T., Fukunaga, M., et al (2009). Decoupling of the brain's default mode network during deep sleep. Proceedings of the National Academy of Sciences of the United States, 106, 11376-11381.
Ioannides, A., Kostopoulos, K., Liu, L., & Fenwick, P. (2009). MEG identifies dorsal medial brain activations during sleep. NeuroImage, 44, 455-468.
Jung, C. (1974). Dreams. Princeton, NJ: Princeton University Press.
Jus, A., Jus, K., Villenueve, A., Pires, A., Lachance, R., Fortier, J., et al (1973). Studies on dream recall in chronic schizophrenic patients after prefrontal lobotomy. Biological Psychiatry, 6, 275-293.
Kahn, D., & Hobson, J. A. (2005). State-dependent thinking: A comparison of waking and dreaming thought. Consciousness & Cognition, 14, 429-438.
Kamiya, J. (1961). Behavioral, subjective, and physiological aspects of drowsiness and sleep. In D. W. Fiske & S. R. Maddi (Eds.), Functions of varied experience (pp. 145-174). Homewood, IL: Dorsey.
Kaufmann, C., Wehrle, R., Wetter, T. C., Holsboer, F., Auer, D., Pollmacher, T., et al (2006). Brain activation and hypothalamic functional connectivity during human non-rapid eye movement sleep: An EEG/fMRI study. Brain, 129, 655-667.
Kerr, N., & Foulkes, D. (1981). Right hemispheric mediation of dream visualization: A case study. Cortex, 17, 603-610.
Kerr, N., Foulkes, D., & Jurkovic, G. (1978). Reported absence of visual dream imagery in a normally sighted subject with Turner's syndrome. Journal of Mental Imagery, 2, 247-264.
Klinger, E. (1971). Structure and functions of fantasy. New York: Wiley-Interscience.
Klinger, E. (1990). Daydreaming. Los Angeles: Jeremy P. Tarcher.
Klinger, E. (1999). Thought flow: Properties and mechanisms underlying shifts in content. In J. Singer & P. Salovey (Eds.), At play in the fields of consciousness (pp. 29-50). Hillsdale, NJ: Erlbaum.
Klinger, E. (2009). Daydreaming and fantasizing: Thought flow and motivation. In K. Markman, W. Klein, & J. Suhr (Eds.), Handbook of imagination and mental simulation (pp. 225-239). New York: Psychology Press.
Kozmova, M., & Wolman, R. (2006). Self-awareness in dreaming. Dreaming, 16, 196-214.
Larson-Prior, L., Zempel, J., Nolan, T., Prior, F., Snyder, A., & Raichle, M. (2009). Cortical network functional connectivity in the descent to sleep. Proceedings of the National Academy of Sciences of the United States, 106, 4489-4494.
Laufs, H. (2008). Endogenous brain oscillations and related networks detected by surface EEG-combined fMRI. Human Brain Mapping, 29, 762-769.
Laufs, H., Walker, M. C., & Lund, T. (2007). Brain activation and hypothalamic functional connectivity during human non-rapid eye movement sleep: An EEG/fMRI study -- its limitations and an alternative approach. Brain, 130, 75-84.
Maquet, P., Degueldre, C., Delfiore, G., Aerts, J., Peters, J., Luxen, A., et al (1997). Functional neuroanatomy of human slow wave sleep. Journal of Neuroscience, 17(8), 2807-2812.
Maquet, P., Laureys, S., Peigneux, P., Fuchs, S., Petiau, C., Phillips, C., et al (2000). Experience-dependent changes in cerebral activation during human REM sleep. Nature Neuroscience, 3(8), 831-836.
Maquet, P., Peters, J., Aerts, J., Delfiore, G., Dequerldre, C., Luxen, A., et al (1996). Functional neuroanatomy of human rapid-eye-movement sleep and dreaming. Nature, 383, 163-166.
Maquet, P., Ruby, P., Maudoux, A., Albouy, G., Sterpenich, V., Dang-Vu, T., et al (2005). Human cognition during REM sleep and the activity profile within frontal and parietal cortices: A reappraisal of functional neuroimaging data. Progress Brain Research, 150, 219-227.
Mason, M., Norton, M., Van Horn, J., Wenger, D., Grafton, S., & Macrae, N. (2007). Wandering minds: The default network and stimulus-independent thought. Science, 315, 393-395.
McNamara, P., McLaren, D., & Durso, K. (2007). Representation of the self in REM and NREM dreams. Dreaming, 17, 113-126.
McNamara, P., McLaren, D., Smith, D., Brown, A., & Stickgold, R. (2005). A 'Jekyll and Hyde' within: Aggressive versus friendly interactions in REM and NREM dreams. Psychological Science, 16, 130-136.
Meier, B. (1993). Speech and thinking in dreams. In C. Cavallero & D. Foulkes (Eds.), Dreaming as cognition (pp. 58-76). New York: Harvester Wheatsheaf.
Nielsen, T. A. (2000). Covert REM sleep effects on REM mentation: Further methodological considerations and supporting evidence. Behavioral and Brain Sciences, 23, 1040-1057.
Nir, Y., & Tononi, G. (2010). Dreaming and the brain: From phenomenology to neurophysiology. Trends in Cognitive Sciences, 14, 88-100.
Nofzinger, E., Berman, S., Fasiczka, A., Miewald, J., Meltzer, C., Price, J., et al (2001). Effects of bupropion SR on anterior paralimbic function during waking and REM sleep in depression: Preliminary findings using (18F)-FDG PET. Psychiatry Research, 106, 95-111.
Nofzinger, E., Buysse, D., Germain, A., carter, C., Luna, B., Price, J., et al (2004). Increased activation of anterior paralimbic and executive cortex from waking to rapid eye movement sleep in depression. Archives of General Psychiatry, 61, 695-702.
Nofzinger, E., Mintun, M., Wiseman, M., Kupfer, D., & Moore, R. (1997). Forebrain activation in REM sleep: an FDG PET study. Brain Research, 770, 192-201.
Occhionero, M. (2004). Mental processes and the brain during dreams. Dreaming, 14, 54-64.
Pace-Schott, E. (2007). The frontal lobes and dreaming. In D. Barrett & P. McNamara (Eds.), The new science of dreaming: Bioloigical aspects (Vol. 1, pp. 115-154). Westport, CT: Praeger/Greewood.
Pace-Schott, E. (2010). The neurobiology of dreaming. In M. Kryger, T. Roth, & W. Dement (Eds.), Principles and Practices of Sleep Medicine (Fifth ed., pp. 563-575). Philadelpha: Elsevier Saunders.
Poza, I., & Masso, J. (2006). Total dream loss secondary to left temporo-occipital brain injury. Neurologia, 21, 152-154.
Rechtschaffen, A. (1978). The single-mindedness and isolation of dreams. Sleep, 1, 97-109.
Rechtschaffen, A. (1997). Postscript, 1995: The single-mindedness and isolation of dreams. In M. Myslobodsky (Ed.), The mythomanias: The nature of deception and self-deception (pp. 219-223). Mahwah, NJ: Erlbaum.
Rechtschaffen, A., Verdone, P., & Wheaton, J. (1963). Reports of mental activity during sleep. Canadian Psychiatric Association Journal, 8, 409-414.
Reinsel, R., Antrobus, J., & Wollman, M. (1992). Bizarreness in dreams and waking fantasy. In J. Antrobus & M. Bertini (Eds.), The neuropsychology of sleep and dreaming (pp. 157-184). Hillsdale, NJ: Erlbaum.
Schacter, D., Addis, D., & Buckner, R. (2008). Episodic simulation of future events: Concepts, data, and applications. Annals of the New York Academy of Sciences, 1124, 39-60.
Schneider, A., & Domhoff, G. W. (1995). The quantitative study of dreams. http://www.dreamresearch.net/.
Smallwood, J., Baracaia, S., Lowe, M., & Obsonsawin, M. (2003). Task unrelated thought whilst encoding information. Consciousness and Cognition, 12, 452-484.
Smallwood, J., Nind, L., & O'Connor, R. (2009). When is your head at? An exploration of the factors associated with the temporal focus of the wandering mind. Consciousness and Cognition, 18, 118-125.
Smallwood, J., & Schooler, J. (2006). The restless mind. Psychological Bulletin, 132, 946-958.
Snyder, F. (1970). The phenomenology of dreaming. In L. Madow & L. Snow (Eds.), The psychodynamic implications of the physiological studies on dreams (pp. 124-151). Springfield, IL: Thomas.
Snyder, F., Karacan, I., Tharp, V., & Scott, J. (1968). Phenomenology of REM dreaming. Psychophysiology, 4, 375.
Solms, M. (1997). The neuropsychology of dreams: A clinico-anatomical study. Hillsdale, NJ: Erlbaum.
Solms, M. (2000). Dreaming and REM sleep are controlled by different brain mechanisms. Behavioral and Brain Sciences, 23, 843-850.
Spreng, R. N., Stevens, W. D., Chamberlain, J., Gilmore, A., & Schacter, D. (2010). Default network activity, coupled with the frontoparietal control network, supports goal-directed cognition. NeuroImage, 53, 303-317.
Strauch, I., & Meier, B. (1996). In search of dreams: Results of experimental dream research. Albany, NY: State University of New York Press.
Szpunar, K. (2010). Episodic future thought: An emerging concept. Perspectives on Psychological Science, 5, 142-162.
Vogel, G. (1991). Sleep-onset mentation. In S. Ellman & J. Antrobus (Eds.), The mind in sleep: Psychology and psychophysiology (2nd ed., pp. 125-136). New York: Wiley & Sons.
Wamsley, E., Hirota, Y., Tucker, M., Smith, M., Doan, T., & Antrobus, J. (2007). Circadian and ultradian influences on dreaming: A dual rhythm model. Brain Research Bulletin, 71, 347-354.
Whitty, C., & Lewin, W. (1957). Vivid day-dreaming: An unusual form of confusion following anterior cingulectomy. Brain, 80, 72-76.
Wolman, R. N., & Kozmova, M. (2007). Last night I had the strangest dream: Varieties of rational thought processes in dream reports. Consciousness & Cognition, 16, 838-849.
Yu, C. (2006). The brain mechanims of dreaming. Unpublished doctoral dissertation, University of Cape Town, Cape Town, South Africa.
Zadra, A., & Domhoff, G. W. (2011). The content of dreams: Methods and findings. In M. Kryger, T. Roth, & W. Dement (Eds.), Principles and Practices of Sleep Medicine (Fifth ed., pp. 585-594). Philadelphia: Elsevier Saunders.
Go back to the Dream Library index.