Content uploaded by Tristan Bekinschtein
Author content
All content in this area was uploaded by Tristan Bekinschtein
Content may be subject to copyright.
1
'Cognitive processing during the transition to sleep'
Authors: Louise Goupil and Tristan A. Bekinschtein
Medical Research Council Cognition and Brain sciences Unit, Cambridge
Abstract
Dramatic physiological and behavioural changes occur during the transition from wakefulness to
sleep. The process is regarded as a grey area of consciousness between attentive wakefulness and
slow wave sleep. Although neurophysiological changes, as the emergence of EEG grapho-
elements (spindles, K-complexes), changes in power spectra and coherence, thalamocortical
connectivity in fMRI, and single neuron changes in firing patterns, have been well characterised
during the falling asleep process, little attempt has been made to link these modifications to the
cognitive and behavioural dynamics of the transition. We revise here the body and brain
physiology, behaviour and phenomenology of these changes of consciousness and propose an
experimental framework to integrate the two aspects of consciousness that interact in the
transition, wakefulness and awareness.
Word count: 6973.
Tristan Bekinschtein (trisbek@gmail.com) and Louise Goupil (louisegoupil@hotmal.fr)
MRC Cognition and Brain Sciences Unit
15 Chaucer Road, CB2 7EF
Cambridge, UK.
Telephone: +44 1223 355294
Fax: +44 1223 359062
2
General introduction
Every night we descend unhurriedly in Hypnos arms. How exactly we lose consciousness
while falling asleep is far for being characterised. A few studies point to a slow, in the range of
minutes, cascade of events (Ogilvie, 2001), while others try to find the “point of sleep onset”.
This transition from full wakefulness to light sleep manifests in all levels, from molecular to
phenomenological, and has its most famous expression in the change of the “Berger wave”
(alpha wave) described by Berger when inventing the EEG 80 years ago (Berger, 1940).
Moreover, every day in thousands of hospitals and clinics around the world, patients fall asleep
and most of the data collected between wakefulness and the onset of sleep stage 2 is discarded.
Despite long clinical tradition of looking at this transition of consciousness, little scientific
research has been done to determine its dynamics. Our aim here was to review the scarce and
sparse literature about the process of falling asleep and to build bridges between different levels
of explanation and paradigms.
Awareness of the environment and of the self (i.e., content of consciousness) and
wakefulness (i.e., level of consciousness) are two aspects of consciousness that are regularly used
to separate arousal level from capacity of conscious awareness (i.e. ability to make a volitional
response). Classical accounts of consciousness assume that you need to be awake in order to be
aware; however, not being aware because you are asleep is the most common, nevertheless
interesting case of simple dissociation between these aspects. During NREM (slow-wave sleep)
you are difficult to awake and you are unaware (no contents of consciousness), whereas in REM-
sleep many people seem to have contents of consciousness (dreams and mentation) but remain
difficult to arouse. In both states (REM and NREM) you are asleep (and unconscious) and do
not behaviourally respond to command, but it seems that awareness is dissociated. Thus, it is
clear that differentiating wakefulness levels (asleep/awake states) is by no means the same thing
as differentiating awareness levels (conscious/unconscious states). Moreover, the interaction
between wakefulness and awareness has hardly been studied. What are the commonalities and
differences between the transition from being aware to unaware and the transition from being
awake to asleep?
What is sleep, then?
3
Sleep is normally described as a recurring state of reduced or lack of consciousness,
inactivity of voluntary muscles, decreased ability to react to stimuli (as compared to quiet
wakefulness), but is more easily reversible than coma, vegetative state or deep sedation. This
modern definition is certainly more accurate than those commonly found in textbooks a few
decades ago, when sleep was assumed to be simply “a state of inactivity occurring passively when
organs became fatigued” (Pelayo and Guilleminault, 2009). More orderly, a short account to
decide what is sleep can be found by using these four criteria: a) a state of lethargy (lassitude,
inactivity); b) a specific posture (lying down in humans); c) reduced response to stimulation
(higher threshold of arousal); d) reversibility, we can wake up from sleep (different from
disorders of consciousness).
Sleep is also described in terms of the two-process model of sleep regulation, the
interaction of the homeostatic Process S and the circadian Process C (Borberly et al. 1982, 2005).
The ultradian process completes the picture by representing the two main sleep stages (REM and
NREM). The model is good in the scale of days and can predict the contribution of homeostasis
and circadian Clock in sleep patterns. However, for the cognitive dynamics of the sleep onset
process -the purpose of this review- there are no established models that we could bring to help
in the explanation, although we will discuss some simple models that could soon become a tool
to look at transitions.
Emergence and physiological mechanisms of sleep cycles
In the history of evolution, before animals began to sleep they were already cycling.
First, there was circadian rhythmicity, and sleep evolved later, possibly when the community of
neurons (in simple chordates) became big and complex enough to profit from new functions
complementing the circadian clock.
Wakefulness cycles exist since the moment a molecular feedback loop evolved hundreds
years ago under the pressure of the light-dark cycle (Dvornyk et al., 2003). In mammals the
central circadian clock that time us is in the base of the hypothalamus, a few thousands cell in the
suprachiasmatic nucleus (Schwartz, 2009). Sleep is under strong circadian regulation but its
molecular and cellular mechanisms have become more widespread that the central circadian
clock itself (Zimmerman et al., 2008). Wake-sleep cycling seems to be controlled by the
interaction between the anterior hypothalamus (where the central clock sits) and more ventral
4
and lateral parts. The ventrolateral preoptic nucleus possibly sets off sleep onset by mutual
inhibition of arousal systems in the brainstem, the posterior hypothalamus and basal forebrain.
These networks are most likely closely modulated by the orexinergic arousal system of the lateral
hypothalamus (Pace-Schott and Hobson, 2002). These structures are not directly accessible in
humans, posing limitations to unravel the cellular and molecular mechanisms of changes of
consciousness states.
Physiological modifications at sleep onset
Sleep cycle progression is reflected in every biological level of organisms. Thus, a set of
changes can be observed at the physiological level while approaching sleep.
The most notorious and studied of these modifications is probably the appearance of
slow rolling eye movements (SEMs) during the state of drowsiness, which ceases before the first
sleep spindles (Oglivie, 2001; Magosso et al., 2007; De Gennaro et al., 2000; Hori et al., 1982).
These SEMs have been closely related to the EEG spectral changes associated with sleep onset
(SO), and peak just before the cessation of behavioural responsiveness (Ogilvie, 2001). They are
also simultaneous to another sleep related physiological modification, namely the decrease
observed in skin potential negativity (Susmakova et al., 2008; Hori et al., 1982).
Thermoregulation is also modulated. The core body temperature decreases before sleep
onset, along with an increase in peripheral temperature (Van Den Heuvel et al., 1998).
Furthermore, the balance between the two components of the autonomous nervous system is
reversed during the sleep onset period, with the parasympathetic influence becoming dominant,
resulting in a decreased heart rate, even before the onset of classic sleep stage 1 (Pivik and
Busby, 1996; Baharav et al., 1995). Respiratory activity is also reduced, the decrease happens in
parallel to the alpha to theta EEG power dominance shift characteristic of sleep onset (Worsnop
et al. 1998), and to the slowing of reaction times (Ogilvie and Wilkinson, 1989).
All these physiological modifications appear thereabouts in parallel before the onset of
classical sleep stage 1. Hence, it is tempting to suggest that they are subserved by common
underlying mechanisms of sleep onset anticipation.
5
Neurophysiology and the brain signatures of the process of falling asleep
The change in conscious state at sleep onset is accompanied by complex
neurophysiological changes, classically studied through EEG/polysomnographic recordings. The
transition is characterized by neurophysiological changes that have some of its most famous
expression in the power changes of the EEG. But there is more to oscillatory brain activity than
meets the eye.
Spontaneous electroencephalographic fluctuations
The brain signatures of sleep onset are well characterised in scalp EEG and give rise to
the definition of sleep stages that are now accepted by the clinical sleep societies worldwide
(Silber et al., 2007). Nevertheless staging still has a high degree of subjectivity since it relies on
visual scoring; and despite humans are excellent at pattern recognition, the variability of experts
rating the same sleep polysomnographic recordings is huge (Danker-Hopfe et al., 2009). As it
happens with any defined measure, there are arbitrary components and the classic sleep staging
system is no exception. To further characterize the transitions it is necessary to make use of finer
grained measures. Under the current criteria, 20-30 seconds are the minimum units to define the
stage, but it is clear from any EEG measures and from the timing of the cognitive processes that
this period may contain several microstates (Hori, 1985). Since several time scales of changes can
be defined in the transition between being asleep and being awake (tonic, phasic and transient)
(Makeig et al.,et al., 1996), research on transitions is better place when using shorter and richer
measures.
Qualitative EEG patterns of the transition
Hori and collaborators acknowledged the time scale limitation of using 30sec epochs
more than 20 years ago, and created a visual scale relying on the graphoelements defined in the
classic scale. Theta and alpha waves, vertex shape waves, k-complexes, spindles and slow wave
rhythms were used to generate a scale (the Hori Scale) taking 5 seconds as the minimum unit and
using a broader coverage of the scalp.
6
Nine Stages were defined in this manner, from drowsy (H1 and H2), to classical sleep
stage 2 onset (H9) and the appearance of sleep spindles. In between, the heterogeneous period
classically defined as sleep stage 1 was detailed in six stages (H3 to H8) characterised by different
EEG patterns (Hori, 1994). Along the falling asleep process, spontaneous EEG activity present a
stage by stage evolution along this nine step scale, but this succession is not strict and is not
uniform across all subjects, as some stages are unstable and do not present in the sequence (i.e.
H4, H6, and H1H2H3 for subjects with suppressed alpha) (Tanaka et al., 1996; Oken et al.,
2006).
The Hori system represents an improvement in the characterisation of the sleep onset
period compared to classical scoring methods in that it permits a refined description, sensible to
microstates. This makes it a potentially useful tool for clinicians. Nonetheless, because it relies on
non stationary and artificial material (short epochs) and on subjective and non systematic pattern
recognition, it is somehow not fully satisfying for a scientific study of the transition in different
states of consciousness.
Quantitative electrophysiological measures of the transitions
Despite the strong bias of “clinical sleep” to define sleep in terms of qualitative eyeball
EEG patterns, EEG spontaneous activity at sleep onset has been frequently studied through
power spectra analysis. This non-stage based approach, based on mathematical tools, meets the
criterion of objectivity and allows a dynamic and continuous description of the changes.
The three major sleep rhythms (Steriade, 1993) are increasingly reflected in EEG activity
while entering sleep: slow oscillations (< 1 Hz), delta, and spindle activity. More precisely, delta
and theta activity increases substantially, while high frequency activities decrease (see table 1). In
particular during the wake-sleep transition, alpha activity shows a dramatic decrease, the peak in
the topography becomes more anterior, and its peak frequency shifts (Tanaka et al., 1997). A rise
in sigma activity, linked to the synaptic generation of sleep spindles in thalamic networks
(Steriade, 1993) is, on the contrary, observed at late stages of the sleep onset process (Gennaro et
al., 2003; Tanaka et al., 1998).
Yet another interesting set of measures arises from the analysis of the frequency bands
coherence, providing rich information on the synchronisation of brain oscillations, and the
7
changing relationships between different regions of the cortex at sleep onset. In particular, these
studies signal a modification of the functional association between frontal and posterior areas
(Ogilvie, 2001). Globally, coherence increases while progressing towards sleep, but shows a
transient decrease around Hori stage H4, that could correspond to the switch between alpha and
theta dominance. While a significant decrease in alpha fronto-occipital and inter-frontal
coherence is observed during drowsiness (Cantero et al., 1999), delta and theta coherence
increase in antero-central areas from H6 (Tanaka, 1999). On the other hand, sigma coherence
emerges in the central area around H8, before spreading to the whole cortex (Tanaka, 1998).
A promising tool for a systematic measurement of EEG spontaneous signals change, and
automatic detections of sleep onset, is the use of non linear mathematical methods (e.g.
correlation dimension, fractal exponents). These measures highlight signal complexity decreases
at sleep onset (Susmakova, 2008, 2006; Acharya, 2005).
Altogether, these EEG measures show that spontaneous brain signal is dramatically
modified while progressing towards sleep. The loss of consciousness and vigilance drop is
paralleled by the transition from a high frequency and complex mode, to a slow oscillatory and
highly synchronized mode.
Alpha
(8 –12 Hz)
Theta
(4 –8 Hz)
Delta
(1 –4 Hz)
Sigma
Slow Fast
(12 –13.5 Hz) (13 –15 Hz)
Relative
power spectra
Topography
Anteriorization
Move from
occipital to
frontal regions
Parietal and
temporal.
Firstly frontal
Gradually
spreads to the
whole scalp
Fronto –central
Centro - parietal
Coherence
Fronto-occipital
& inter-frontal
decrease
Occipito-frontal and centro-parietal
firstly decrease
Then from H6 increase in antero –
central areas
Increase
Increase
Hori 1
Hori 2
Hori 3
Hori 4
Hori 5
Hori 6
Hori 7
Hori 8
Hori 9
Drowsy Wakefulness
Classical Sleep Stage I
Sleep
Stage II
8
Table 1: Spontaneous electroencephalographic modifications at sleep onset. EEG relative
power spectra, topography and coherence evolution along sleep stages for relevant frequency
bands. Sources: Hori et al. 1985, 1994 ; Tanaka et al. 1997, 1999, 1998 ; Morikawa 1997 ; Cantero
et al.1999 ; Ogilvie 2001 ; Badia et al. 1994 ; Oken et al. 2006 ; Susmakova et al. 2008.
Evoked EEG potentials in the transition
Another way to look at brain signal is to observe its reactions to perturbations, instead of
its spontaneous evolution. This is done in electroencephalography by computing event related
potentials (ERPs), and in particular in the sleep domain ERPs provoked by the presentation of
auditory stimuli (auditory related potentials, ARPs).
Like spontaneous activity, neurophysiological response to stimuli is modified early in the
transition. Thus, ARPs (P1, N1, P2, N2 components) recorded after tones presentations during
the sleep onset period show gradually increasing modifications with the progression towards
sleep. While the amplitude of N1 is gradually attenuated, the amplitude of P1 and P2 increase
(De Lugt et al., 1996; Ogilvie, 2001; Noldy et al., 1988). De Lugt and collaborators proposed that
these modifications result from the attenuation of a negative slow wave with the progression
towards sleep (alternatively it could be that a slow positive wave is added to the waveform). This
slow wave could reflect differential mode of processing depending on vigilance states. Results
concerning the N2 component are less clear: some studies found increased amplitude as sleep
progresses (Ornitz et al., 1967; Ogilvie, 2001; Harsh et al., 1994), but others didn’t find
modifications of this component at sleep onset (De Lugt et al., 1996; Noldy et al., 1988). Later
components (N3, P3) on the contrary show consistent and dramatic increases with sleep
progression (Ogilvie, 2001). Furthermore, the end of the transitional period sees the emergence
of specific components (N350, N550) that are related to transient sleep patterns such as the
evoked K complex characteristics of sleep stage 2 (Harsh et al., 1994; Atienza et al., 2001).
A few studies, using classical oddball paradigms during the sleep onset period, have
focused on the P300, a neurophysiological signature of the detection of infrequent stimuli. They
found that with sleep progression, P300 amplitude is gradually reduced, and that this reduction is
tightly linked to behavioural responsiveness and target detection (Cote et al., 2000, Harsh et al.,
1994). Studies that focused on the MMN, that is supposed to reflect automatic change detection
mechanisms, show a gradual diminution of the MMN with sleep progression (Nittono et al.,
9
2001; Sabri et al., 2003). According to Nittono and collaborators, the MMN vanishes around
Hori stage 5 (Nittono et al., 2001). However, other studies still find a MMN in response to large
deviants throughout sleep stage 1, and even in sleep stage 2 (Atienza et al., 2002; Sabri et al.,
2003; Ruby et al., 2008). These differences probably result from methodological divergences (rate
of stimuli presentation, intensity of the tones, sleep stages definitions, computation of the MMN,
confusions with the N1 component of the ARP...). These studies nonetheless suggest a gradual
attenuation of change detection systems while subjects fall asleep. Whether this impairment is
linked to the attenuation of environmental signals at the thalamic level, or to further cognitive
steps of stimuli processing (e.g. sensory memory, comparison mechanisms...) is far from clear.
Hemodynamic activity during the transition
Passive transitions with spatial neuroimaging
In light sleep stage 2 (regarded as the stage of true loss of self-conscious awareness) there
are several areas showing decreased activity, as measured by BOLD signal, when compared to
rested wakefulness. Thalamus and hypothalamus, cingulate cortex, right insula and adjacent
regions of the temporal lobe, the inferior parietal lobule and the inferior/middle frontal gyri, all
showed deactivation in stage 2 (Kaufmann et al., 2006). Interestingly, in sleep stage 1 there were
commonalities to stage 2 but when compared directly (direct contrast) S1 showed more activity
than S2 in the middle and MedFG, supramarginal gyrus, superior temporal gyri, cingulate cortex,
supplementary motor area and paracentral lobule. In contrast, S2 showed more activity in the
cerebellum, the parahippocampal gyrus and the hippocampus. Although difficult to interpret
directly, these activations and deactivation seem partially replicable (Kajimura et al., 1999), S1
(Kjaer et al., 2002), S2 (Andersson et al., 1998; Maquet et al., 1992, 1997), and they may become
more powerful when compared to TMS-EEG and intracranial studies of transitions.
In another series of studies the relationship between EEG power bands and localized
brain activity started to be established. Specifically, delta activity covaried negatively with rCBF
(PET) in the thalamus and brainstem reticular formation, cerebellum, anterior cingulate, and
orbitofrontal cortex (Hofle et al., 1997). In the second analysis, when the effect of delta was
removed, a significant negative covariation between spindle activity and the residual rCBF was
evident in the medial thalamus. In a study using EEG-fMRI Laufs and collaborators (Laufs et al.,
2007) found a negative correlation in alpha band with BOLD changes in the precuneus,
10
prefrontal, and temporal-parietal cortices. Thalamus activation was correlated with sigma power
and negatively with central alpha power. The authors tend to interpret these covariations with
the active inhibition of thalamocortical relay neurons in association with slow wave rhythms and
spindles, as well as the neural substrates underlying the progressive attenuation of sensory
awareness, motor responsiveness, and arousal that occur during slow wave sleep. We extend this
interpretation to current frameworks of integrative processing by suggesting that dissociation of
networks during the transition, with decreasing frontoparietal activity from awake to S1 to S2 fits
nicely with the hypothesis supporting loss of integration in the attentional network as we fall
sleep.
More support for the thalamocortical role in information integration and sleep
transitions comes from the analysis of the local and long-distance functional connectivity. While
from awake to light sleep the thalamocortical connectivity brusquely decreases, the corticortical
interaction increases (Spoormaker et al., 2010). Moreover, local processing seemed to be almost
random in light sleep while slow wave sleep correlated with a high degree of local clustering.
Active transitions with spatial neuroimaging
Single neuron studies have frequently reported a decrease in activity in primary cortices
in response to stimuli during SWS in mammals (Issa and Wang, 2008). Nevertheless, the results
are mixed for the auditory modality, activity was weaker during sleep than during wakefulness
(Murata & Kameda, 1963; Brugge & Merzenich, 1973), but some found the opposite effect, with
even increased activity during sleep (Peña et al., 1999; Edeline et al., 2001). These findings are
supported by an early fMRI imaging were auditory cortex activation did not change between
sleep and wakefulness when presented with by simple tones and the subject's name (Portas et al.,
2000). On the other hand, others have shown activation in the auditory regions of the temporal
lobe decreased during sleep (Czisch et al., 2004), more specifically, BOLD decreases were higher
during stage 2 but vanished in slow wave sleep. Deactivations in auditory areas were associated
with increased number of K-complexes and delta power (Czisch et al., 2004). Following these
results Issa and Wong (Issa and Wong, 2008) found, in monkeys, that modulation was
heterogeneous across neurons such that responses in sleep could be enhanced, suppressed, or
unchanged compared with wakefulness. This variability in cortical effects in primary and
secondary cortices founds its counterpart in EEG and behavioural responses in human (see
11
above). One possible interpretation is that BOLD correlates more with synaptic activity than
spiking activity measured extracellularly (Logothetis, 2008) so if imaging signals are dominated by
the depressed input coming from thalamus in light sleep, they may not reflect active cortical
processing during sleep.
Another interesting result arises from the study of sleep spindles. These may be a key
component in information processing during Sleep stage 2 (Schabus et al., 2007), thought to be
associated with the alteration of consciousness in this stage. Spindles are 11 to 15Hz oscillations
lasting for 0.4 - 3.5 seconds, arising from rebound spike bursts in thalamocortical neurons
postinhibition by reticulothalamic neurons (Steriade & McCarley, 2005). While together the
spindles activate the thalami, anterior cingulate cortex, insular cortex and superior temporal gyri,
when differentiating the slow and fast spindles, slow spindles seem to involve activity in the
superior frontal gyrus, while fast spindles recruited a set of cortical regions involved in
sensorimotor processing, as well as the mesial frontal cortex and hippocampus.
Intracranial recordings and TMS induced sleep waves
To complement the scalp EEG findings on the process of falling asleep a recent study
showed that the thalamus precedes the cortex by several minutes in the progressive deactivation
leading to deep sleep (Magnin et al., 2010). The authors use a measure called dimension of
activation (Shen et al., 2003), this technique provides an index of EEG signal complexity, which
is higher in wakefulness than during slow-wave sleep and independent of the amplitude of the
signal. DA permits the temporal analysis of activation changes in each electrode. In the original
studies in lab animals Steriade and collaborators (Steriade et al., 1993) did not find an earlier area
originating the development of slow wave rhythms but they suggested this rhythm starts in
mesopontine nuclei or locus coeruleus neurons with a relay in the thalamus which spreads the
message to the cortex. Magnin and collaborators add the temporal dimension to this mechanism
by elegantly showing the early deactivation of the thalamus, as measured by local field potentials
in humans, is followed by different areas of the cortex. This effect of earlier sleep onset by the
thalamus seems to be true for the initial microstages from awake to early onset of the Sleep Stage
2, while changes in DA from stage 2 to 3 or 4 do not show delays. Another major finding is that
those delays found in transition from wakefulness to light sleep do not appear when the patient
regains consciousness, in the transition from sleep to awake. The transitions are not symmetrical.
12
In a key series of studies Tononi and collaborators used a perturbational approach to
evaluate the integration capacity of the brain in different sleep stages (Massimini et al., 2009).
They used a combined TMS-EEG technique and applied faint TMS pulses to the frontal cortex.
This allowed them to show that the brain wave elicited by the pulse travelled to distant areas
during wakefulness (even crossing hemispheres) but remained local in Stage 2. In stage 1 there
was a middle ground with partial cortical neuronal activity, processing of sensory inputs and
sometimes displays patterns of synchronous activity (Massimini et al., 2005). According to
Tononi’s model what happens as consciousness fades is that the brain looses integration and
differentiation, and breaks down in modules, local cortical processing, or generates a common
nonspecific response in the form of a (sleep) slow wave (see chapter XXX in this volume).
A recent experiment by this same group of researchers, (Vyazovskiy et al., 2011) showed that in
freely behaving rats awaken for a long period, cortical neurons can briefly and locally go “off-
line” (stop firing altogether and display high-amplitude slow waves) as in sleep. These striking
results suggest that the process of falling asleep is even more widespread in time and space than
previously thought, as independent groups of neuron can already exhibit a sleep like pattern in a
fully awake state. Experiments using a combination of high density EEG, intracranial recordings,
behavioural and responsiveness measures are needed to extend the results to humans and further
characterize these “local sleep” events.
Behavioural dynamics of the process of falling asleep
At sleep onset, subjects progressively lose the ability to respond to external stimuli. The
complete loss of responsiveness constitutes a behavioural sleep onset, assimilated by a few
authors to the loss of awareness.
Behavioural sleep onset has been classically studied through active or passive paradigms.
In active paradigms, subjects must press a button after an auditory cue. Behavioural sleep onset
is then achieved when a consistent absence of responses to environmental stimuli is observed. In
passive paradigms, subjects must exert a continuous pressure on a button. Behavioural sleep
onset then corresponds to a release of the pressure. These two methods correlate, but according
to Ogilvie (1985, Phd Thesis), the loss of responsiveness measured by passive method is more
sensible and occur before the one measured by active methods. This could be linked to the
awakening power of the tones commonly presented in active paradigms.
13
Studies using the active paradigm point out a gradual augmentation of reaction times
with the progression towards sleep (Oglivie & Wilkinson, 1989; Hori et al., 1994). This
slowdown is observed from early stages of drowsiness, and carries on after the first sleep
spindles where responses can still be observed. Hori and collaborators showed that a clear linear
relationship exists between their 9 EEG stages and the slowing of reaction times, validating the
relevance of their graphoelements based classifications in describing the sleep onset period.
In parallel to this slowing, responses become progressively intermittent (Makeig et al.,
1995, 2000; Ogilvie 1989). Subjects start to miss targets from sleep stage 1 (Ogilvie and
Wilkinson, 1989) and its associated EEG spectral power changes described above: delta, theta
bands and sigma power increases (Makeig et al., 1995; Oken et al., 2006). Moreover,
performances in more complex tasks (e.g. deviation from baseline in video games) are less and
less accurate (Makeig et al., 2000).
What is responsible for this gradual loss of responsiveness to external events? Are
responses already limited at early stages by a decreased ability of environmental stimuli
perception? Or are the following steps of response production, i.e. ability to decide, ability to
program and produce a response, impaired? The truth probably resides in between, with a
progressive impairment of each step of cognitive processing while falling asleep.
Thereby, environmental stimuli perception seems gradually impaired at sleep onset. For
instance, a decrease in sensory thresholds along sleep stages progression has been early
documented (Bonnet et al., 1986). The thalamic deactivation at sleep onset has been involved in
a progressive reduction of external information’s transmission to the cortex, leading to a
decreased perception of the environment.
Studies on the P300 provide interesting information on that issue (Harsh et al., 1994,
Cote et al., 2000). The P300 evoked by a deviant stimulus embedded in a train of standard,
decreases as the sleep onset period progress and responsiveness to stimuli decrease. Moreover,
its amplitude is related to the success in target detection. For instance, Cote et al. (2000)
observed that during sleep stage 1 the P300 was absent when subjects failed to detect the target,
and present even if attenuated at frontal sites when he was responding to the target. Behavioural
responses were thus linked to the presence or absence of this change detection neural signature.
This result points out a gradual disappearance of change detection systems at sleep onset.
14
Higher cognitive processes (decision making, motor planning...) are probably also
impaired at sleep onset, involving a gradual loss of the capability to respond. This would be
coherent with the reduced brain activity in structures involved in cognitive performances and
action preparation (prefrontal cortex, locus coeruleus...) (Kaufman et al., 2006; Magnin et al.,
2010; Usher et al., 1999). However, studies combining behavioural measures and imaging
methods during the sleep onset period still lack to document this issue. Nonetheless, sleep
deprivation studies combining PET and behavioural assessment already suggested a tight link
between sleep propensity, decreased activity in thalamic and cortical areas (prefrontal and
parietal), and reduced alertness and cognitive performances (Thomas et al., 2000). Conducting
studies of this kind during the sleep onset period may be the key to disentangle the progressive
disengagement of the different levels of cognitive processing at sleep onset.
Phenomenology of falling asleep
In parallel to the extensively studied physiological and behavioural modifications, one’s
phenomenology is dramatically modified while falling asleep. Although in every day’s life, one’s
subjective feeling is the main criterion for wakefulness judgement, sleep onset period
phenomenology, apart from the very famous hypnagogic phenomenon, received relatively little
scientific attention. The behavioural attitude (eyes closed, lying down, retreat into oneself...) and
physiological transition into a different vigilance state impact on subject’s mind. During the
falling asleep process, subjects pass through successive intermediate conscious states.
In a pioneer study associating polysomnography (PSG) and subjective reports, Foulkes
and Vogel (1965) asked the participants "What were you experiencing just before I called you?",
while they were falling asleep. Further questions allowed the investigation of the perceived state
of vigilance and the quality and characteristics of their phenomenological experience. 212 reports
collected at different junctures of the sleep onset period showed that physiological (eye
movements patterns) and neurophysiological (alpha rhythm propensity) changes observed at
sleep onset were paralleled by mental activity modifications. With the progression towards sleep,
subject’s reports became higher in fantasy, while voluntary control of thoughts was decreasing.
A handful of studies based on a combination between prompted (generally by an
auditory cue) subjects reports and PSG recordings followed this first attempt to characterize
15
sleep onset period phenomenology (Gibson et al., 1982; Hori et al., 1994; Tanaka et al., 1997;
Wackermann et al., 2002; Yang et al., 2010, Lehmann et al., 1995), permitting now the
integration of the psychological findings, physiological and behavioural data.
Hypnagogic Imagery
Probably the most striking phenomenological feature of SO is the emergence of hypnagogic
imagery. These dreamlike hallucinations are symptomatic of the falling asleep period to such an
extent that authors often refer to SO as the hypnagogic period.
Dreamlike mentation can be reported even before the classical onset of sleep stage I, and
until a few minutes after the onset of sleep stage II (Hori 1985, 1982, 1994, Foulkes 1965). Hori
classification permits a better survey of hypnagogic activity in time: the percentage of hypnagogic
reports is closely related to the 9 Hori sleep stages. This relationship is a U-shape function with a
peak at H5 (Hori 1994).
Hypnagogic imagery is often compared to dream imagery. These two phenomenon share
various common features, like the presence of the different sensory modalities, the intrusive
character of images (uncontrolled irruption into consciousness stream), their bizarreness, the
abundance of symbols (Schacter, 1976) ... Like dreams, hypnagogic images are mostly visuals.
Auditory sensations are also present to a lesser extent (noises, music, speech...), as well as
kinaesthetic feelings (falling sensations...) (Schacter, 1976; Foulkes and Vogel, 1965; Hori, 1994).
As we previously discussed, during the hypnagogic period the thalamo-cortical network is
progressively deactivated. Moreover, there is a decoupling between thalamic and cortical
deactivations (Magnin et al., 2010). This neurophysiological pattern (partial cortical activation
and thalamic deactivation) could favour the development of hallucinatory images due to isolation
of the brain from the environment and an alteration of cortical functionality, in a similar way that
during REM sleep, dreams would be facilitated by a cortical reactivation despite preserved
thalamic blockade.
Certain cortical regions, notably primary and secondary sensory areas (occipital visual areas,
temporal auditory areas), remain active during the sleep onset period. Moreover, Hofle et al.
(1997) showed that delta activity, which diminishes from frontal to posterior areas, is positively
16
correlated with rCBF in visual and auditory cortices. Thus, delta activity development in various
cortical areas could correspond to the emergence of sensory images.
Furthermore, the sensory modalities presented in hypnagogic imagery follow a sequential
pattern that seems coherent with neurophysiological data. Thereby, while kinaesthetic images
peak in H1, visual images are maximally reported in H5 or H6 (Nielsen & Germain, 2001; Hori
1994). Nielsen and Germain (2001) also showed that kinaesthetic images are associated with
frontal delta power, and visual images with left central and temporal delta activity. Thus, this data
suggest a strong association between delta expansion from frontal to posterior regions, and
temporal hypnagogic imagery sensory features.
Loss of external world awareness
In parallel to hypnagogic imagery, subjects while falling asleep report a gradual loss of
awareness of their surrounding environment. This effect is exhibited early in the sleep onset
period, from the apparition of SEMs, and ends with a total loss of external world awareness after
immersion in sleep stage 2 (Yang et al., 2010; Foulkes & Vogel, 1965; Wackermann et al., 2002;
Gibson et al.,1982).
This decrease of environmental awareness obviously relates to the loss of responsiveness
observed at the behavioural level (see above), and probably reinforces the immersion into
hypnagogic images. It could result from the decreased thalamic activity at sleep onset (Kaufman
et al., 2006; Hofle et al., 1997).
Disappearing control of thoughts
While progressing towards sleep, subjects also report a gradual loss of control over their
thoughts. The appearance of hypnagogic images seems spontaneous, the coherence, temporal
continuity and logic of thoughts are altered. This deterioration process is already present early in
the sleep onset period (from the first SEMs) but dramatically increases with the first sleep
spindles (Yang, 2010).
17
Various neurophysiological variables could be related to this cognitive impairment, for
instance the decrease observed in the orbito-frontal cortex at the entrance in sleep stage II
(Kaufman et al., 2006; Maquet, 1997; Hofle et al., 1997).
Loss of reality orientation and of time perception
In addition a loss of “reality-orientation” or ability to distinguish internally generated
images from external perceptions is observed with the progression towards sleep (Foulkes &
Vogel, 1965; Yang et al., 2010). Time perception seems also impaired. Indeed, subjects falling
asleep are poor at estimating sleep latencies, or the time elapsed between two interactions with
the experimenter (Wackermann et al., 2002 ; Ogilvie 2001 ; Gibson et al., 1982). These two
phenomena, also characteristic of dreams, could be a result of the combination between loss of
external awareness and disappearing control of thoughts at sleep onset.
Sleep Perception
It is commonly admitted that a consistent perception of sleep is found only a few
minutes after classical sleep stage II onset (Schacter, 1976 ; Davis et al., 1937 ; Yang et al., 2010 ;
Ogilvie 2001...). Before that final breakdown of vigilance, sleepiness and asleep / awake self
reports increase gradually as the sleep onset period progresses (Oken et al., 2006 ; Wackermann
et al., 2002 ; Hori et al., 1994 ; Yang et al., 2010). For instance, a close relationship exists between
asleep/awake reports and the 9 Hori EEG stages (Hori et al., 1994).
But on which subjective basis are people judging that they are asleep or not? Which
characteristics of their phenomenological experience do they use? Various characteristics of the
sleep onset period phenomenology have been involved in sleep perception. For instance, Gibson
el al. (1982) showed that correct subjective estimations of sleep stages correlated mainly with
three phenomenological variables: external awareness, temporal awareness and the control over
thoughts... According to a recent study by Yang et al. (2010), the loss of control over thinking
processes would be the determinant subjective variable for sleep perception.
Discussion
18
How do we fall asleep? How do we regain consciousness? These are central questions to
the understanding of consciousness and nevertheless bare attempts have been done to address
them in depth from a cognitive point of view. Several studies have been done on drowsiness
(Makeig et al., 1996; Huang et al., 2009; Kar et al., 2010) due to the importance of preventing
traffic and work-related accidents. These studies look primarily at the errors made by the
participants but do not frame the results in key questions for the study of consciousness: When
do we lose the capacity to consciously respond? When do we lose the capacity to consciously
take a decision? What sleep elements predict the changes in cognitive processes during the
transition?
The first line of enquiries about transitions was started by Hori and collaborators and
continued by Ogilvie et al in the 80’. Both teams recognised the lack of detail in determining the
transitions and looked to characterise sleep onset with different EEG and behavioural measures.
The definition of the sleep EEG grapho-elements every 5 seconds instead of 30 seconds (Hori et
al.) in the transition from wake to sleep broke with the classic assumption of the sleep medicine
establishment that the conscious state is stationary during those 30 seconds (Retchaffen & Kale,
1968). This paved the way for more research challenging the criterion for wakefulness (and the
loss of wakefulness). Ogilvie stated in the 1988 and 1989 that “if the criterion for wakefulness is
cognitive response to external stimulation, only in EEG Stages 3, 4, and REM can accurate
distinctions between sleep and wakefulness be made. If EEG is the criterion, then the data
suggest that cognitive response is possible during Stages 1 and 2 sleep.” They equated motor
response to simple tones to being awake and found that the probability of response dramatically
decreased from stage 1 to stage 2 and became close to zero in stage 3 (and 4). We have now
advanced in defining what a conscious response is, or maybe it would be more accurate to say we
have moved the boundary of unconscious responses to even inhibition of a response (van Gaal
et al., 2008; Hughes et al., 2009; Boy et al., 2010).
Under the current frameworks for the study of consciousness a few criteria must be met
for a response to qualify as conscious awareness movement. The problem underlying the
definition of what constitutes a conscious response, arising from a conscious decision, comes
from recent shift in the paradigm of volition and awareness (Page & Baars, 2007). In the last
decade there has been a myriad of investigations showing motor responses without awareness of
the decision (Dijksterhuis & Aarts, 2010) and furthermore actions may be initiated even though
we remain unconscious of the goals or motivation of our behaviour (Custer & Aarts, 2010). This
19
line of thought has moved the awareness line from action -a movement in response to stimuli is
performed- to intention –a brain process causing the action-, and finally to goal –a brain process
guided by belief that sets up the intentions. This theoretical framework does not take into
account those mental processes occurring while consciousness fades into unconsciousness and
therefore they are difficult to frame. Moreover, the batch of behavioural experiments performed
during the transition is too modest to even start to fit them in these models. It might be better to
think the future experiments in line with these complex accounts of consciousness keeping in
mind the limited cognitive capacity of the people when they enter drowsiness.
Another avenue to explore when thinking about transitions is to keep the concept of
awareness at its core (Dehaene and Naccache, 2001) by assuming that we are dealing with
perceptual awareness. This simplifies the problem and allows us to experimentally test whether
specific ERP or fMRI signatures at different points in the transition do show parieto-frontal
activity causing top-down amplification (Dehaene et al.,al., 2006). These specific neural
signatures, if not too affected by the arousal changes, should help in the interpretation as to
when a response (brain response or movement) is conscious, preconscious or unconscious.
Maybe a complementary model is that one that does not rely on the ability to respond or
purposeful behaviour as such but in a particular mode of information integration (Tononi, 2004).
This framework relies in two conditions needed to attain consciousness; first, to what extent
different regions of the thalamocortical system can interact causally (integration), and second,
whether they can produce specific responses (information). The authors have been refining the
theory and what and how to measure to further develop these ideas and make them suitable to
be tested by a variety of methods and paradigms (see chapter XXX in this volume).
The way forward to test the limits of conscious processing during the transition
Our proposal to study the transitions of consciousness, both in sleep or sedation studies,
includes behavioral, electrophysiological and brain imaging measures with and without
stimulation and instructions. This multifaceted approach will allow the fields of sleep, sedation
and consciousness to have a common ground. The transitional approach, continuous testing of
behaviour and neurophysiology during the transitions, aims at integrating the behavioural
measures and neuromarkers of the cognitive sciences (perception, awareness, attention) with the
physiological and brain markers of wakefulness.
20
In an archetypical experiment under this experimental framework the normal transition
from wakefulness to SWS would be followed by high density EEG or EEG-fMRI, plus a series
of physiological markers such as EMG, ECG and respiration rate (to complete the
polysomnography). The participant would be assigned to an active or passive task; if active it
may be from a range of executive functions tests going from a simple detection task to a stop
signal task or a crossmodal judgement task. The inclusion of a parametric experimental design on
the wake-sleep transition is a novel conceptual advance in the theoretical treatment of the axis of
consciousness, wakefulness and awareness that should help move the field forward. These
perceptual and cognitive perturbations will allow mapping the dynamics of the different
cognitive processes in behaviour, physiology and neurophysiology. The parsing of the data in
single trial analysis with the combination of prestate/pretrial information, evoked responses and
behavioural performance should permit us to start answering the key questions in transitions of
consciousness (loss of the capacity of detection, discrimination, action, intention, goal,
inhibition). Furthermore, the incorporation of integration measures in EEG and fMRI
(coherence, local and global interactions, connectivity, causality, etc) will complete the picture by
providing different accounts of the brain capacity before the stimulus (predictive of brain or
muscle response), after the stimulus (as evoked neural markers of a cognitive process), and as
covariates or regressors. Finally, the amalgamation of different accounts of sleep changes,
neurophysiology of sedation and experimental theories of consciousness should become
complementary in the quest for characterization of how consciousness fade, how it is regained,
what is the different capacities of the brain during these changes and how can this impact in the
clinical and social aspects of sleep and wakefulness.
21
References
Acharya U.R., Faust O., Kannathal N., Chua T. and Laxminarayan S. Non-linear analysis of
EEG signals at various sleep stages. Comput Methods Programs Biomed, 80: 37-45,
2005.
Andersson J.L., Onoe H., Hetta J., Lidstrom K., Valind S., Lilja A., Sundin A., Fasth K.J.,
Westerberg G., Broman J.E., Watanabe Y. and Langstrom B. Brain networks affected
by synchronized sleep visualized by positron emission tomography. J Cereb Blood
Flow Metab, 18: 701-715, 1998.
Atienza M., Cantero J.L. and Dominguez-Marin E. Mismatch negativity (MMN): an objective
measure of sensory memory and long-lasting memories during sleep. Int J
Psychophysiol, 46: 215-225, 2002.
Atienza M., Cantero J.L. and Escera C. Auditory information processing during human sleep
as revealed by event-related brain potentials. Clin Neurophysiol, 112: 2031-2045,
2001.
Baars B. and Gage N. Cognition, Brain and Consciousness. 2007, London: Elsevier/Academic
Press.
Baharav A., Kotagal S., Gibbons V., Rubin B.K., Pratt G., Karin J. and Akselrod S. Fluctuations
in autonomic nervous activity during sleep displayed by power spectrum analysis of
heart rate variability. Neurology, 45: 1183-1187, 1995.
Berger H. Psyche. 1940, Jena: Gustav Fischer.
Bonnet M.H. Auditory thresholds during continuing sleep. Biol Psychol, 22: 3-10, 1986.
Borbely A.A. A two process model of sleep regulation. Hum Neurobiol, 1: 195-204, 1982.
Borbély A.A., Achermann P., Meir H.K., Thomas R. and William C.D. Sleep Homeostasis and
22
Models of Sleep Regulation. pp. 405-417. Principles and Practice of Sleep Medicine
(Fourth Edition). Philadelphia, W.B. Saunders, 2005.
Boy F., Husain M., Singh K.D. and Sumner P. Supplementary motor area activations in
unconscious inhibition of voluntary action. Exp Brain Res, 206: 441-448, 2010.
Brugge J.F. and Merzenich M.M. Responses of neurons in auditory cortex of the macaque
monkey to monaural and binaural stimulation. J Neurophysiol, 36: 1138-1158, 1973.
Cantero J.L., Atienza M., Salas R.M. and Gomez C.M. Alpha EEG coherence in different brain
states: an electrophysiological index of the arousal level in human subjects. Neurosci
Lett, 271: 167-170, 1999.
Cote K.A., De Lugt D.R. and Campbell K.B. Changes in the scalp topography of event-related
potentials and behavioral responses during the sleep onset period.
Psychophysiology, 39: 29-37, 2002.
Custers R. and Aarts H. The unconscious will: how the pursuit of goals operates outside of
conscious awareness. Science, 329: 47-50, 2010.
Czisch M., Wehrle R., Kaufmann C., Wetter T.C., Holsboer F., Pollmacher T. and Auer D.P.
Functional MRI during sleep: BOLD signal decreases and their electrophysiological
correlates. Eur J Neurosci, 20: 566-574, 2004.
Danker-Hopfe H., Anderer P., Zeitlhofer J., Boeck M., Dorn H., Gruber G., Heller E., Loretz E.,
Moser D., Parapatics S., Saletu B., Schmidt A. and Dorffner G. Interrater reliability for
sleep scoring according to the Rechtschaffen & Kales and the new AASM standard. J
Sleep Res, 18: 74-84, 2009.
Davis H., Davis P.A., Loomis A.L., Harvey E.N. and Hobart G. Changes in Human Brain
Potentials During the Onset of Sleep. Science, 86: 448-450, 1937.
23
De Gennaro L. and Ferrara M. Sleep spindles: an overview. Sleep Med Rev, 7: 423-440, 2003.
De Gennaro L., Ferrara M., Ferlazzo F. and Bertini M. Slow eye movements and EEG power
spectra during wake-sleep transition. Clin Neurophysiol, 111: 2107-2115, 2000.
De Lugt D.R., Loewy D.H. and Campbell K.B. The effect of sleep onset on event related
potentials with rapid rates of stimulus presentation. Electroencephalogr Clin
Neurophysiol, 98: 484-492, 1996.
Dehaene S., Changeux J.P., Naccache L., Sackur J. and Sergent C. Conscious, preconscious,
and subliminal processing: a testable taxonomy. Trends Cogn Sci, 10: 204-211, 2006.
Dehaene S. and Naccache L. Towards a cognitive neuroscience of consciousness: basic
evidence and a workspace framework. Cognition, 79: 1-37, 2001.
Dijksterhuis A. and Aarts H. Goals, attention, and (un)consciousness. Annu Rev Psychol, 61:
467-490, 2010.
Dvornyk V., Vinogradova O. and Nevo E. Origin and evolution of circadian clock genes in
prokaryotes. Proc Natl Acad Sci U S A, 100: 2495-2500, 2003.
Edeline J.M., Manunta Y. and Hennevin E. Auditory thalamus neurons during sleep: changes
in frequency selectivity, threshold, and receptive field size. J Neurophysiol, 84: 934-
952, 2000.
Foulkes D. and Vogel G. Mental Activity at Sleep Onset. J Abnorm Psychol, 70: 231-243,
1965.
Germain A. and Nielsen T.A. EEG Power Associated with Early Sleep Onset Images Differing
in Sensory Content. Sleep Res Online, 4: 83-90, 2001.
Gibson E., Perry F., Redington D. and Kamiya J. Discrimination of sleep onset stages:
behavioral responses and verbal reports. Percept Mot Skills, 55: 1023-1037, 1982.
24
Harsh J., Voss U., Hull J., Schrepfer S. and Badia P. ERP and behavioral changes during the
wake/sleep transition. Psychophysiology, 31: 244-252, 1994.
Hofle N., Paus T., Reutens D., Fiset P., Gotman J., Evans A.C. and Jones B.E. Regional cerebral
blood flow changes as a function of delta and spindle activity during slow wave sleep
in humans. J Neurosci, 17: 4800-4808, 1997.
Hori T. Electrodermal and Electro-oculographic Activity in a Hypnagogic State.
Psychophysiology, 19: 668-672, 1982.
Hori T. Spatiotemporal changes of EEG activity during waking-sleeping transition period. Int J
Neurosci, 27: 101-114, 1985.
Hori T., Hayashi M. and Morikawa T. Topographical EEG changes and the hypnagogic
experience. pp. 237-253. In: Ogilvie R.D. & Harsh J.R. (Eds.) Sleep Onset: Normal and
Abnormal Processes. Washington DC, American Psychological Association, 1994.
Huang R.S., Jung T.P. and Makeig S. Tonic Changes in EEG Power Spectra during Simulated
Driving. San Diego, USA, Proceedings of the 5th International Conference on
Foundations of Augmented Cognition, Springer-Verlag, 2009.
Hughes G., Velmans M. and De Fockert J. Unconscious priming of a no-go response.
Psychophysiology, 46: 1258-1269, 2009.
Issa E.B. and Wang X. Sensory responses during sleep in primate primary and secondary
auditory cortex. J Neurosci, 28: 14467-14480, 2008.
Kajimura N., Uchiyama M., Takayama Y., Uchida S., Uema T., Kato M., Sekimoto M.,
Watanabe T., Nakajima T., Horikoshi S., Ogawa K., Nishikawa M., Hiroki M., Kudo Y.,
Matsuda H., Okawa M. and Takahashi K. Activity of midbrain reticular formation and
neocortex during the progression of human non-rapid eye movement sleep. J
25
Neurosci, 19: 10065-10073, 1999.
Kar S., Routray A. and Nayak B.P. Functional network changes associated with sleep
deprivation and fatigue during simulated driving: Validation using blood biomarkers.
Clin Neurophysiol, 2010.
Kaufmann C., Wehrle R., Wetter T.C., Holsboer F., Auer D.P., Pollmacher T. and Czisch M.
Brain activation and hypothalamic functional connectivity during human non-rapid
eye movement sleep: an EEG/fMRI study. Brain, 129: 655-667, 2006.
Kjaer T.W., Law I., Wiltschiotz G., Paulson O.B. and Madsen P.L. Regional cerebral blood flow
during light sleep--a H(2)(15)O-PET study. J Sleep Res, 11: 201-207, 2002.
Laufs H., Walker M.C. and Lund T.E. '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: 75-76, 2007.
Lehmann D., Grass P. and Meier B. Spontaneous conscious covert cognition states and brain
electric spectral states in canonical correlations. Int J Psychophysiol. 19: 41-52, 1995.
Magnin M., Rey M., Bastuji H., Guillemant P., Mauguiere F. and Garcia-Larrea L. Thalamic
deactivation at sleep onset precedes that of the cerebral cortex in humans.
Proceedings of the National Academy of Sciences, 2010.
Magosso E., Ursino M., Provini F. and Montagna P. Wavelet analysis of
electroencephalographic and electro-oculographic changes during the sleep onset
period. Annual International Conference of the IEEE Engineering in Medicine and
Biology Society, 2007.
Makeig S. and Jung T.P. Changes in alertness are a principal component of variance in the
EEG spectrum. Neuroreport, 7: 213-216, 1995.
26
Makeig S. and Jung T.P. Tonic, phasic, and transient EEG correlates of auditory awareness in
drowsiness. Brain Res Cogn Brain Res, 4: 15-25, 1996.
Makeig S., Jung T.P. and Sejnowski T.J. Awareness during drowsiness: dynamics and
electrophysiological correlates. Can J Exp Psychol, 54: 266-273, 2000.
Maquet P., Degueldre C., Delfiore G., Aerts J., Peters J.M., Luxen A. and Franck G. Functional
neuroanatomy of human slow wave sleep. J Neurosci, 17: 2807-2812, 1997.
Maquet P., Dive D., Salmon E., Sadzot B., Franco G., Poirrier R. and Franck G. Cerebral
glucose utilization during stage 2 sleep in man. Brain Res, 571: 149-153, 1992.
Massimini M., Boly M., Casali A., Rosanova M. and Tononi G. A perturbational approach for
evaluating the brain's capacity for consciousness. Prog Brain Res, 177: 201-214,
2009.
Massimini M., Ferrarelli F., Huber R., Esser S.K., Singh H. and Tononi G. Breakdown of
cortical effective connectivity during sleep. Science, 309: 2228-2232, 2005.
Murata K. and Kameda K. The Activity of Single Cortical Neurones of Unrestrained Cats
During Sleep and Wakefulness. Arch Ital Biol, 101: 306-331, 1963.
Nittono H., Momose D. and Hori T. The vanishing point of the mismatch negativity at sleep
onset. Clin Neurophysiol, 112: 732-739, 2001.
Noldy N., McGarry P. and Campbell K. Late auditory evoked potentials as indicators of sleep
onset. pp. 277-280. In: W K., Obal F., H S. & Visser P. (Eds.) Sleep '86. Stuttgart,
Pontenegal Press, 1988.
Ogilvie R.D. The process of falling asleep. Sleep Med Rev, 5: 247-270, 2001.
Ogilvie R.D., Wilkinson R.T. and Allison S. The detection of sleep onset: behavioral,
physiological, and subjective convergence. Sleep, 12: 458-474, 1989.
27
Oken B.S., Salinsky M.C. and Elsas S.M. Vigilance, alertness, or sustained attention:
physiological basis and measurement. Clin Neurophysiol, 117: 1885-1901, 2006.
Ornitz E.M., Ritvo E.R., Carr E.M., La Franchi S. and Walter R.D. The effect of sleep onset on
the auditory averaged evoked response. Electroencephalogr Clin Neurophysiol, 23:
335-341, 1967.
Pace-Schott E.F. and Hobson J.A. The Neurobiology of Sleep: Genetics, cellular physiology
and subcortical networks. Nat Rev Neurosci, 3: 591-605, 2002.
Pelayo R. and Guilleminault C. History of Sleep Research. pp. 3-6. In: Stickgold R. & Walker
M.P. (Eds.) The neuroscience of sleep. Amsterdam, Boston, Academic Press/Elsevier,
2009.
Pena J.L., Perez-Perera L., Bouvier M. and Velluti R.A. Sleep and wakefulness modulation of
the neuronal firing in the auditory cortex of the guinea pig. Brain Res, 816: 463-470,
1999.
Pivik R.T. and Busby K. Heart rate associated with sleep onset in preadolescents. J Sleep Res,
5: 33-36, 1996.
Portas C.M., Krakow K., Allen P., Josephs O., Armony J.L. and Frith C.D. Auditory processing
across the sleep-wake cycle: simultaneous EEG and fMRI monitoring in humans.
Neuron, 28: 991-999, 2000.
Rechtschaffen A. and Kale A.. A Manual of Standardized Terminology, Techniques, and
Scoring System for Sleep Stages of Human Subjects. U.S. Govt Printing Office,
Washington DC, 1968.
Ruby P., Caclin A., Boulet S., Delpuech C. and Morlet D. Odd sound processing in the
sleeping brain. J Cogn Neurosci, 20: 296-311, 2008.
28
Sabri M., Labelle S., Gosselin A. and Campbell K.B. Effects of sleep onset on the mismatch
negativity (MMN) to frequency deviants using a rapid rate of presentation. Brain Res
Cogn Brain Res, 17: 164-176, 2003.
Schabus M., Dang-Vu T.T., 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. and Maquet P. Hemodynamic cerebral correlates of sleep
spindles during human non-rapid eye movement sleep. Proc Natl Acad Sci U S A, 104:
13164-13169, 2007.
Schacter D.L. The hypnagogic state: a critical review of the literature. Psychol Bull, 83: 452-
481, 1976.
Schwartz W.J. Circadian oscillations in the suprachiasmatic nucleus. pp. 939-944. In: Squire
L.R. (Ed.), Encyclopedia of Neuroscience. Oxford, Academic Press, 2009.
Shen Y., Olbrich E., Achermann P. and Meier P.F. Dimensional complexity and spectral
properties of the human sleep EEG. Electroencephalograms. Clin Neurophysiol, 114:
199-209, 2003.
Silber M.H., Ancoli-Israel S., Bonnet M.H., Chokroverty S., Grigg-Damberger M.M.,
Hirshkowitz M., Kapen S., Keenan S.A., Kryger M.H., Penzel T., Pressman M.R. and
Iber C. The visual scoring of sleep in adults. J Clin Sleep Med, 3: 121-131, 2007.
Spoormaker V.I., Schroter M.S., Gleiser P.M., Andrade K.C., Dresler M., Wehrle R., Samann
P.G. and Czisch M. Development of a large-scale functional brain network during
human non-rapid eye movement sleep. J Neurosci, 30: 11379-11387, 2010.
Steriade M., Contreras D., Curro Dossi R. and Nunez A. The slow (< 1 Hz) oscillation in
reticular thalamic and thalamocortical neurons: scenario of sleep rhythm generation
29
in interacting thalamic and neocortical networks. J Neurosci, 13: 3284-3299, 1993.
Steriade M. and W. M.R. Brain Control of Wakefulness and Sleep. 2005, New York: Kluwer
Academic.
Susmakova K. Correlation Dimension versus Fractal Exponent During Sleep Onset.
Measurement Sci Rev, 6: 58-64, 2006.
Susmakova K. and Krakovska A. Discrimination ability of individual measures used in sleep
stages classification. Artif. Intell. Med., 44: 261-277, 2008.
Tanaka H., Hayashi M. and Hori T. Statistical features of hypnagogic EEG measured by a new
scoring system. Sleep, 19: 731-738, 1996.
Tanaka H., Hayashi M. and Hori T. Topographical characteristics and principal component
structure of the hypnagogic EEG. Sleep, 20: 523-534, 1997.
Tanaka H., Hayashi M. and Hori T. Topographic mapping of electroencephalography
coherence in hypnagogic state. Psychiatry Clin Neurosci, 52: 147-148, 1998.
Tanaka H., Hayashi M. and Hori T. Topographic mapping of EEG spectral power and
coherence in delta activity during the transition from wakefulness to sleep.
Psychiatry Clin Neurosci, 53: 155-157, 1999.
Thomas M., Sing H., Belenky G., Holcomb H., Mayberg H., Dannals R., Wagner H., Thorne D.,
Popp K., Rowland L., Welsh A., Balwinski S. and Redmond D. Neural basis of alertness
and cognitive performance impairments during sleepiness. I. Effects of 24 h of sleep
deprivation on waking human regional brain activity. J Sleep Res, 9: 335-352, 2000.
Tononi G. An information integration theory of consciousness. BMC neuroscience, 5: 42,
2004.
Usher M., Cohen J.D., Servan-Schreiber D., Rajkowski J. and Aston-Jones G. The role of locus
30
coeruleus in the regulation of cognitive performance. Science, 283: 549-554, 1999.
Van den Heuvel C.J., Noone J.T., Lushington K. and Dawson D. Changes in sleepiness and
body temperature precede nocturnal sleep onset: evidence from a
polysomnographic study in young men. J Sleep Res, 7: 159-166, 1998.
Van Gaal S., Ridderinkhof K.R., Fahrenfort J.J., Scholte H.S. and Lamme V.A. Frontal cortex
mediates unconsciously triggered inhibitory control. J Neurosci, 28: 8053-8062, 2008.
Vyazovskiy V.V.,Olcese U., Hanlon E.C., Nir Y., Cirelli C. and Tononi G. Local sleep in awake
rats. Nature, 472: 443–447, 2011.
Wackermann J., Putz P., Buchi S., Strauch I. and Lehmann D. Brain electrical activity and
subjective experience during altered states of consciousness: ganzfeld and
hypnagogic states. Int J Psychophysiol, 46: 123-146, 2002.
Worsnop C., Kay A., Pierce R., Kim Y. and Trinder J. Activity of respiratory pump and upper
airway muscles during sleep onset. J Appl Physiol, 85: 908-920, 1998.
Yang C.M., Han H.Y., Yang M.H., Su W.C. and Lane T. What subjective experiences determine
the perception of falling asleep during sleep onset period? Conscious Cogn, 19: 1084-
1092, 2010.
Zimmerman J.E., Naidoo N., Raizen D.M. and Pack A.I. Conservation of sleep: insights from
non-mammalian model systems. Trends Neurosci, 31: 371-376, 2008.
