www.elsevier.com/locate/brainres

Brain Research 1050

Research Report

Theta activity in the waking EEG is a marker of sleep propensity in the rat

Vladyslav V. Vyazovskiy1, Irene Tobler*

Institute of Pharmacology and Toxicology, University of Zurich, Winterthurerstr., 190 CH-8057 Zurich, Switzerland

Accepted 10 May 2005

Available online 22 June 2005

Abstract

In humans, EEG power in the theta frequency band (5–8 Hz) during quiet waking increases during sleep deprivation (SD), and predicts

the subsequent homeostatic increase of sleep slow-wave activity (SWA; EEG power between 0.5 and 4.0 Hz). These findings indicate that

theta power in waking is an EEG variable, which reflects the rise in sleep propensity. In rodents, a number of short sleep attempts, as well as

SWA in the waking EEG increase in the course of SD, but neither variable predicts the subsequent homeostatic increase of EEG SWA during

recovery sleep. To investigate whether there is an EEG marker for sleep propensity also in rodents, the EEG of the rat was recorded during 6

h SD in the first half of the light period (SDL, n = 7). During SDL, power of the waking EEG showed an increase in the delta (1.5–4 Hz) and

low theta (5–6.5 Hz) band. Based on the neck muscle EMG, wakefulness was subdivided into active (high EMG activity) and quiet (low

EMG activity) waking. During quiet waking, the theta peak occurred at 5.5 Hz, the frequency at which the increase of EEG power during SD

was most pronounced. This increase was due to higher amplitude of theta waves, while wave incidence (frequency) was unchanged.

Correlation analysis showed that the rise in EEG power in the 5–7 Hz band during SD predicted the subsequent enhancement of SWA in

non-rapid eye movement sleep. The analysis of data of a further batch of rats which were sleep deprived for 6 h after dark onset (SDD, n = 7)

revealed a significant increase in theta-wave amplitude during the SD and a tendency for a similar, positive correlation between the increase

of theta power (5–7 Hz) and subsequent SWA. The results indicate that in rats, as in humans, a specific waking EEG frequency, i.e., theta

power in quiet waking is a marker of sleep propensity.

D 2005 Elsevier B.V. All rights reserved.

Theme: Neural basis of behavior

Topic: Biological rhythms and sleep

Keywords: Theta EEG activity; Waking EEG; Sleep homeostasis; Sleep regulation; EMG; Behavior

1. Introduction

Sleep propensity rises progressively during waking and

declines monotonically during sleep. The level of EEG slow-

wave activity (SWA; EEG power between 0.5 and 4.0 Hz)

increases proportionally to the duration of prior wakefulness,

0006-8993/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.brainres.2005.05.022

Abbreviations: ANOVA, analysis of variance; EEG, electroencephalo-

gram; EMG, electromyogram; NREM, non-rapid eye movement; SEM,

standard error of the mean; SD, sleep deprivation; SDL, sleep deprivation in

the light period; SDD, sleep deprivation in the dark period; SWA, slow-

wave activity

* Corresponding author. Fax: +41 01 635 57 07.

E-mail address: tobler@pharma.unizh.ch (I. Tobler).1 Present address: University of Wisconsin-Madison, Psychiatric Institute

and Clinics, 6001 Research Park Blvd., Madison, WI 53719, USA.

and thus reflects the homeostatic aspect of sleep regulation

[5,11,26]. This fundamental property of sleep has been

demonstrated for a wide variety of different species [6,25].

In the two-process model of sleep regulation, the time

course of the rise of the homeostatic Process S was derived

from the level of SWA in non-rapid eye movement (NREM)

sleep preceding and following a waking episode [1,5]. In

humans, a marker of Process S was identified also in the

waking EEG. Thus, an increase of delta or theta power

occurred in the course of wakefulness [2,3,9,12,22,28].

Further evidence for a link between theta activity in the

waking EEG and sleep homeostasis was provided by a

between-subject correlative study [12]. The rate of increase

of theta power during 40 h of waking correlated positively

with the increase of SWA in NREM sleep. Moreover, theta

(2005) 64 – 71

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–71 65

activity in the waking EEG paralleled subjective sleepiness

[2,12,15,20,21,24].

In animals, an EEG correlate for sleep pressure during

waking has not been reported. Rodents exhibit a variety of

waking behaviors including active exploration and quies-

cence with low locomotion. These differences in behavior

are reflected in the EEG [29]. During active waking,

associated with voluntary movements, the EEG is

desynchronized, dominated by fast frequencies and regular

theta activity, while quiet waking is characterized by a

mixed pattern with slower waves [14,16,29].

The changes in arousal level and EEG activity are

interrelated because both depend on the tonic depolarizing

input from the brainstem [18,23]. Several neurotransmitters

play a major role in the ‘‘arousal brain system’’ [18]. The

neurons comprising the ascending reticular activating sys-

tem, as well as thalamocortical neurons, use glutamate as a

transmitter; pontomesencephalic neurons use acetylcholine,

while locus coeruleus neurons, projecting in a diffuse manner

from the brainstem to the entire forebrain, are adrenergic [18].

Activation of these nuclei results in the blockage of

synchronized low-frequency oscillations in the thalamocort-

ical system that are typical for NREM sleep, and in the

generation of fast activity, typical for wakefulness [23].

We hypothesized that an increase of sleep propensity

could be manifested in the EEG during quiet wakefulness

also in rodents in the course of prolonged wakefulness. To

test this hypothesis, we subdivided the waking epochs during

a 6-h sleep deprivation (SD) into quiet waking and active

waking on the basis of EMG activity, analyzed the changes in

the power spectrum, and investigated whether the changes

were related to the increase of SWA during recovery sleep.

2. Materials and methods

2.1. Animals

The local governmental commission for animal research

approved the experiments. Adult male albino rats of the

Sprague–Dawley strain (n = 14) with a mean body weight

277.5 T 5.1 (SEM) g were used. The animals were kept

individually in Macrolon cages (53� 34� 37 cm) with food

and water available ad libitum, and maintained on a 12-h

light–12-h dark cycle (light from 8.00 to 20.00 h; 7 W

OSRAM Dulux EL energy saving lamp, approximately 30

lx). Ambient temperature wasmaintained at 21–22 -C.Underdeep pentobarbital anesthesia (Nembutal sodium, 80 mg/kg

i.p., volume approximately 0.5 ml), the rats were implanted

with gold-plated miniature screws (0.9 mm diameter) that

served as EEG electrodes. The right parietal electrode was

implanted 5.5 mm lateral to the midline, 2.5 mm posterior to

the bregma, and referenced to the electrode above the

cerebellum (2 mm posterior to the lambda, on midline).

Two gold wires (diameter 0.2 mm) inserted into the

neck muscles served to record the electromyogram (EMG).

The electrodes were connected to stainless steel wires that

were fixed to the skull with dental cement. At least 10 days

were allowed for recovery. The rats were connected to the

cable and amplifier via a swivel one day before the

baseline was recorded.

2.2. Experimental protocol and data acquisition

The EEG and the EMG were recorded continuously. A

24-h baseline day was followed by 6 h sleep deprivation

(SD) starting at light onset (8:00 a.m., SDL, n = 7) and 18

h recovery. To investigate whether the timing of the

intervention would play a role, a second batch of rats was

sleep deprived for 6 h beginning at dark onset (8:00 p.m.,

SDD, n = 7). SD was performed by introducing a variety

of objects (e.g., nesting material, pieces of wood, paper,

and tissue) into the cage, and by tapping on the cage

whenever the animal appeared drowsy or the EEG

exhibited slow waves. Halfway through the SD, the cages

were exchanged to provide additional stimulation [17,30].

Due to the increased occurrence of EEG artifacts during

this manipulation, the second 2-h interval of SD (hours 3–

4) was excluded from the analysis. The rats were never

disturbed when they were spontaneously awake, or during

feeding and drinking.

The EEG and the EMG signals were amplified

(amplification factor approximately 2000), conditioned by

analog filters (high-pass filter: �3 dB at 0.016 Hz; low-

pass filter: �3 dB at 40 Hz, less than �35 dB at 128 Hz)

sampled with 512 Hz, digitally filtered (EEG: low-pass

FIR filter 25 Hz; EMG: band-pass FIR filter 20–50 Hz)

and stored with a resolution of 128 Hz. EEG power spectra

were computed for 2-s epochs by a Fast Fourier Transform

(FFT) routine (linear detrending, Hanning window, 0.5 Hz

resolution). Frequencies between 0.5 and 25 Hz were

analyzed. Period–amplitude analysis was performed on the

5–7 frequency band. The raw EEG signal was band-pass

filtered (5–7 Hz, 4th order Chebyshev type II filter,

forward and reverse filtering), and the amplitude of

positive peaks of individual theta waves, and their number

per second (frequency) were computed.

2.3. Vigilance states

Two vigilance states, NREM sleep and waking were

scored for 4-s epochs. They were determined off-line by

visual inspection of the parietal EEG and the EMG records

and EEG power in the slow-wave range (0.75–4.0 Hz).

Special care was taken to exclude from the waking EEG

those epochs during SD where clear signs of NREM sleep

(i.e., delta waves concomitant with low EMG activity)

were evident. Epochs containing EEG artifacts were

excluded from spectral analysis (Table 1). The sleep

deprivation procedure was effective, since only 0.3 T 0.1

(n = 7) and 1.8 T 0.9 min (n = 7) of NREM sleep occurred

during SDL and SDD, respectively.

Table 1

Waking during 6 h sleep deprivation and EMG amplitude

Amount of waking (min) EMG amplitude (AV)

Waking during SD Baseline

Total ‘‘Low EMG’’ ‘‘High EMG’’ ‘‘Low EMG’’ NREMS REMS

SDL (n = 7) 309.8 T 7.3 51.9 T 1.4 43.8 T 5.0 17.9 T 1.9* 6.4 T 0.6** 4.9 T 0.7‘

SDD (n = 7) 289.9 T 9.6 49.5 T 2.0 46.1 T 4.6 18.6 T 1.5* 6.2 T 0.8** 4.8 T 0.6‘

Mean amount of total artifact free waking, and waking with low EMG activity during 6 h SD in the light (SDL, n = 7) and in the dark (SDD, n = 7) period

(TSEM), and mean EMG amplitude during the two SDs and 24-h baseline. Significant differences between EMG amplitude: ‘‘high EMG’’ vs. ‘‘low EMG’’ (*),

NREM sleep vs. ‘‘low EMG’’ epochs (**), NREM sleep vs. REM sleep (‘) (P < 0.01; two-tailed paired t tests).

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–7166

2.4. Wakefulness with low EMG and high EMG

Our aim was to investigate the EEG separately during

active (high EMG activity) and quiet (low EMG activity)

waking. All 2-s epochs of wakefulness during the 6-h SDs

were subdivided on the basis of the log-transformed EMG

amplitude into ‘‘low EMG’’ and ‘‘high EMG’’ epochs. Log

transformation was used to ensure a normal distribution of the

EMG amplitude. The mean peak amplitude of positive half

waves of the raw EMG signal was computed. ‘‘Low EMG’’

epochs were defined as epochs where the EMG amplitude

was below the mean value by at least one standard deviation.

This criterion was considered sufficient to exclude epochs

with high locomotor activity from ‘‘low EMG epochs’’ and

vice versa. The remaining epochs were considered ‘‘high

EMG’’ epochs (Table 1). This subdivision resulted in 16.8 T0.6% (SEM) and 17.1 T 0.5% of ‘‘low EMG’’ epochs during

SDL and SDD, respectively.

2.5. Statistics and analyses

Both EMG amplitude and EEG spectra were analyzed for

2-s epochs. The time course of the EEG spectrum was

assessed with one-way ANOVAs for repeated measures,

factor Ftime interval_. Contrasts were tested by post hoc two-tailed t tests. The relationship between the changes of EEG

power during SDL and SDD, and the SWA increase during

recovery sleep, was assessed by linear correlation analyses.

The change of waking EEG power was computed by

comparing the first and last 2 h of the 6-h SD. The change

in SWA after SD was defined as its level in NREM sleep

during the first 0.5 h (SDL) or 1.5 h (SDD) expressed as

percentage of the SWA level during NREM sleep in the last 3

h (SDL) or 2 h (SDD) immediately before the SD (different

interval length was used for SDL and SDD to ensure that a

sufficient amount of sleep would contribute to the data,

despite the circadian differences in the amount of sleep

between the two experiments). The increase of SWA was

computed as a ratio between SWA after the SD and the

interval immediately preceding the SD in order to take into

account the different sleep–wake history prior to SD, that

influences the initial level of SWA as well as a circadian

influence. The amount of NREM sleep contributing to this

analysis was in minutes: SDL before SD: 66.7 T 7.8, after SD:23.1 T 0.6; SDD before SD: 64.1 T 3.1, after SD: 57.2 T 4.8.

The software package MATLAB (The Math Works, Inc.,

Natick, MA, USA) was used for analysis, and SAS (SAS

Institute, Inc., Cary, NC, USA) for statistics.

3. Results

3.1. The largest increase in EEG power during SD was in

the theta band

The waking EEG spectrum during SD was characterized

by a prominent peak at 7.5 Hz (Fig. 1, right panel, indicated

by a vertical line). During SDL, EEG power computed for 2-h

intervals increased significantly in the 1.5–6.5 Hz range in

the course of SD, while power at the peak frequency did not

change (1-way ANOVA for repeated measures, factor F2-hinterval_: three intervals). Fig. 2 (left panel) illustrates the

absolute EEG spectra of the first and last 2-h SD. The

comparison of EEG power during the first and last 2 h showed

the largest increase in frequencies between 5 and 6 Hz (low

theta range). A significant increase was present in three

frequency bands (1.5–6.5 Hz, 10–14 Hz, and 17–25 Hz;

Fig. 2, middle panel). The changes were similar, but smaller

during SDD (Fig. 2, right panel).

3.2. Most prominent changes in EEG power were observed

during quiet waking and not during active waking

To investigate whether the increase of low-frequency theta

activity (5–6 Hz) is different depending on the intensity of

waking behaviors, EEG spectra of the 6-h SDL were com-

puted separately for ‘‘low EMG’’ and ‘‘high EMG’’ epochs.

Regular theta activity was evident during waking with high

EMG activity, while a mixed pattern, dominated by slower

theta waves, was typical for waking with low EMG activity

(Fig. 1, left panels). ‘‘Low EMG’’ epochs were characterized

by significantly lower mean EMG amplitude (Table 1) and

reduced standard deviation of the EMG amplitude (mean

standard deviation of ‘‘high EMG’’ epochs: SDL, 30.6 T 3.7,SDD, 33.3 T 3.8; of ‘‘low EMG’’ epochs: SDL, 11.7 T 1.4;

SDD, 12.6 T 1.2) compared to ‘‘high EMG’’ epochs.

The theta-peak frequency occurred at 5.5 Hz during

epochs with low EMG activity and at 7.5 Hz in epochs with

high EMG activity (Fig. 3, left panel, delimited by vertical

lines within the panel). EEG power during ‘‘low EMG’’

Fig. 1. Left panels: examples of 4-s raw EEG during active and quiet waking and corresponding EMG activity during a baseline day in an individual rat (right

parietal derivation). Right panel: EEG power density spectrum in total waking during 6 h sleep deprivation in the light period (n = 7). Mean EEG spectrum for

all artifact free 2-s epochs of waking.

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–71 67

epochs showed a significant increase from hours 1–2 to 5–

6 of SDL in all frequencies above 1 Hz except for the 7.5 Hz

bin (Fig. 3, middle panel). In contrast, in waking with high

EMG activity, EEG power at the theta-peak frequency did

not change during SDL, and a minor increase was present in

the delta and lower theta frequencies (Fig. 3, right panel).

3.3. Theta-wave amplitude increases during sleep

deprivation

To investigate in more detail whether theta-wave

amplitude or incidence contribute to the increase in EEG

power in the theta band, period–amplitude analysis was

Fig. 2. Left panel: EEG power spectra in waking during hours 1–2 and 5–6 of 6

Middle and right panels: relative EEG spectra during hours 5–6 as percentage of h

significant F values of one-way ANOVA for repeated measures, performed on log

Hz denote the theta-peak frequency during quiet and active waking (see Fig. 3).

performed on the 5–7 Hz band. Theta-wave amplitude

increased from hours 1–2 to hours 5–6 in both SDL and

SDD (Fig. 4). In contrast, the number of theta waves per

second did not change significantly.

3.4. Positive correlation between the increase of EEG power

during SD and SWA in NREM sleep during recovery

During the first 30 min after SDL, EEG power in NREM

sleep was significantly increased in a broad band encom-

passing 0.5–8.5 Hz, with a peak between 1 and 3 Hz (not

shown). We tested the hypothesis that for SDL the increase

of EEG power in ‘‘low EMG’’ waking is associated with the

h sleep deprivation (SD) in the light period (n = 7). Mean values (TSEM).

ours 1–2 of SD Light and SD Dark. Vertical bars below the panels represent

-transformed spectral values. Vertical lines within the panels at 5.5 and 7.5

Fig. 3. EEG power spectra in waking; left panel: during 6 h sleep deprivation (SD) in the light period (n = 7). Mean EEG spectra for 2-s epochs of all waking

epochs (i.e., including waking with low EMG and high EMG activity, TSEM). Vertical bars in the lower panel represent significant ( P < 0.05) F values of one-

way ANOVA for repeated measures performed on log-transformed spectral values. Middle and right panels: during hours 1–2 and 5–6 of 6 h SD. Mean values

(TSEM) separately for ‘‘low EMG’’ epochs and ‘‘high EMG’’ epochs. Vertical bars in the lower panel represent significant F values of one-way ANOVA for

repeated measures, performed on log-transformed spectral values. Vertical lines within the panels delimit the theta-peak frequencies at 5.5 and 7.5 Hz.

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–7168

rise of SWA in recovery NREM sleep. The increase of EEG

power during waking in SD hours 5–6 was expressed as

percentage of SD hours 1–2. The increase of SWA (0.5–4.0

Hz) was defined as its level in the first 30 min of recovery

sleep, expressed as percentage of the last 3 h preceding SD

(see Materials and methods). The mean absolute values of

SWA in NREM sleep (n = 7) during these intervals in AV2/

0.5 Hz were: baseline 179.1 T 38.5; recovery: 427.7 T 73.5

(P < 0.01, paired t test).

Fig. 4. Mean values (TSEM) of theta-wave (theta frequency defined as 5–7

Hz) amplitude (maximal voltage of positive half-waves), and incidence

(number of theta waves per second) are shown for ‘‘low EMG’’ epochs

during hours 1–2 and 5–6 of SDL (n = 7) and SDD (n = 7). Asterisks

above the bars indicate a significant increase during SD (**P < 0.01; *P <

0.05, two-tailed paired t test).

Correlation analysis revealed a cluster of significant bins

at 5–7 Hz (Fig. 5). In this frequency band, the increase of

EEG power during quiet waking correlated positively with

the subsequent increase of SWA during recovery sleep. This

frequency band (5–7 Hz) corresponded to the peak on the

power spectrum, indicating that changes of the dominant

oscillation of the waking EEG are associated with sleep

homeostasis. EEG power in this frequency band increased

significantly during hours 5–6 of SDL compared to hours

1–2 (109.2 T 9.8 vs. 81.7 T 7.4 AV2, paired t test, P < 0.05).

This increase correlated positively with the increase of SWA

Fig. 5. Linear correlation analysis between the increase of EEG power in

‘‘low EMG’’ epochs during 6 h sleep deprivation (SD) in the light period (n =

7), and subsequent increase of SWA (0.5–4 Hz) in NREM sleep. Increase of

EEG power during SD: hours 5–6 as percentage of hours 1–2. Increase of

SWA: first 30 min of sleep after SD as % of last 3 h before SD. r values are

shown for each frequency bin. Lower panel shows frequency bins where the

correlation was significant (P < 0.05) or reached a tendency (P < 0.1).

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–71 69

in NREM sleep during recovery (Fig. 6; left panel, r value =

0.76, P < 0.05). Also, the increase of theta-wave amplitude

from hours 1–2 to hours 5–6 of SDL showed a significant

correlation with the increase of SWA during recovery (r =

0.78, P = 0.04).

A similar computation for the rats subjected to SDD

showed that during low EMG epochs, theta activity (5–7

Hz) also increased in the course of the SD but did not

change significantly (97.7 T 11.0 vs. 84.4 T 12.9 AV2, n.s.).

However, a tendency for a positive correlation was present

(r = 0.70, P = 0.08; Fig. 5, right panel). Mean absolute

values of SWA in NREM sleep before and after SDD in

AV2/0.5 Hz were: baseline 102.5 T 7.0; recovery: 319.0 T42.9 (as for SDL, the increase was significant; P < 0.01,

paired t test). The correlation between the increase in theta

power during waking and EEG power in NREM sleep was

higher when it was computed for a broader ‘‘low’’-

frequency range (0.5–5.5 Hz; r = 0.74, P = 0.058). The

correlation for ‘‘high EMG’’ epochs was neither significant

for SDL nor SDD.

To control for a circadian influence, the time course of

theta activity in ‘‘low EMG’’ waking epochs was computed

also for the first 6 h of the baseline light and dark period,

where no increase of sleep propensity was expected. One-

way ANOVA, factor F2-h interval_ was not significant for

either deprivation, indicating that the increase of theta

activity during the SD was not merely reflecting a circadian

variation.

4. Discussion

We demonstrate for the first time in a rodent that an EEG

parameter during wakefulness reflects sleep propensity and

is related to the subsequent increase in sleep intensity. EEG

power in the theta band increased during SD, and correlated

positively with the level of SWA in NREM sleep during

Fig. 6. Scatter plot of the correlation between the increase of theta activity (5–7 Hz

(n = 7) and the dark period (n = 7), and the increase of SWA (0.5–4 Hz) in NR

recovery. The increase of EEG power at the theta-peak

frequency was restricted to epochs of quiet waking. This

effect was statistically significant for SDL and showed a

trend for SDD.

The results are consistent with findings in humans

[2,3,9,12,24,28]. Early studies in the rat showed that delta

activity in the waking EEG increases during SD and remains

above baseline values during recovery [8,13]. Also, in mice,

low-frequency EEG power was increased during 6 h SD

[17]. In mice, exploratory behavior is characterized by a

higher theta-peak frequency than quiet waking [16]. Also, in

the present study, the theta-peak frequency was higher

during waking epochs with high EMG (Fig. 3, left panel).

However, neither the peak frequency nor power at the peak

frequency changed in the course of SD, indicating that it

reflects locomotion or exploratory activity, rather than sleep

homeostasis.

EEG power in the low frequencies between 0.5 and 1 Hz

was enhanced during epochs with high EMG activity

compared to ‘‘low EMG’’ epochs (Fig. 3, left panel). This

increase was probably due to movement artifacts, and not

based on neurophysiological mechanisms, since physiolog-

ical delta activity generation is incompatible with active

wakefulness [23]. EEG power in these low-frequency bins

did not show any change in the course of SD, indicating that

low frequencies in the waking EEG are unrelated to sleep

propensity.

After both SD experiments, SDL and SDD, SWA during

NREM sleep increased significantly. This result is consistent

with previous reports [13,26]. An important finding resulted

from the distinction of the EEG spectrum belonging to

active and quiet waking. The subdivision was based on

EMG activity and resulted in significantly lower mean EMG

amplitude and standard deviation of the EMG amplitude.

Larger EMG variability during ‘‘high EMG’’ epochs is a

result of phasic activity, associated with active movements,

in contrast to quiet waking dominated by tonic EMG. Only

) in ‘‘low EMG’’ epochs during 6 h sleep deprivation (SD) in the light period

EM sleep after SD. Straight line depicts linear regression.

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–7170

the increase of theta power during quiet waking predicted

the subsequent increase of SWA in NREM sleep.

The results were similar to the finding obtained in

humans. Thus, the increase of theta power during 40 h of

waking was maximal at 6.5 Hz, and the rate of increase of

power between 5 and 8 Hz correlated positively with the

increase of SWA in NREM sleep [12]. The design in the

study performed in human subjects [12] ensured that the

waking EEG corresponded to quiet waking.

It had been proposed previously that the increase of delta

activity in the waking EEG during SD reflects sleep

propensity in mice [17]. Also, in humans, EEG delta power

increased during a prolonged waking episode [9,12,28].

However, although in the present study power increased

over a large range of frequencies, including also the delta

frequency band, a significant correlation with the subse-

quent increase of SWA was restricted to a cluster of

frequency bins (5–7 Hz). This particular band encompassed

the peak frequency of the EEG during quiet waking (Fig. 3).

This result indicated that theta activity, the main oscillatory

activity of the waking EEG, is associated with sleep

homeostasis while other frequencies are not consistently

changed.

It has been proposed that the increase of theta power

during SD in humans could represent a compensatory

process counteracting the increasing sleep pressure [12]. An

interesting alternative is that the buildup of theta activity

during waking could reflect a process that has to be

compensated during sleep with delta waves. The need for

such a compensation is expected to be higher when SD is

performed in the light period, when rats normally sleep most

of the time, compared to SD in the dark period. In support of

this notion, theta power in the rat did not increase

significantly during the SD in the dark period, when the

nocturnal rat is awake spontaneously most of the time [7].

There was only a weak association between theta activity

and SWAwhen the rats were sleep deprived during the dark

period, when the SD was performed under relatively low

initial sleep pressure. Also, in humans, only a moderate

change of theta power occurred during the light period,

when the subjects were awake spontaneously, and not yet

sleep deprived [12,21]. Period–amplitude analysis provided

an interesting finding, revealing that it was the amplitudes of

theta waves which increased, both during SDL and SDD,

indicating that the amplitude of theta waves in the waking

EEG may be a sensitive measure for the duration of

wakefulness, i.e., of sleep propensity.

Further studies are needed to clarify whether the

generation of theta activity is similar in different species,

and which mechanisms underlie the relation of theta power

in waking and subsequent SWA in NREM sleep. In rodents,

theta activity is believed to originate from the hippocampus

[4], but prominent theta activity has been recorded also

locally from several extrahippocampal regions, including

cortical areas (reviewed in [19]). In humans, an independent

generation of theta activity in the hippocampus and in the

neocortex was postulated [10]. Thus, it is likely that both in

humans and rodents there are multiple theta generators.

It has been proposed recently that neuronal potentiation

occurs during waking necessitating synaptic downscaling

during subsequent sleep [27]. EEG power in the theta band

during quiet waking might reflect the level of synaptic

potentiation, or the degree of ‘‘saturation’’ of synaptic

potentiation.

In conclusion, the data indicate that theta power and

theta-wave amplitude during quiet wakefulness can be used

as markers for the buildup of the homeostatic process of

sleep regulation in the rat.

Acknowledgments

We thank Drs. A.A. Borbely and P. Achermann for

advice and comments on the manuscript. The study was

supported by Swiss National Science Foundation grants

3100A0-100567, 3100-053005.97, and 3100.062112.00.

References

[1] P. Achermann, D.J. Dijk, D.P. Brunner, A.A. Borbely, A model of

human sleep homeostasis based on EEG slow-wave activity: quanti-

tative comparison of data and simulations, Brain Res. Bull. 31 (1993)

97–113.

[2] D. Aeschbach, J.R. Matthews, T.T. Postolache, M.A. Jackson, H.A.

Giesen, T.A. Wehr, Dynamics of the human EEG during prolonged

wakefulness: evidence for frequency-specific circadian and homeo-

static influences, Neurosci. Lett. 239 (1997) 121–124.

[3] D. Aeschbach, J.R. Matthews, T.T. Postolache, M.A. Jackson, H.A.

Giesen, T.A. Wehr, Two circadian rhythms in the human electro-

encephalogram during wakefulness, Am. J. Physiol. 277 (1999)

R1771–R1779.

[4] B.H. Bland, The physiology and pharmacology of hippocampal

formation theta rhythms, Prog. Neurobiol. 26 (1986) 1–54.

[5] A.A. Borbely, A two process model of sleep regulation, Hum.

Neurobiol. 1 (1982) 195–204.

[6] A.A. Borbely, P. Achermann, Homeostasis of human sleep and models

of sleep regulation, in: M.H. Kryger, T. Roth, W.C. Dement (Eds.),

Principles and Practice of Sleep Medicine, W.B. Saunders Co.,

Philadelphia, 2005, pp. 405–417.

[7] A.A. Borbely, H.U. Neuhaus, Daily pattern of sleep, motor activity

and feeding in the rat: effects of regular and gradually extended

photoperiods, J. Comp. Physiol. 124 (1978) 1–14.

[8] A.A. Borbely, I. Tobler, M. Hanagasioglu, Effect of sleep deprivation

on sleep and EEG power spectra in the rat, Behav. Brain Res. 14

(1984) 171–182.

[9] C. Cajochen, J.K. Wyatt, C.A. Czeisler, D.-J. Dijk, Separation of

circadian and wake duration-dependent modulation of EEG activation

during wakefulness, Neuroscience 114 (2002) 1047–1060.

[10] J.L. Cantero, M. Atienza, R. Stickgold, M.J. Kahana, J.R. Madsen, B.

Kocsis, Sleep-dependent theta oscillations in the human hippocampus

and neocortex, J. Neurosci. 23 (2003) 10897–10903.

[11] S. Daan, D.G.M. Beersma, A.A. Borbely, Timing of human sleep:

recovery process gated by a circadian pacemaker, Am. J. Physiol. 246

(1984) R161–R178.

[12] L.A. Finelli, H. Baumann, A.A. Borbely, P. Achermann, Dual

electroencephalogram markers of human sleep homeostasis: correla-

tion between theta activity in waking and slow-wave activity in sleep,

Neuroscience 101 (2000) 523–529.

V.V. Vyazovskiy, I. Tobler / Brain Research 1050 (2005) 64–71 71

[13] P. Franken, D.-J. Dijk, I. Tobler, A.A. Borbely, Sleep deprivation in

rats: effects on EEG power spectra, vigilance states, and cortical

temperature, Am. J. Physiol. 261 (1991) R198–R208.

[14] P. Franken, A. Malafosse, M. Tafti, Genetic variation in EEG activity

during sleep in inbred mice, Am. J. Physiol. 275 (4 Pt. 2) (1998)

R1127–R1137.

[15] J.A. Horne, S.D. Baulk, Awareness of sleepiness when driving,

Psychophysiology 41 (2004) 161–165.

[16] R. Huber, T. Deboer, I. Tobler, Prion protein: a role in sleep

regulation? J. Sleep Res. 8 (Suppl. 1) (1999) 30–36.

[17] R. Huber, T. Deboer, I. Tobler, Effects of sleep deprivation on sleep

and sleep EEG in three mouse strains: empirical data and simulations,

Brain Res. 857 (2000) 8–19.

[18] B.E. Jones, Basic mechanisms of sleep–wake states, in: M.H. Kryger,

T. Roth, W.C. Dement (Eds.), Principles and Practice of Sleep

Medicine, W.B. Saunders Co., Philadelphia, 2000, pp. 134–154.

[19] M.J. Kahana, D. Seelig, J.R. Madsen, Theta returns, Curr. Opin.

Neurobiol. 11 (2001) 739–744.

[20] G. Kecklund, T. Akerstedt, Sleepiness in long distance truck driving:

an ambulatory EEG study of night driving, Ergonomics 36 (1993)

1007–1017.

[21] H.P. Landolt, J.V. Retey, K. Tonz, J.M. Gottselig, R. Khatami, I.

Buckelmuller, P. Achermann, Caffeine attenuates waking and sleep

electroencephalographic markers of sleep homeostasis in humans,

Neuropsychopharmacology 29 (2004) 1933–1939.

[22] S. Makeig, T.P. Jung, T.J. Sejnowski, Awareness during drowsiness:

dynamics and electrophysiological correlates, Can. J. Exp. Psychol. 54

(2000) 266–273.

[23] M. Steriade, The corticothalamic system in sleep, Front. Biosci. 8

(2003) d878–d899.

[24] A.M. Strijkstra, D.G.M. Beersma, B. Drayer, N. Halbesma, S. Daan,

Subjective sleepiness correlates negatively with global alpha (8–12

Hz) and positively with central frontal theta (4–8 Hz) frequencies in

the human resting awake electroencephalogram, Neurosci. Lett. 340

(2003) 17–20.

[25] I. Tobler, Phylogeny of sleep regulation, in: M.H. Kryger, T. Roth,

W.C. Dement (Eds.), Principles and Practice of Sleep Medicine, W.B.

Saunders Co., Philadelphia, 2005, pp. 77–90.

[26] I. Tobler, A.A. Borbely, Sleep EEG in the rat as a function of prior

waking, Electroencephalogr. Clin. Neurophysiol. 64 (1986) 74–76.

[27] G. Tononi, C. Cirelli, Sleep and synaptic homeostasis: a hypothesis,

Brain Res. Bull. 62 (2003) 143–150.

[28] L. Torsvall, T. Akerstedt, Sleepiness on the job: continuously

measured EEG changes in train drivers, Electroencephalogr. Clin.

Neurophysiol. 66 (1987) 502–511.

[29] C.H. Vanderwolf, Cerebral activity and behavior: control by central

cholinergic and serotonergic systems, Int. Rev. Neurobiol. 30 (1988)

225–340.

[30] V.V. Vyazovskiy, P. Achermann, A.A. Borbely, I. Tobler, Inter-

hemispheric coherence of the sleep electroencephalogram in mice

with congenital callosal dysgenesis, Neuroscience 124 (2004)

481–488.