The e€ect of arousals during sleep onset on estimates of sleep
onset latency
S IMON SMITH and JOHN TR INDER
Department of Psychology, University of Melbourne, Parkville Victoria 3052, Australia
Accepted in revised form 30 November 1999; received 2 January 1999
INTRODUCTION
People with insomnia frequently overestimate the time it takes
them to fall asleep when compared with objective measures
such as polysomnography (PSG). Good sleepers, however,
tend to correctly estimate or slightly underestimate the amount
of time taken to fall asleep (Rechtscha€en and Monroe 1969;
Bixler et al. 1973; Carskadon et al. 1976; Frankel et al. 1976;
Lutz et al. 1977; Coates et al. 1983; Hauri and Olmstead
1983; Edinger and Fins 1995; Perlis et al. 1997). Frequent
transient arousals, too brief to be recognized by conventional
sleep-stage scoring or be subjectively experienced by sleepers
(Knab and Engel 1988; Perlis et al. 1997) have been associated
with daytime symptoms of sleepiness and poor cognitive
functioning (Stepanski et al. 1984; Bonnet 1985; Guilleminault
et al. 1993). In this study, the extent to which transient
arousals influence the perception of sleep onset was assessed.
The frequency of brief awakenings, or microarousals, has
been scored for clinical purposes to determine the extent of
sleep fragmentation (Mathur and Douglas 1995; Martin et al.
1997). A number of definitions for measuring sleep fragmen-
tation have been proposed (Atlas Task Force of the American
Sleep Disorders Association 1992
1
; Cheshire et al. 1992;
Mathur and Douglas 1995). In general, such definitions regard
a return to alpha frequencies in the electroencephalogram
(EEG) or a conspicuous increase in submental electromyo-
gram (EMG) tone as indicative of arousal. The current
standard definition is that recommended by the American
Sleep Disorders Association (ASDA), and requires a minimum
duration of alpha activity of 3 s. A greater number of arousals
can be identified with shorter duration criteria (Martin et al.
1997). However, the most clinically useful definition for
microarousals has yet to be determined as current definitions
yield only a moderate relationship between arousal frequency
and daytime sequalae.
Correspondence: Dr S. Smith, Department of Psychology, University
of Melbourne, Parkville, Victoria 3052, Australia. Tel.: 61 3
9344 4549; fax: 61 3 9347 6618; e-mail: s.smith7@pgrad.unimelb.
edu.au.
J. Sleep Res. (2000) 9, 129–135
SUMMARY
It is well established that insomniacs overestimate sleep-onset latency. Furthermore,
there is evidence that brief arousals from sleep may occur more frequently in insomnia.
This study examined the hypothesis that brief arousals from sleep influence the
perception of sleep-onset latency. An average of four sleep onsets was obtained from
each of 20 normal subjects on each of two nonconsecutive, counterbalanced,
experimental nights. The experimental nights consisted of a control night (control
condition) and a condition in which a moderate respiratory load was applied to increase
the frequency of microarousals during sleep onset (mask condition). Subjective
estimation of sleep-onset latency and indices of sleep quality were assessed by self-
report inventory. Objective measures of sleep-onset latency and microarousals were
assessed using polysomnography. Results showed that sleep-onset latency estimates
were longer in the mask condition than in the control condition, an e€ect not reflected
in objective sleep-stage scoring of sleep-onset latency. Furthermore, an increase in the
frequency of brief arousals from sleep was detected in the mask condition, and this is a
possible source for the sleep-onset latency increase perceived by the subjects. Findings
are consistent with the concept of a physiological basis for sleep misperception in
insomnia.
KEYWORDS insomnia, microarousal, sleep fragmentation, sleep onset latency
Ó 2000 European Sleep Research Society 129

Most investigations of sleep fragmentation have arisen from
attempts to model sleep apnoea, and utilize estimates of
daytime sleepiness or performance on tasks as outcome
measures (Stepanski et al. 1987; Roehrs et al. 1994; Chugh
et al. 1996; Kimo€ 1996). However, a recent review by Perlis
et al. (1997) indicates that an increase in high-frequency EEG
activity across the sleep-onset period may underlie the subject-
ive underestimation of sleep duration in insomnia patients,
suggesting that sleep fragmentation might have an impact on
the self-perception of whether a person is asleep or not.
However, to date, there have been no published studies that
have adequately evaluated the e€ects of microarousals on
either the subjective perception of sleep duration or on the
perception that an individual is asleep.
Previous studies of sleep fragmentation have often used
auditory evoked arousal or ca€eine-induced hyperarousal
methods to simulate disrupted sleep and produce a model of
insomnia (e.g. Philip et al. 1994
2
; Series et al. 1994; Alford et al.
1996; Bonnet and Arand 1996; Badr et al. 1997; Martin et al.
1997). A possible alternative method for producing an increase
in microarousals is to exploit endogenous arousal mechanisms
associated with respiratory control during sleep. Respiratory
variables appear to be involved in initiating or maintaining
arousals during sleep, particularly during the sleep-onset
period (Gleeson et al. 1989; Trinder et al. 1992; Kay et al.
1994; Mathur and Douglas 1995
3,4
; Douglas and Martin 1996).
This occurs because alternation in state during sleep onset
produces fluctuations in ventilation and gas levels (Trinder
et al. 1992), which in turn result in arousals and even greater
instability in the respiratory system (Dunai et al. 1996).
Consistent with this, increasing breathing diculty by the
addition of an inspiratory resistance, or by temporary occlu-
sion of inspiration, leads to increased arousals (Carskadon and
Dement 1981). Increasing respiratory load therefore increases
state instability, and provides a naturalistic method of
increasing the frequency of microarousals during sleep onset,
thus providing a model for investigating features of insomnia
in normal individuals.
In this study, subjective indices of sleep parameters were
collected in conjunction with PSG measures of sleep in a
manipulation designed to increase state instability, and there-
fore the frequency of microarousals across the sleep-onset
period. It was hypothesized that the subjects would be able to
identify a period of state instability produced by respiratory
load, and that the duration of their subjective estimates of sleep
onset latency (SOL) would be increased by this perception.
Furthermore, it was hypothesized that PSG measures of sleep
onset would not be a€ected by an increase in microarousals.
METHODS
Subjects
Subjects were recruited from a university student population.
Twenty subjects (18 females and two males) with a mean age of
22.9 ‹ 3.2 y, participated in the experimental nights. The
subjects reported being nonsmokers (currently and in the past),
free of respiratory problems (including asthma) and free of
sleep problems. They were not on any medication, and were
asked to refrain from alcohol and ca€eine consumption for at
least 12 h prior to their participation. Subjects were screened
for psychopathology known to a€ect sleep using the Beck
Anxiety Inventory (Beck et al. 1988) and the Beck Depression
Inventory (Beck 1987). Habitual sleep was assessed for
one week prior to testing using the Pittsburgh Sleep Diary
(Monk et al. 1994), to ensure that subjects had regular sleep/
wake patterns.
Design
The experiment consisted of two within-subject variables (the
recording condition and the method of measuring SOL). The
recording condition involved two levels, a control condition
and an experimental condition. In the control condition
(Control), subjects slept with routine PSG recording equip-
ment and with body position unrestricted. In the experimental
recording condition (Mask), subjects wore respiratory appa-
ratus with light resistance and the body position remained
supine. The order of presentation of the conditions was
counterbalanced across subjects. The two methods of measur-
ing SOL were objective SOL estimates based on the PSG
record, and the subjective SOL estimates of the subjects. The
subjective measures comprised three interrelated estimates of
SOL.
Procedure
Subjects reported to the sleep laboratory one week prior to
testing for screening, briefing and distribution of sleep diaries.
Subjects underwent an adaptation night in the laboratory prior
to data collection to control for ‘first night’ phenomena
(Toussaint et al. 1995). Data collection nights were noncon-
secutive.
Subjects in the Mask condition wore a Hans Rudolph
silicon-rubber anaesthetic mask, which covered the nose and
mouth and was held in place by a head strap. The mask was
attached to a two-way nonrebreathing valve (Hans Rudolph
series 2600). The inspiratory resistance of the mask and leaf
valve assembly, measured at 0.25 L/s, was 1.25 cmH
2
O/L/s.
To maximize the number of sleep onsets recorded, a multiple
sleep onset procedure was used. Subjects were woken after a
variable period of uninterrupted Stage 2 sleep or deeper. This
period was of three lengths: short (9–11 min), medium (14–
16 min) and long (18–22 min). The times were presented in
random order. Subjects were given the following instructions:
‘The lights will be turned o€ and you should go to sleep. After
attaining some amount of sleep, you will be woken and asked to
estimate the time spent asleep, and a number of other questions.
This process will re-occur several times across the night.’
When woken, each subject was asked to complete subjective
measures of SOL. Subjects read the statement ‘We often drift
in and out of light sleep before going into continuous, or
130 S. Smith and J. Trinder
Ó 2000 European Sleep Research Society, J. Sleep Res., 9, 129–135