Content uploaded by Lampros Perogamvros
Author content
All content in this area was uploaded by Lampros Perogamvros on May 17, 2019
Content may be subject to copyright.
1
BIOLOGICAL SCIENCES, Neuroscience
Fear in dreams and in wakefulness: evidence for day/night affective
homeostasis.
Sterpenich V1,2,5, Perogamvros, L1, 3, 4, 5, 6,*, Tononi G4, Schwartz S1,2,*
1 Department of Neurosciences, Faculty of Medicine, University of Geneva, 9 chemin des
mines, 1202 Geneva, Switzerland
2 Swiss Center for Affective Sciences, University of Geneva, 9 chemin des mines, 1202
Geneva, Switzerland
3 Sleep Laboratory, Division of Pneumology, Geneva University Hospitals, Rue Gabrielle-
Perret-Gentil 4, 1205 Geneva, Switzerland
4 Wisconsin Institute for Sleep and Consciousness, Department of Psychiatry, University of
Wisconsin - Madison, 6001 Research Park Blvd Madison, Wisconsin, USA
5 These authors contributed equally
6 Lead Contact
*Correspondence: Virginie.Sterpenich@unige.ch Lampros.Perogamvros@unige.ch
Keywords: dreaming, wakefulness, sleep, emotion regulation, fear, EEG, fMRI, amygdala,
insula
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
2
ABSTRACT
Recent neuroscientific theories have proposed that emotions experienced in dreams
contribute to the resolution of emotional distress and preparation for future affective
reactions. We addressed one emerging prediction, namely that experiencing fear in dreams
is associated with more adapted responses to threatening signals during wakefulness. Using
a stepwise approach across two studies, we identified brain regions activated when
experiencing fear in dreams and showed that frightening dreams modulated the response of
these same regions to threatening stimuli during wakefulness. Specifically, in Study 1, we
performed serial awakenings in 18 participants recorded throughout the night with high-
density EEG and asked them whether they experienced any fear in their dreams. Insula and
midcingulate cortex activity increased for dreams containing fear. In Study 2, we tested 89
participants and found that those reporting higher incidence of fear in their dreams showed
reduced emotional arousal and fMRI response to fear-eliciting stimuli in the insula, amygdala
and midcingulate cortex, while awake. Consistent with better emotion regulation processes,
the same participants displayed increased medial prefrontal cortex activity. These findings
support that emotions in dreams and wakefulness engage similar neural substrates, and
substantiate a link between emotional processes occurring during sleep and emotional brain
functions during wakefulness.
SIGNIFICANT STATEMENT
Highly debated while pivotal to current theoretical models of dreaming, the relationship
between emotion processing during wakefulness and in dreams remains elusive. In a first
study, we used high-density EEG recordings and observed that regions involved in fear
processing (i.e. the insula and midcingulate cortex) were activated during fear-related
dreams. This first finding demonstrates that emotions in dreams engage similar neural
circuits as during wakefulness. In a second study, using fMRI, we show that higher incidence
of fear in dreams was associated with reduced emotional arousal and brain responses
indicative of better emotion regulation during wakefulness. Together, these results strongly
support the idea that experiencing fear in a secure environment, as in dreams, relates to
more adapted responses to threatening events in real life.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
3
INTRODUCTION
Converging evidence from human and animal research suggests functional links between
sleep and emotional processes (Wagner U et al. 2006; Walker MP and E van der Helm 2009;
Perogamvros L and S Schwartz 2012; Boyce R et al. 2016). Chronic sleep disruption can lead
to increased aggressiveness (Kamphuis J et al. 2012) and negative mood states (Zohar D et
al. 2005), while affective disorders such as depression and PTSD are frequently associated
with sleep abnormalities (e.g., insomnia and nightmares). Experimental evidence indicates
that acute sleep deprivation impairs the prefrontal control over limbic regions during
wakefulness, hence exacerbating emotional responses to negative stimuli (Yoo SS et al.
2007). Neuroimaging and intracranial data further established that, during human sleep,
emotional limbic networks are activated (e.g., Maquet P et al. 1996; Braun AR et al. 1997;
Nofzinger EA et al. 1997; Schabus M et al. 2007; Corsi-Cabrera M et al. 2016). Together these
findings indicate that sleep physiology may offer a permissive condition for affective
information to be reprocessed and reorganized. Yet, it remains unsettled whether such
emotion regulation processes also happen at the subjective, experiential level during sleep,
and may be expressed in dreams. Several influential theoretical models formalized this idea.
For example, the Threat Simulation Theory postulated that dreaming may fulfill a
neurobiological function by allowing an offline simulation of threatening events and
rehearsal of threat-avoidance skills, through the activation of a fear-related amygdalocortical
network (Revonsuo A 2000; Valli K et al. 2005). Such mechanism would promote adapted
behavioral responses in real life situations (Valli K and A Revonsuo 2009). Other models
suggested that dreaming would facilitate the resolution of current emotional conflict
(Cartwright R et al. 1998; Cartwright R et al. 2006), the reduction of next-day negative mood
(Schredl M 2010) and extinction learning (Nielsen T and R Levin 2007). Although these two
main theories differ, because one focuses on the resolution of current emotional distress
(e.g. fear extinction; Nielsen T and R Levin 2007) and the other on the optimization of waking
affective reactions (Revonsuo A 2000; Perogamvros L and S Schwartz 2012), both converge
to suggest that experiencing fear in dreams leads to more adapted responses to threatening
signals during wakefulness (Scarpelli S et al. 2019). The proposed mechanism is that
memories from a person’s affective history are replayed in the virtual and safe environment
of the dream so that they can be reorganized (Nielsen T and R Levin 2007; Perogamvros L
and S Schwartz 2012). From a neuroscience perspective, one key premise of these
theoretical models is that experiencing emotions in dreams implicates the same brain
circuits as in wakefulness (Hobson JA and EF Pace-Schott 2002; Schwartz S 2003). Preliminary
evidence from two anatomical investigations showed that impaired structural integrity of
the left amygdala was associated with reduced emotional intensity in dreams (De Gennaro L
et al. 2011; Blake Y et al. 2019).
Like during wakefulness, people experience a large variety of emotions in their dreams,
with REM dreaming being usually more emotionally-loaded than NREM dreams (Smith MR et
al. 2004; Carr M and T Nielsen 2015). While some studies found a relative predominance of
negative emotions, such as fear and anxiety, in dreams (Merritt J et al. 1994 ; Roussy F et al.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
4
2000), others reported a balance of positive and negative emotions (Schredl M and E Doll
1998), or found that joy and emotions related to approach behaviors may prevail (Fosse R et
al. 2001; Malcolm-Smith S et al. 2012). When performing a lexicostatistical analysis of large
datasets of dream reports, a clear dissociation between dreams containing basic, mostly
fear-related, emotions and those with other more social emotions (e.g. embarrassment,
excitement, frustration) was found, highlighting distinct affective modes operating during
dreaming, with fear in dreams representing a prevalent and biologically-relevant emotional
category (Revonsuo A 2000; Schwartz S 2004). Thus, if fear-containing dreams serve an
emotion regulation function, as hypothesized by the theoretical models, the stronger the
recruitment of fear-responsive brain regions (e.g. amygdala, cingulate cortex, insula; see
Phan KL et al. 2002) during dreaming, the weaker the reaction of these same regions to
actual fear-eliciting stimuli during wakefulness should be. This compensatory or homeostatic
mechanism may also be accompanied by an enhanced recruitment of emotion regulation
brain regions (such as the medial prefrontal cortex, mPFC, which is implicated in fear
extinction) during wakefulness (Quirk GJ et al. 2003; Phelps EA et al. 2004; Yoo SS et al.
2007; Dunsmoor JE et al. 2019).
Here, we collected dream reports and functional brain measures using high-density EEG
(hdEEG) and functional MRI (fMRI) across two studies to address the following questions: (i)
do emotions in dreams (here fear-related emotions) engage the same neural circuits as
during wakefulness, and (ii) is there a link between emotions experienced in dreams and
brain responses to emotional stimuli during wakefulness. By addressing these fundamental
and complementary topics, we aim at clarifying the grounding conditions for the study of
dreaming as pertaining to day/night affective homeostasis.
METHODS
Study 1: Neural correlates of fear in dreams
Participants
Eighteen healthy participants were included in Study 1 (4 males, age 39.77 ± 13.12 years, 25-
63 [mean ± SD, range]). From these 18 participants, twelve (N=12) were used for the analysis
of fear vs. no fear conditions in NREM sleep (N2 stage), while eight (N=8) were used for the
analysis of fear vs. no fear conditions in REM sleep. All participants had no history of
neurological or psychiatric disorder. Signed informed consent was obtained from all
participants before the experiment, and ethical approval for the study was obtained from
the University of Wisconsin-Madison Institutional Review Board.
Procedure
Serial awakenings during sleep
Dream sampling during sleep was accomplished using the ‘serial awakening’ method, as
described in detail elsewhere (Siclari F et al. 2017) (Fig. 1A). In brief, participants were
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
5
awakened several times during the night while sleeping and were asked to describe ‘the last
thing going through your mind prior to the alarm sound’, and then underwent a structured
interview via intercom. Among other questions, they were asked to name any specific
emotion that they experienced, and to report the presence/absence of fear or anxiety.
Awakenings were performed at intervals of at least 20 minutes in N2 sleep or REM sleep
using an alarm sound. Participants must have been asleep for a minimum of 10 minutes and
must have been in a stable sleep stage for a minimum of 5 minutes before any experimental
awakening was triggered.
EEG recordings
Recordings were made at the University of Wisconsin (Wisconsin Institute for Sleep and
Consciousness) using a 256-channel high-density EEG (hdEEG) system (Electrical Geodesics,
Inc., Eugene, Ore.) combined with Alice Sleepware (Philips Respironics, Murrysville, PA).
Additional polysomnography channels were used to record and monitor eye movements and
submental electromyography during sleep. Sleep scoring was performed over 30s epochs
according to standard criteria (Iber C et al. 2007).
EEG Preprocessing
The EEG signal was sampled at 500 Hz and band-pass filtered offline between 1 and 50 Hz.
The EEG data were high-pass filtered at 1Hz instead of lower frequencies as there were
sweating artifacts in some of the participants which caused intermittent high-amplitude
(>300uV) slow frequency oscillatory activity around 0.3 Hz. Noisy channels and epochs
containing artifactual activity were visually identified and removed. To remove ocular,
muscular, and electrocardiograph artifacts, we performed Independent Component Analysis
(ICA) using EEGLAB routines (Schenck CH et al. 1993; Jung TP et al. 2000). The previously
removed noisy channels were interpolated using spherical splines (EEGLAB). Finally, EEG
data was referenced to the average of all electrodes.
EEG Signal analysis
Source localization
The cleaned, filtered and average-referenced EEG signal corresponding to the 20s before the
alarm sound was extracted and analysed at the source level. Source modelling was
performed using the GeoSource software (Electrical Geodesics, Inc., Eugene, Ore.). A 4-shell
head model based on the Montreal Neurological Institute (MNI) atlas and a standard
coregistered set of electrode positions were used to construct the forward model. The
source space was restricted to 2447 dipoles in 3-dimensions that were distributed over
7x7x7 mm cortical voxels. The inverse matrix was computed using the standardized low-
resolution brain electromagnetic tomography (sLORETA) constraint. A Tikhonov
regularization procedure (λ=10-1) was applied to account for the variability in the signal-to-
noise ratio (Pascual-Marqui RD 2002). We computed spectral power density using the
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
6
Welch's modified periodogram method (implemented with the pwelch function in MATLAB
(The Math Works Inc, Natick, MA) in 2s Hamming windows (8 segments, 50% overlap) to
decompose the source signals into frequency bands of interest before taking the norm
across dimension to produce a single power value for each dipole.
Statistical Analysis
Statistical analysis was carried out in MATLAB. To compare brain activity between trials with
fear and those without, source-space power was averaged within standard frequency bands
(Delta: 1-4Hz, Theta: 5-8Hz, Alpha: 8.5-12Hz, Sigma: 12.5-17Hz, Beta: 17.5-24Hz, Gamma: 25-
50Hz). We then averaged the power values within trials with fear and those trials without
fear for each participant and for each frequency band separately. Group level analyses used
paired two-sample t-tests (two-tailed) between the fear and no fear conditions, performed
separately for each frequency band, and thresholded at corrected p< 0.05 using non-
parametric threshold-free cluster enhancement (TFCE) (weighing parameters E=0.5 and H=2)
(Mensen A and R Khatami 2013).
Study 2: Modulation of brain responses to aversive stimuli during wake as a function of
fear in dreams
Participants
A total of 127 healthy individuals (45 males, age 22.00 ± 3.15 years, 18-37 [mean ± SD,
range]) participated in Study 2. All participants had no history of neurological or psychiatric
disorder. Signed informed consent was obtained from all participants before the experiment,
and ethical approval for the study was obtained from the University of Geneva (Switzerland)
and University of Liège (Belgium). Among these 127 participants, 13 participants did not
report any dream in their dream diary before the experimental session (see below). For the
fMRI analyses, we also excluded 25 participants who were presented with emotional (and
neutral) words, unlike all other participants who saw emotional pictures. The final group of
89 participants included 58 females, 21.5 ± 2.4 (mean ± SD) year-old. Participants took part
in one of 3 different experiments (Exp. 1: N=28, age 21.36 ± 2.70 years, 16 men, University of
Geneva; Exp. 2: N=19, age 22.16 ± 2.48 years, 19 men, University of Geneva, Exp. 3: N=42,
age 21.31 ± 2.08, 23 men, University of Liège). All participants filled out the same
questionnaires on sleep quality (PSQI, Tzourio-Mazoyer N et al. 2002), daytime sleepiness
(Epworth sleepiness scale, ESS, Pascual-Marqui RD 2002), depression (Beck Depression Scale,
BDI, Mensen A and R Khatami 2013), anxiety (State-Trait Anxiety Inventory Trait, STAI-T,
Spielberger CD et al. 1970), and also kept the same sleep and dream diary. These results
correspond to a secondary analysis because part of the data was already presented
elsewhere (Sterpenich V, C Piguet, et al. 2014; Sterpenich V et al. 2017), yet without
exploring any of the dream measurements. Note that the sleep and dream diaries were
designed specifically for this analysis and the stimuli of the three different studies
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
7
correspond to similar visual items (emotional and neutral pictures, see below), eliciting
similar changes in emotional arousal and in local brain activition (Fig. S1).
Collection and analysis of dream data
During the week preceding the fMRI session, participants were asked to fill out a sleep and
dream diary at home (Fig. 2A). Every morning, they responded to a mini-questionnaire about
the content of the dreams they may have had during the preceding night, and had the
possibility to write down their dreams in more details. Among other questions in the mini-
questionnaire, participants were asked to report the presence/absence of specific emotions
in their dreams (anger, disgust, confusion, embarrassment, fear, sadness, joy, frustration).
Note that joyful dreams may be slightly over-represented because joy was the only positive
emotion among the emotions assessed. We computed the percentage of nights with dreams
containing specific emotions (for example, 3 nights with dreams containing fear over a total
of 5 nights with dreams leads to a percentage of fear in dreams of 60%). We entered the
percentage for each emotion from each participants (8 values by participant) into a Principal
Component Analysis (PCA) to i) identify main affective components (or dimensions)
according to variance explained and ii) characterize each participant by component (or
factor) scores, which we then used as regressors in an fMRI design matrix. Specifically, here
we used the individual scores on the second PCA component that contrasted basic negative
emotions, in particular fear, to non-basic social negative emotions (such as embarrassment
and frustration) (see Schwartz S 2004 for a similar data structure). Because of the
imbalanced number of options for negative (n=7) vs. positive (n=1) emotions in the original
data, we did not investigate the first PCA component. Indeed, and unsurprisingly, the latter
contrasted the only positive emotion to all other (negative) emotions.
Functional MRI session
Emotional tasks
Data from three different fMRI experiments were included in the analysis. Two of these
three sets of data have already been reported elsewhere, but none of these former
publications concerned emotions in dreams (Sterpenich V, C Piguet, et al. 2014; Sterpenich V
et al. 2017). Common to these three experiments was that participants were exposed to
aversive and neutral images, and that dream data were collected using the exact same
instructions and dream diary. Below, we briefly describe the task used in each experiment,
focusing only on those aspects that are relevant to the purpose of the present work.
In Experiment 1, participants were presented with conditioned (aversive) and
unconditioned faces. Stimuli were presented in an intermixed, random order, one at a time
(2.5 s each) followed by a varying interval (ISI: 4–5.5 s, mean=4.75) (see Sterpenich V, C
Piguet, et al. 2014 for more details). In Experiment 2, 60 aversive, 60 funny and 60 neutral
pictures were presented to participants. Each picture was displayed for 3s, preceded by a
fixation cross of 1 s. Participants had a maximum of 2s to judge the valence of each stimuli
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
8
on a 7 point scale (from -3: very negative, 0: neutral, +3: very funny). In Experiment 3,
participants saw 90 faces displaying a negative expression and 90 faces with a neutral
expression (see Nielsen T 2017 for more details). Each face was displayed for 3 s, and then
participants judged it for emotional valence and arousal. Each trial started with a fixation
cross for 1.5-s duration and ended as soon as participants responded, resulting in a jitter
between trials (range 5.03-12.1 s). For the analysis of each experiment, we compared
activity elicited by the presentation of aversive vs. neutral stimuli.
All visual stimuli were presented on a back projection screen inside the scanner bore using
an LCD projector, which the participant could comfortably see through a mirror mounted on
the head coil. Responses were recorded via an MRI-compatible response button box (HH-1 ×
4-CR, Current Designs Inc., USA).
Pupillary size
During all fMRI sessions, eye movements and pupil diameter were measured continuously
using an MRI-compatible long-range infrared eye tracking system (Applied Science
Laboratories, Bedford, MA, USA; sampling rate: 60 Hz). Pupil size variation was used as an
index of emotional arousal during the tasks (Bradley MM et al. 2008). Pupillary responses
were analyzed during epochs of 5 s following the onset of the presentation of pictures. For
each epoch, baseline pupil size was estimated as the average pupil measurement during the
second preceding the presentation of the picture, and was then subtracted from all values of
this epoch. Trials with more than 30% of signal loss were discarded. The pupillary values
were z-scored for each task to take into account the difference in luminance of the different
pictures and background. For each trial type (aversive and neutral) and for each task, we
analyzed the mean signal value over the 5 s epochs using a t-test (Fig. S1A). Data from 50
participants were discarded because of poor quality of the recording or technical problems,
and finally data from 77 participants were included in this analysis. The mean pupil diameter
value for aversive stimuli was subtracted from that for neutral stimuli for each participant to
obtain an individual physiological emotional response to fear-eliciting stimuli. To test for a
potential link with emotions in dreams, we correlated these individual pupil reactivity values
with the frequency of fear in dream.
MRI acquisition
For Experiments 1 and 2, MRI data were acquired on a 3 T whole body MR scanner (Tim Trio,
Siemens, Erlangen, Germany) using a 12-channel head coil. For Experiment 3, data were
acquired on a 3T head-only magnetic resonance scanner (Allegra, Siemens, Erlangen,
Germany). Functional images were acquired with a gradient-echo EPI sequence with the
following parameters. For Experiment 1: repetition time (TR): 2200ms, echo time (TE): 30ms,
flip angle (FA): 85°, field of view (FOV): 235 mm, 36 transverse slices, voxel size: 1.8 × 1.8 x
3.4 mm. For Experiment 2: TR: 2100 ms, TE: 30 ms, FA: 80°, FOV: 205 mm, 36 transverse
slices, voxel size: 3.2 x 3.2 x 3.8 mm. For Experiment 3: TR: 2460 ms, TE: 40 ms, FA: 90°, FOV:
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
9
220 mm, 32 transverse slices, voxel size: 3.4 x 3.2 x 3.4 mm. T1-weighted structural images
were also acquired in each participant.
MRI analysis
Functional volumes were analyzed by using Statistical Parametric Mapping 8 (SPM8;
www.fil.ion.ucl.ac.uk/spm/software/spm8) implemented in Matlab (The MathWorks Inc,
Natick, Massachusetts, USA). Functional MRI data were corrected for head motion, slice
timing and were spatially normalized to an echo planar imaging template conforming to the
Montreal Neurological Institute (MNI) template (voxel size, 3 × 3 × 3 mm). The data were
then spatially smoothed with a Gaussian kernel of 8 mm full width at half maximum
(FWHM). For each participant, we used a General Linear Model (GLM) approach to estimate
brain responses at every voxel, and computed the main contrast of interest (aversive vs.
neutral). For each experiment, any other conditions (e.g. positive items in Exp. 2) were
modeled as variable of no interest. Movement parameters estimated during realignment
were also added as regressors of no interest. The resulting individual maps of t-statistics (the
contrast images for each individual) were then used in second-level random-effects analyses.
We used one-sample t-tests for testing common effects of aversive vs. neutral stimuli. For
the regression analysis, individual scores on the second PCA component (higher for more
fearful basic negative emotion in dreams; see Collection and analysis of dream data section)
were used as a regressor at the group level for the contrast aversive vs. neutral stimuli.
Statistical inferences were corrected for multiple comparisons according to the Gaussian
random field theory at p<0.05 Family wise error (FWE) corrected i) on the entire volume or
ii) using correction within predefined anatomical regions (small volume correction, SVC),
including the amygdala, insula, and the midcingulate cortex using the toolbox Anatomy
(Tzourio-Mazoyer N et al. 2002).
RESULTS
Study 1: Identifying the neural correlates of fear in dreams
In Study 1, we performed serial awakenings in 18 participants recorded throughout the night
with hdEEG (256 channels) and identified brain regions activated prior to awakenings from a
dream containing fear (vs. without fear; Fig. 1A). We only analyzed the EEG data from those
participants who reported at least one dream containing fear and one without fear, within
the same recording night and for a given sleep stage (N2 or REM). We conducted spectral
analyses for the 20-seconds EEG epochs preceding each awakening and compared epochs
associated with the presence of fear in dreams with those without fear. We found significant
modulations in the delta (1-4 Hz) and gamma (25-50Hz) ranges, which we interpreted in
terms of underlying local neuronal population activity as follows. Increased (respectively
decreased) low frequency power in the delta range (<4Hz) corresponds to neuronal
inhibition (respectively activation) (Tononi G and M Massimini 2008; De Gennaro L et al.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
10
2011; Pigorini A et al. 2015), while increased high EEG frequencies in the gamma range
reflect increases in neuronal firing (Steriade M et al. 1996; Le Van Quyen M et al. 2010) and
positively correlate with local BOLD fluctuations (Murta T et al. 2015).
For NREM, we obtained a total of 79 awakenings in N2 sleep from 12 participants
(average per night 6.58 ± 2.13 [mean ± SD, range]). Of these awakenings, 57 yielded reports
of dream experience (from which 18 without recall of any content), while 22 yielded no
report. Fear was present in 14 reports (average per subject 1.16 ± 0.37 [mean ± SD, range])
and absent in 25 reports (average per subject 2.08 ± 1.03 [mean ± SD, range]). N2 reports
with presence of fear (compared to without fear) were associated with decreased delta
power (1-4Hz) in the right insula, and increased gamma power (25-50Hz) in the bilateral
insular cortex (Fig. 1B). No significant changes were found for other frequency bands.
For REM sleep, we obtained a total of 32 awakenings from 8 participants (average per
night 4.00 ± 0.86 [mean ± SD, range]). Of these awakenings, 28 yielded reports of dream
experience (from which 1 without recall of any content), while 4 yielded no report. Fear was
present in 12 reports (average per subject 1.50 ± 0.50 [mean ± SD, range]) and absent in 15
reports (average per subject 1.87 ± 1.05 [mean ± SD, range]). REM reports with presence of
fear compared to those without fear were associated with decreased delta power (1-4Hz) in
the bilateral insula and midcingulate cortex, thus partly replicating the results from N2 (Fig.
1B). No significant changes were found for other frequency bands. Although EEG source
reconstruction should be considered with caution, the present data also suggest that
experiencing fear in NREM and REM dreams could involve different portions of the insula,
with slightly more anterior dorsal insula activity during REM (Fig 1B). Together these results
demonstrate that the occurrence of frightening dreams coincided with increased activation
of insula cortex during both NREM and REM sleep, and of the midcingulate cortex during
REM sleep.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
11
Figure 1. A. Participants under hdEEG were awakened several times while sleeping in the
sleep laboratory and were asked to report “the last thing going through their [your] mind
prior to the alarm sound” and then underwent a structured interview via intercom.
Participants were also asked whether they felt fear. Twenty-second epochs of EEG recording
prior to each awakening were then sorted as a function of the presence or absence of fear in
the dream. B. Brain maps showing modulations of delta and gamma power during N2 sleep
and delta power during REM sleep when comparing trials with fear to those without fear.
Only significant differences at the p<.05 level, obtained after correction for multiple
comparisons, are shown at the source level (two-tailed, paired t-tests, TFCE corrected).
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
12
Study 2: Linking awake brain responses to aversive stimuli and fear in dreams
After we established that fear in dreams implicated the insula and midcingulate cortex
known to contribute to the processing of aversive stimuli during wakefulness (Study 1), we
asked whether frequently experiencing fear in dreams related to individual autonomic and
neural sensitivity to fear during wakefulness. To this end, we collected a large amount of
dream reports from 127 participants and correlated individual scores on a fear component
(see below) with pupillary responses and fMRI responses to aversive stimuli during
wakefulness recorded in the same participants (Fig. 2A). All participants filled the same sleep
and dream diary over 7.38 (± 4.90; mean ± SD) nights with the same instructions before an
MRI session (with a minimum of 3 nights). From the sleep diary, participants reported a
mean sleep duration of 7.93 (± 0.76 SD) hours per night and a sleep quality of 7.02 (± 0.97
SD; on a 10-point scale from 0 - very bad sleep to 10 - very good sleep quality). From the
dream diary, an average of 5.47 nights (± 3.63 SD) were associated with the feeling of having
dreamt, and participants answered specific questions related to the content of their dreams
after 3.75 (± 2.66 SD) nights. In particular, they were asked to indicate whether an emotion
was present/absent for 8 distinct emotions (see Table S1). We performed a principal
component analysis (PCA) on these data and found that the variance in emotion ratings was
best explained by two main components. The first component (20.39% of the variance)
distinguished between negative and positive (here “joy”) emotions, thus representing
emotional valence. The second component (16.40% of the variance) contrasted emotions
related to basic negative emotions (fear, anger, sadness, disgust), with fear having the
strongest contribution, and those related to non-basic, social negative emotions (i.e.,
embarrassment, confusion, frustration; see (Schwartz S 2004) for a similar finding). The
predominance of fear reported in dreams provides a first confirmation of our initial
hypothesis about the expression of this emotion during sleep. Accordingly, we shall call this
second “fear component” in the remainder of the manuscript. We did not use the first
component in our fMRI analysis primarily because it does not directly address our main
hypothesis about fear, and because of the imbalance between the number of negative and
positive emotions participants had to choose from (7 vs. 1; Table S1), as also captured by the
PCA. The individual PCA scores (or loadings) for the fear component did not correlate with
sleep quality (PSQI, R2 <0.001, P=0.51), sleep duration (sleep diary, R2=0.005, P=0.44),
sleepiness (ESS, R2=0.02, P=0.12), depression (BDI, R2=0.008, P=0.34), anxiety (STAI-T,
R2=0.001, P=0.77), or frequency of dreaming (percentage of nights with dreams, R2=0.003,
P=0.54). This pattern of results supports that the PCA fear component might represent an
individual affective measure that is largely independent of other sleep-related variables and
waking affective dimensions.
Study 2 aimed at establishing whether neurophysiological responses to fear-eliciting
emotions during wakefulness correlated with fear in dreams. We first analyzed changes in
pupillary size recorded during the presentation of aversive (vs. neutral; see Methods; Fig.
S1A) stimuli as an index of emotional arousal response (Bradley MM et al. 2008), and
showed that this index correlated with the individual PCA scores for the fear component
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
13
(Spearman’s rho=0.25, P=0.034). We next tested whether we could observe a similar
relationship between emotions in dreams and brain responses collected during the
presentation of the same aversive stimuli (vs. neutral). Using fMRI data collected on 89
participants (from 3 experiments, see Methods for more details), we first confirmed that
fear-eliciting stimuli activated a set of expected brain regions, including the amygdala, the
insula, and occipital regions (Table S2, Fig. S1B). Then, we used the individual PCA scores for
the fear component as a regressor in a whole-brain regression analysis (for the contrast
aversive vs. neutral stimuli, see Methods). This analysis revealed that participants with a high
propensity to experience fear in dreams had increased activation of the medial prefrontal
cortex cortex (Fig. 2B, Table S3) when facing aversive stimuli while awake. Conversely, and as
predicted by theoretical models (see Introduction), the same analysis yielded negative
correlations with activity in the right amygdala, right insula, and midcingulate cortex (Fig.
2B). Taken together, these results show that individuals who reported a high prevalence of
fear-related emotions in their dreams had stronger fear inhibition during wakefulness.
Critically, associated neural decreases implicated the insula and midcingulate, which were
both strongly activated during fearful dreams in Study 1, thus further supporting reciprocal
links between sleep and wake emotional functioning.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
14
Figure 2. A. During one week, participants filled a dream diary at home, and their responses
to aversive stimuli were assessed using fMRI. To test for a link between fear in dreams and
brain responses to fear-eliciting stimuli during wakefulness, the individual propensity to
experience fear in dreams (second PCA component, Table S1) was used as a regressor in a
whole-brain analysis. B. Responses of the medial prefrontal cortex to aversive stimuli were
greater in those individuals who frequently experienced fear in dreams (top panel), while
activation of the amygdala, insula, and midcingulate cortex decreased in the same
individuals (middle and bottom panels). Significant whole-brain regression results are
displayed on the mean structural image.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
15
DISCUSSION
Here we investigated the neural correlates of fear in dreams and their relation to brain
responses to threatening stimuli during wakefulness. In Study 1, we found that experiencing
fear (vs. no fear) in dreams was associated with the activation of the insula and midcingulate
cortex (the latter during REM dreams; Fig. 1B), which were both also activated when
experiencing fear during wakefulness (Study 2; Fig. S1B, Table S2), as also classically reported
in previous research (Pereira MG et al. 2010; Alves FH et al. 2013; Casanova JP et al. 2016).
We recently reported that specific dream contents—such as faces, places, movement,
speech, and thoughts— engage similar cortical networks as during wakefulness
(Perogamvros L et al. 2017; Siclari F et al. 2017). Here we show, for the first time to our
knowledge, that a specific emotional state, fear, activated the insula and midcingulate
during both dreaming and awake consciousness. The consistency of the present results
across brain states and their correspondence with classical work on brain structures involved
in fear are encouraging. Importantly, hdEEG (especially with 256 electrodes as in our study)
can have sufficient accuracy in source localization (e.g. associating dreams containing faces
with activation of the fusiform face area (Siclari F et al. 2017), even for the detection of
signal originating from deep cerebral structures (Seeber M et al. 2019). However, combined
EEG/fMRI studies are certainly needed to further elucidate the exact contribution of
subcortical structures, such as the amygdala, especially during REM sleep (Maquet P et al.
1996; De Gennaro L et al. 2011; Corsi-Cabrera M et al. 2016).
The insula, especially its anterior part, may contribute to social–emotional experience and
associated visceral states, possibly giving rise to conscious feelings (Critchley HD et al. 2004;
Chang LJ et al. 2013), and participates in the emotional response to distressing cognitive or
interoceptive signals (Reiman EM et al. 1997). Insula activation during dreaming could thus
reflect the integration of internally-generated sensory, affective and bodily information
culminating in a subjective feeling of danger, as we further discuss below. During REM sleep,
we also found an activation of the midcingulate cortex, a region known to be critically
involved in behavioral/motor responses to dangers (Pereira MG et al. 2010). Because REM
sleep is characterized by activation across sensory and motor cortices (Schwartz S and P
Maquet 2002), while muscle atonia prevents the overt expression of motor behaviors, this
sleep stage could provide a well-suited physiological condition for the (re)activation of
threatening situations with associated emotional and motor reactions.
Based on the results from Study 1, which suggested an anatomo-functional
correspondence between fear in dreams and wakefulness, we then asked whether
frequently experiencing fear in dreams may relate to the individual’s neurophysiological
sensitivity to fear during wakefulness. In Study 2, we analyzed awake fMRI response to
aversive stimuli as a function of whether participants reported a high incidence of fear in
their dreams. Here we thus considered fear in dreams as an individual trait, which we determined
based on the analysis of a large dataset of dreams. Of note, this measure of fear in dreams did
not correlate with depression, anxiety, sleep quality, and the frequency of dream recall (all
p>0.05), which supports the specificity of fear in dreams with respect to other classical
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
16
dimensions of mood or sleep/dreams. We found decreased activity in the insula, amygdala and
midcingulate cortex, while activity of the mPFC cortex was increased (Fig. 2B). Along with the
insula, both the amygdala and the cingulate cortex have been associated with fear and the
perception of negative emotions during wakefulness (Phan KL et al. 2002), as we also
demonstrated (Fig. S1B, Table S2). On the other hand, the mPFC is believed to regulate the
response to threatening stimuli by modulating the activity of the amygdala (Quirk GJ et al.
2003; Phelps EA et al. 2004). Specifically, the mPFC exerts an inhibitory control on fear
expression by decreasing amygdala output and has been associated with extinction learning
(i.e., when a neutral conditioned stimulus that previously predicted an aversive
unconditioned stimulus no longer does so, and conditioned response subsequently
decreases) (Kalisch R et al. 2006; Herry C et al. 2010). Consistent with the proposal that
dreaming may serve an emotion regulation function (Kramer M 1991; Hartmann E 1996;
Cartwright R et al. 2006; Perogamvros L et al. 2013), participants who frequently (but not
excessively, see below) experienced frightening dreams showed a stronger inhibition of the
amygdala, potentially mediated by the mPFC. This interpretation is further supported by the
pupillary results showing that participants who frequently reported fear in their dreams had
reduced autonomic responses to aversive stimuli during wakefulness, suggesting a better
ability to regulate defensive and alerting reactions to threatening signals in those individuals.
In the domain of sleep research (irrespective of dreaming), REM sleep was found to play a
role in emotional memory consolidation (Nishida M et al. 2009; Goldstein AN and MP Walker
2014; Sterpenich V, C Schmidt, et al. 2014), especially fear memory consolidation (Pace-
Schott EF et al. 2015) and successful fear/safety recall (Menz MM et al. 2016), while both
NREM (Hauner KK et al. 2013; He J et al. 2015) and REM sleep stages (Pace-Schott EF et al.
2015; Menz MM et al. 2016) have been found to promote the retention and generalization
of extinction learning. In addition, it was proposed that specifically REM sleep contributes to
the attenuation of the emotional tone of waking-life memories (Walker MP and E van der
Helm 2009; Vallat R et al. 2017). Importantly, total sleep deprivation may cause a reduction
of mPFC control over the limbic system, resulting in an accentuation of emotional responses
to negative stimuli (Yoo SS et al. 2007) and an impairment of extinction recall (Straus LD et
al. 2017). We have previously suggested that these findings from sleep studies may
putatively extend to or even depend on concomitant dreaming (Perogamvros L and S
Schwartz 2012; Perogamvros L et al. 2013). In particular, the exposure to feared stimuli
(objects, situations, thoughts, memories, and physical sensations) in a totally safe context
during dreaming would thus resemble desensitization therapy (Levin R and TA Nielsen 2007).
Besides, several studies have demonstrated that dreaming of negative waking-life
experiences (e.g. divorce) contributes in a resolution of previous emotional conflicts and a
reduction of next-day negative mood (Cartwright RD et al. 1984; Cartwright R et al. 2006).
While this was not the objective of the present work, we would like to emphasize that the
present data do not allow us to make any inference about whether one occurrence of one
specific emotion in a dream influences emotional state or responsiveness on the following
day.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
17
Contrasting with this beneficial role of negative but benign dreams, recurrent nightmares,
such as those observed in PTSD patients, might represent a failure of the fear extinction
function of dreaming (Nielsen T and R Levin 2007; Nielsen T 2017). Thus, nightmare patients
may be more prone to emotional dysregulation, as suggested by one recent study reporting
decreased mPFC activity during the viewing of negative pictures in these patients (Marquis L
et al. 2016). Furthermore, exerting ineffective emotional regulation strategies (e.g. fear
suppression) and elevated anxiety during wakefulness may lead to increased excitability of
negatively-loaded memories at sleep-onset or even during sleep (Schmidt RE and GH Gendolla 2008;
Malinowski J 2017; Sikka P et al. 2018; Sikka P et al. 2019), namely in conditions where monitoring
from the prefrontal cortex is reduced (Maquet P et al. 1996; Braun AR et al. 1997). Such disruption
in the regulation of emotions during wakefulness and sleep has been proposed as a major
contributing factor to insomnia (Wassing R et al. 2016).
Here, we experimentally show that, beyond sleeping, experiencing negative emotions
specifically during dreaming is associated with better-adapted emotional responses during
waking life. Study 2 combined data from three different experiments testing for brain
responses to aversive stimuli (vs. neutral). This allowed us to include a very large set of
participants, which is needed to exploit interindividual differences, as we do here. In all
three experiments, dream reports were collected using the exact same instructions and
same questionnaires. While we confirmed consistent fMRI and pupillary response results
across the three experiments for the effects of aversive vs. neutral emotions, Study 2 yielded
significant results in brain regions, for which we had strong theory-driven a priori. We
therefore suggest that what may be perceived as a potential limitation (i.e. combining data
from 3 experiments) may actually offer a better generalizability of the present findings to
diverse waking threatening situations.
Taken together, across two complementary studies, we show opposing neural effects of fear
experience in dreams and during wakefulness. These results support recent theoretical
claims that dreaming (beyond sleep) benefits emotion regulation processes, by achieving a
form of overnight affective simulation or recalibration (e.g. through extinction learning and
generalization), which would foster adapted emotional responses to dangerous real-life
events (Revonsuo A 2000; Nielsen T and R Levin 2007; Perogamvros L and S Schwartz 2012).
Studying the role of positive emotions (e.g. positive social interactions) in dreams (especially
in NREM dreams, (McNamara P et al. 2005) and their potential links with emotional brain
responses during wakefulness may be needed to further corroborate or refine existing
theoretical models. Finally, based on our results, we would like to suggest that future studies
should address how sleep and dreaming may influence exposure and extinction-based
therapies for affective disorders (Pace-Schott EF et al. 2018).
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
18
ACKNOWLEDGMENTS
This study was supported by the Swiss National Science Foundation Grants 155120 (to L.P.)
and 320030-159862 (to S.S.), NIH/NCCAM P01AT004952 (to G.T.), NIH/NIMH
5P20MH077967 (to G.T.), Tiny Blue Dot Inc. grant MSN196438/AAC1335 (to G.T.).
AUTHOR CONTRIBUTIONS
VS, LP, GT, and SS designed the experiments, VS and LP conducted the experiments, VS, LP,
and SS analyzed the data, VS, LP, GT and SS wrote the paper.
DECLARATION OF INTERESTS
The authors declare no competing interests.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
19
REFERENCES
Alves FH, Gomes FV, Reis DG, Crestani CC, Correa FM, Guimaraes FS, Resstel LB. 2013. Involvement of
the insular cortex in the consolidation and expression of contextual fear conditioning. Eur J Neurosci
38:2300-2307.
Blake Y, Terburg D, Balchin R, van Honk J, Solms M. 2019. The role of the basolateral amygdala in
dreaming. Cortex; a journal devoted to the study of the nervous system and behavior 113:169-183.
Boyce R, Glasgow SD, Williams S, Adamantidis A. 2016. Causal evidence for the role of REM sleep
theta rhythm in contextual memory consolidation. Science 352:812-816.
Bradley MM, Miccoli L, Escrig MA, Lang PJ. 2008. The pupil as a measure of emotional arousal and
autonomic activation. Psychophysiology 45:602-607.
Braun AR, Balkin TJ, Wesenten NJ, Carson RE, Varga M, Baldwin P, Selbie S, Belenky G, Herscovitch P.
1997. Regional cerebral blood flow throughout the sleep-wake cycle. An H2(15)O PET study. Brain : a
journal of neurology 120 ( Pt 7):1173-1197.
Carr M, Nielsen T. 2015. Daydreams and nap dreams: Content comparisons. Conscious Cogn 36:196-
205.
Cartwright R, Agargun MY, Kirkby J, Friedman JK. 2006. Relation of dreams to waking concerns.
Psychiatry Res 141:261-270.
Cartwright R, Luten A, Young M, Mercer P, Bears M. 1998. Role of REM sleep and dream affect in
overnight mood regulation: a study of normal volunteers. Psychiatry Res 81:1-8.
Cartwright RD, Lloyd S, Knight S, Trenholme I. 1984. Broken dreams: a study of the effects of divorce
and depression on dream content. Psychiatry 47:251-259.
Casanova JP, Madrid C, Contreras M, Rodriguez M, Vasquez M, Torrealba F. 2016. A role for the
interoceptive insular cortex in the consolidation of learned fear. Behav Brain Res 296:70-77.
Chang LJ, Yarkoni T, Khaw MW, Sanfey AG. 2013. Decoding the role of the insula in human cognition:
functional parcellation and large-scale reverse inference. Cereb Cortex 23:739-749.
Corsi-Cabrera M, Velasco F, Del Rio-Portilla Y, Armony JL, Trejo-Martinez D, Guevara MA, Velasco AL.
2016. Human amygdala activation during rapid eye movements of rapid eye movement sleep: an
intracranial study. J Sleep Res 25:576-582.
Critchley HD, Wiens S, Rotshtein P, Ohman A, Dolan RJ. 2004. Neural systems supporting
interoceptive awareness. Nat Neurosci 7:189-195.
De Gennaro L, Cipolli C, Cherubini A, Assogna F, Cacciari C, Marzano C, Curcio G, Ferrara M,
Caltagirone C, Spalletta G. 2011. Amygdala and hippocampus volumetry and diffusivity in relation to
dreaming. Hum Brain Mapp 32:1458-1470.
Dunsmoor JE, Kroes MCW, Li J, Daw ND, Simpson HB, Phelps EA. 2019. Role of Human Ventromedial
Prefrontal Cortex in Learning and Recall of Enhanced Extinction. J Neurosci 39:3264-3276.
Fosse R, Stickgold R, Hobson JA. 2001. The mind in REM sleep: reports of emotional experience. Sleep
24:947-955.
Goldstein AN, Walker MP. 2014. The role of sleep in emotional brain function. Annual review of
clinical psychology 10:679-708.
Hartmann E. 1996. Outline for a theory on the nature and functions of dreaming. Dreaming 6:147-
170.
Hauner KK, Howard JD, Zelano C, Gottfried JA. 2013. Stimulus-specific enhancement of fear
extinction during slow-wave sleep. Nat Neurosci 16:1553-1555.
He J, Sun HQ, Li SX, Zhang WH, Shi J, Ai SZ, Li Y, Li XJ, Tang XD, Lu L. 2015. Effect of conditioned
stimulus exposure during slow wave sleep on fear memory extinction in humans. Sleep 38:423-431.
Herry C, Ferraguti F, Singewald N, Letzkus JJ, Ehrlich I, Luthi A. 2010. Neuronal circuits of fear
extinction. Eur J Neurosci 31:599-612.
Hobson JA, Pace-Schott EF. 2002. The cognitive neuroscience of sleep: neuronal systems,
consciousness and learning. Nat Rev Neurosci 3:679-693.
Iber C, Ancoli-Israel S, Chesson A, Quan SF. 2007. The AASM manual for the scoring of sleep and
associated events: Rules, terminology and technical specifications. Westchester.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
20
Jung TP, Makeig S, Humphries C, Lee TW, McKeown MJ, Iragui V, Sejnowski TJ. 2000. Removing
electroencephalographic artifacts by blind source separation. Psychophysiology 37:163-178.
Kalisch R, Korenfeld E, Stephan KE, Weiskopf N, Seymour B, Dolan RJ. 2006. Context-dependent
human extinction memory is mediated by a ventromedial prefrontal and hippocampal network.
Journal of Neuroscience 26:9503-9511.
Kamphuis J, Meerlo P, Koolhaas JM, Lancel M. 2012. Poor sleep as a potential causal factor in
aggression and violence. Sleep medicine 13:327-334.
Kramer M. 1991. The nightmare: A failure in dream function. Dreaming 1:277-285.
Le Van Quyen M, Staba R, Bragin A, Dickson C, Valderrama M, Fried I, Engel J. 2010. Large-scale
microelectrode recordings of high-frequency gamma oscillations in human cortex during sleep. J
Neurosci 30:7770-7782.
Levin R, Nielsen TA. 2007. Disturbed dreaming, posttraumatic stress disorder, and affect distress: a
review and neurocognitive model. Psychol Bull 133:482-528.
Malcolm-Smith S, Koopowitz S, Pantelis E, Solms M. 2012. Approach/avoidance in dreams.
Consciousness and cognition 21:408-412.
Malinowski J. 2017. High Thought Suppressors Dream More of Their Negative Waking-Life
Experiences Than Low Thought Suppressors. Dreaming 27:269-277.
Maquet P, Peters J, Aerts J, Delfiore G, Degueldre C, Luxen A, Franck G. 1996. Functional
neuroanatomy of human rapid-eye-movement sleep and dreaming. Nature 383:163-166.
Marquis L, Blanchette-Carrière C, Carr M, Julien S, Paquette T, Nielsen T. 2016. Decreased activity in
medial prefrontal cortex and anterior cingulate cortex in idiopathic nightmare sufferers during
wakefulness. In. 30th Anniversary Meeting of the Associated Professional Sleep Societies. Denver,
Colorado, U.S.A. p A226-227.
McNamara P, McLaren D, Smith D, Brown A, Stickgold R. 2005. A "Jekyll and Hyde" within -
Aggressive versus friendly interactions in REM and non-REM dreams. Psychological Science 16:130-
136.
Mensen A, Khatami R. 2013. Advanced EEG analysis using threshold-free cluster-enhancement and
non-parametric statistics. Neuroimage 67:111-118.
Menz MM, Rihm JS, Buchel C. 2016. REM Sleep Is Causal to Successful Consolidation of Dangerous
and Safety Stimuli and Reduces Return of Fear after Extinction. J Neurosci 36:2148-2160.
Merritt J, Stickgold R, Pace-Schott E, Williams J, Hobson J. 1994 Emotion profiles in the dreams of
men and women. Consciousness and Cognition 3:46-60.
Murta T, Leite M, Carmichael DW, Figueiredo P, Lemieux L. 2015. Electrophysiological correlates of
the BOLD signal for EEG-informed fMRI. Hum Brain Mapp 36:391-414.
Nielsen T. 2017. The Stress Acceleration Hypothesis of Nightmares. Front Neurol 8:201.
Nielsen T, Levin R. 2007. Nightmares: a new neurocognitive model. Sleep Med Rev 11:295-310.
Nishida M, Pearsall J, Buckner RL, Walker MP. 2009. REM sleep, prefrontal theta, and the
consolidation of human emotional memory. Cerebral cortex 19:1158-1166.
Nofzinger EA, Mintun MA, Wiseman M, Kupfer DJ, Moore RY. 1997. Forebrain activation in REM
sleep: an FDG PET study. Brain Res 770:192-201.
Pace-Schott EF, Bottary RM, Kim SY, Rosencrans PL, Vijayakumar S, Orr SP, Lasko NB, Goetter EM,
Baker AW, Bianchi MT, Gannon K, Hoeppner SS, Hofmann SG, Simon NM. 2018. Effects of post-
exposure naps on exposure therapy for social anxiety. Psychiatry research 270:523-530.
Pace-Schott EF, Germain A, Milad MR. 2015. Effects of sleep on memory for conditioned fear and
fear extinction. Psychol Bull 141:835-857.
Pascual-Marqui RD. 2002. Standardized low-resolution brain electromagnetic tomography (sLORETA):
technical details. Methods Find Exp Clin Pharmacol 24 Suppl D:5-12.
Pereira MG, de Oliveira L, Erthal FS, Joffily M, Mocaiber IF, Volchan E, Pessoa L. 2010. Emotion affects
action: Midcingulate cortex as a pivotal node of interaction between negative emotion and motor
signals. Cogn Affect Behav Neurosci 10:94-106.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
21
Perogamvros L, Baird B, Seibold M, Riedner B, Boly M, Tononi G. 2017. The Phenomenal Contents
and Neural Correlates of Spontaneous Thoughts across Wakefulness, NREM Sleep, and REM Sleep. J
Cogn Neurosci:1-12.
Perogamvros L, Dang-Vu TT, Desseilles M, Schwartz S. 2013. Sleep and dreaming are for important
matters. Front Psychol 4:474.
Perogamvros L, Schwartz S. 2012. The roles of the reward system in sleep and dreaming. Neurosci
Biobehav Rev.
Phan KL, Wager T, Taylor SF, Liberzon I. 2002. Functional neuroanatomy of emotion: a meta-analysis
of emotion activation studies in PET and fMRI. Neuroimage 16:331-348.
Phelps EA, Delgado MR, Nearing KI, LeDoux JE. 2004. Extinction learning in humans: role of the
amygdala and vmPFC. Neuron 43:897-905.
Pigorini A, Sarasso S, Proserpio P, Szymanski C, Arnulfo G, Casarotto S, Fecchio M, Rosanova M,
Mariotti M, Lo Russo G, Palva JM, Nobili L, Massimini M. 2015. Bistability breaks-off deterministic
responses to intracortical stimulation during non-REM sleep. Neuroimage 112:105-113.
Quirk GJ, Likhtik E, Pelletier JG, Pare D. 2003. Stimulation of medial prefrontal cortex decreases the
responsiveness of central amygdala output neurons. J Neurosci 23:8800-8807.
Reiman EM, Lane RD, Ahern GL, Schwartz GE, Davidson RJ, Friston KJ, Yun LS, Chen K. 1997.
Neuroanatomical correlates of externally and internally generated human emotion. Am J Psychiatry
154:918-925.
Revonsuo A. 2000. The reinterpretation of dreams: an evolutionary hypothesis of the function of
dreaming. The Behavioral and brain sciences 23:877-901; discussion 904-1121.
Roussy F, Brunette M, Mercier P, Gonthier I, Grenier J, Sirois-Berliss M, Lortie-Lussier M, De Koninck
J. 2000. Daily events and dream content: Unsuccessful matching attempts. Dreaming 10:77-83.
Scarpelli S, Bartolacci C, D'Atri A, Gorgoni M, De Gennaro L. 2019. The Functional Role of Dreaming in
Emotional Processes. Frontiers in psychology 10:459.
Schabus M, Dang-Vu TT, Albouy G, Balteau E, Boly M, Carrier J, Darsaud A, Degueldre C, Desseilles M,
Gais S, Phillips C, Rauchs G, Schnakers C, Sterpenich V, Vandewalle G, Luxen A, Maquet P. 2007.
Hemodynamic cerebral correlates of sleep spindles during human non-rapid eye movement sleep.
Proc Natl Acad Sci U S A 104:13164-13169.
Schenck CH, Hurwitz TD, Mahowald MW. 1993. Symposium: Normal and abnormal REM sleep
regulation: REM sleep behaviour disorder: an update on a series of 96 patients and a review of the
world literature. J Sleep Res 2:224-231.
Schmidt RE, Gendolla GH. 2008. Dreaming of white bears: the return of the suppressed at sleep
onset. Conscious Cogn 17:714-724.
Schredl M. 2010. Characteristics and contents of dreams. Int Rev Neurobiol 92:135-154.
Schredl M, Doll E. 1998. Emotions in diary dreams. Conscious Cogn 7:634-646.
Schwartz S. 2003. Are life episodes replayed during dreaming? Trends Cogn Sci 7:325-327.
Schwartz S. 2004. What dreaming can reveal about cognitive and brain functions during sleep: A
lexico-statistical analysis of dream reports. Psychological Belgica 44:5-42.
Schwartz S, Maquet P. 2002. Sleep imaging and the neuro-psychological assessment of dreams.
Trends Cogn Sci 6:23-30.
Seeber M, Cantonas LM, Hoevels M, Sesia T, Visser-Vandewalle V, Michel CM. 2019. Subcortical
electrophysiological activity is detectable with high-density EEG source imaging. Nature
communications 10:753.
Siclari F, Baird B, Perogamvros L, Bernardi G, LaRocque JJ, Riedner B, Boly M, Postle BR, Tononi G.
2017. The neural correlates of dreaming. Nat Neurosci 20:872-878.
Sikka P, Pesonen H, Revonsuo A. 2018. Peace of mind and anxiety in the waking state are related to
the affective content of dreams. Sci Rep 8:12762.
Sikka P, Revonsuo A, Noreika V, Valli K. 2019. EEG Frontal Alpha Asymmetry and Dream Affect: Alpha
Oscillations Over the Right Frontal Cortex During REM Sleep and Pre-Sleep Wakefulness Predict
Anger in REM Sleep Dreams. J Neurosci.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
22
Smith MR, Antrobus JS, Gordon E, Tucker MA, Hirota Y, Wamsley EJ, Ross L, Doan T, Chaklader A,
Emery RN. 2004. Motivation and affect in REM sleep and the mentation reporting process. Conscious
Cogn 13:501-511.
Spielberger CD, Gorsuch RL, Lushene RE. 1970. Manual for the State-Trait Anxiety Inventory. Palo
Alto, CA.
Steriade M, Contreras D, Amzica F, Timofeev I. 1996. Synchronization of fast (30-40 Hz) spontaneous
oscillations in intrathalamic and thalamocortical networks. Journal of Neuroscience 16:2788-2808.
Sterpenich V, Ceravolo L, Schwartz S. 2017. Sleep deprivation disrupts the contribution of the
hippocampus to the formation of novel lexical associations. Brain and language 167:61-71.
Sterpenich V, Piguet C, Desseilles M, Ceravolo L, Gschwind M, Van De Ville D, Vuilleumier P, Schwartz
S. 2014. Sleep sharpens sensory stimulus coding in human visual cortex after fear conditioning.
Neuroimage 100:608-618.
Sterpenich V, Schmidt C, Albouy G, Matarazzo L, Vanhaudenhuyse A, Boveroux P, Degueldre C,
Leclercq Y, Balteau E, Collette F, Luxen A, Phillips C, Maquet P. 2014. Memory reactivation during
rapid eye movement sleep promotes its generalization and integration in cortical stores. Sleep
37:1061-1075, 1075A-1075B.
Straus LD, Acheson DT, Risbrough VB, Drummond SPA. 2017. Sleep Deprivation Disrupts Recall of
Conditioned Fear Extinction. Biol Psychiatry Cogn Neurosci Neuroimaging 2:123-129.
Tononi G, Massimini M. 2008. Why does consciousness fade in early sleep? Ann N Y Acad Sci
1129:330-334.
Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, Mazoyer B, Joliot
M. 2002. Automated anatomical labeling of activations in SPM using a macroscopic anatomical
parcellation of the MNI MRI single-subject brain. Neuroimage 15:273-289.
Vallat R, Chatard B, Blagrove M, Ruby P. 2017. Characteristics of the memory sources of dreams: A
new version of the content-matching paradigm to take mundane and remote memories into account.
PLoS One 12:e0185262.
Valli K, Revonsuo A. 2009. The threat simulation theory in light of recent empirical evidence: a
review. Am J Psychol 122:17-38.
Valli K, Revonsuo A, Palkas O, Ismail KH, Ali KJ, Punamaki RL. 2005. The threat simulation theory of
the evolutionary function of dreaming: Evidence from dreams of traumatized children. Conscious
Cogn 14:188-218.
Wagner U, Hallschmid M, Rasch B, Born J. 2006. Brief sleep after learning keeps emotional memories
alive for years. Biological psychiatry 60:788-790.
Walker MP, van der Helm E. 2009. Overnight therapy? The role of sleep in emotional brain
processing. Psychol Bull 135:731-748.
Wassing R, Benjamins JS, Dekker K, Moens S, Spiegelhalder K, Feige B, Riemann D, van der Sluis S,
Van Der Werf YD, Talamini LM, Walker MP, Schalkwijk F, Van Someren EJ. 2016. Slow dissolving of
emotional distress contributes to hyperarousal. Proc Natl Acad Sci U S A 113:2538-2543.
Yoo SS, Gujar N, Hu P, Jolesz FA, Walker MP. 2007. The human emotional brain without sleep--a
prefrontal amygdala disconnect. Current biology : CB 17:R877-878.
Zohar D, Tzischinsky O, Epstein R, Lavie P. 2005. The effects of sleep loss on medical residents'
emotional reactions to work events: a cognitive-energy model. Sleep 28:47-54.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
23
SUPPLEMENTAL INFORMATION
Table S1: PCA analysis of the emotions in dreams.
Type of emotion
Percentage of
emotion in dreams
PCA component 1
(correlation
coefficient)
PCA component 2
(correlation
coefficient)
Fear
19.42 ± 27.38%
-0.40
0.36
Anger
15.63 ± 23.11%
-0.65
0.16
Sadness
9.65 ± 20.10%
-0.41
0.07
Disgust
5.15 ± 13.82%
-0.62
0.04
Frustration
13.21 ± 21.16%
-0.62
-0.33
Confusion
18.05 ± 25.44%
-0.35
-0.53
Embarrassment
8.31 ± 18.59%
-0.08
-0.59
Joy
34.84 ± 32.41%
0.22
-0.64
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
24
Table S2: Functional MRI responses to aversive (vs. neutral) stimuli across the three fMRI
experiments. Results are corrected for multiple comparisons using family-wise error
correction at p<0.05 (FWE-corr) i) on the entire volume or ii) on predefined anatomical
regions of interest (marked as +).
Name
MNI coordinates
Z-score
P
FWE-corr
Aversive > Neutral
Amygdala
20, -4, -16
4.51
<0.001+
Anterior insula
-36, 26, -6
6.24
<0.001
32, 16, -18
5.20
0.003
Inferior frontal gyrus
-32, 14, -26
5.87
<0.001
44,32,-4
5.64
<0.001
Motor cortex
-42, 0, 46
4.95
0.010
Anterior middle temporal
gyrus
50, -2, -24
6.74
<0.001
Posterior middle temporal
gyrus
-54, -58, 10
6.16
<0.001
52, -56, 14
6.06
<0.001
Thalamus
0,-16,-4
6.45
<0.001
Superior temporal sulcus
50, -18, 10
5.75
<0.001
Calcarine/lingual gyrus
-10, -78, 8
4.87
0.015
20, -60, 4
5.19
0.003
Cuneus
14, -100, 6
4.98
0.009
-28, -86, -6
4.90
0.013
Midcingulate cortex
10, 18, 42
4.43
0.002+
Superior frontal sulcus
42, 12, 24
5.04
0.007
Aversive < Neutral
Medial PFC
10, 44, -8
4.77
0.009
Motor cortex
-52, -16, 50
4.91
0.013
Angular gyrus
50, -62, 46
4.99
0.009
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
25
Table S3: Whole-brain correlation between fMRI activity elicited by aversive (vs. neutral)
stimuli during wakefulness and basic (fear) emotions in dreams (as captured by the score on
the second PCA component). Results are corrected for multiple comparisons using family-
wise error correction at p<0.05 (FWE-corr) i) on the entire volume or ii) on predefined
anatomical regions of interest (marked as ᵻ).
Name
MNI
coordinates
Z-score
P
FWE-corr
Positive correlations
Medial PFC
12,56,2
5.59
0.004
Negative correlations
Amygdala
26, 4, -18
3.48
0.008 ᵻ
Amygdala/striatum
22,0,-12
3.48
0.006 ᵻ
Anterior insula
-30, 26, 4
32, 22, -2
3.33
3.65
0.058 ᵻ
0.019 ᵻ
Midcingulate cortex
12, 2, 54
4.68
0.034
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
26
Figure S1. A. Pupil diameter change in response to aversive and neutral stimuli, showing a
significant difference between aversive (plain color) and neutral (hatched) conditions for
each experiment (Exp. 1: t=4.79, p<0.001; Exp. 2: t=5.15, p<0.001; Exp. 3: t=3.49, P=0.002).
B. Regional changes in fMRI signal in response to aversive vs. neutral stimuli across the three
fMRI experiments. Right panels show the parameters estimates for each experiment
separately for aversive and neutral stimuli. For display purposes, whole-brain results are
displayed on the mean structural image at P=0.001 uncorrected.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/534099doi: bioRxiv preprint first posted online Jan. 29, 2019;
