Analysis of the Bispectral Index During Natural Sleep in Children

Benini, Franca MD; Trapanotto, Manuela MD; Sartori, Stefano MD; Capretta, Anna MD; Gobber, Daniela MD; Boniver, Clementina MD; Zacchello, Franco PhD, MD

Section Editor(s): Davis, Peter J.

Anesthesia & Analgesia:
doi: 10.1213/01.ANE.0000157566.41739.2C
Pediatric Anesthesia: Research Report

The Bispectral Index (BIS) is a system used to measure sedation levels. Some investigators recently analyzed changes in the BIS during natural sleep in adults and found that the BIS diminished considerably as sleep became deeper. No such studies have been undertaken to investigate changes in the BIS in sleeping children. The aim of this work was to assess the trend of the BIS in the various stages of sleep in a group of children, performing a descriptive analysis on a limited number of cases. We evaluated 15 children with negative clinical findings and a normal electroencephalogram (EEG) as part of their follow-up for prior episodes of epilepsy, recording the BIS and EEG in a waking state, in the various stages of sleep and on reawaking. For each stage, the mean value, the standard deviation, and the range of BIS values were calculated. The results showed that the BIS decreased progressively as sleep became deeper. The correlation between the stage of sleep and the BIS was significant. On reawaking, a slow increase was apparent in the BIS.

In Brief

IMPLICATIONS: Some studies in adults found that the Bispectral Index (BIS) decreased as sleep became deeper, providing a nonspecific measure of the subject’s level of consciousness. This study investigated the changes of BIS during natural sleep in healthy children.

Author Information

Department of Pediatrics, University of Padova, Padova, Italy

Accepted for publication January 12, 2005.

Address correspondence and reprint requests to Franca Benini, Department of Pediatrics, University of Padova, Padova, Italy. Address e-mail to

Article Outline

The Bispectral Index (BIS) system uses an integrated analysis of the electroencephalogram trace (EEG), obtained with 2 frontal leads (F7, F8), to enable an assessment of the level of sedation and anesthesia (1). The BIS records the continuous EEG signal, breaking it down and eliminating artifacts, then it compares the EEG pattern acquired instant-by-instant with EEG traces saved in its database to provide a single number on a scale of 0 to 100 that correlates with the subject’s level of sedation (2).

Some studies have analyzed changes in the BIS during sleep in adults. Sleigh et al. (3) found that the BIS decreased as sleep became deeper in much the same way as under sedation, thereby demonstrating that this index provides a nonspecific measure of the subject’s level of consciousness. Other authors emphasized that the BIS cannot be used conventionally to differentiate between the different stages of sleep because of the overlap in the values recorded by the index at the moment of the shift from one state to the next (4). Moreover, Tung et al. (5) demonstrated that the BIS is able to detect only the beginning of sleep in adults. However, the BIS variables were constructed based on EEG in sedated and anesthetized adults, not during natural sleep.

Although several studies have considered the applicability of BIS to evaluate sedation and anesthesia in pediatric patients (6–8), none has investigated its changes during natural sleep in children. The aims of the present study were (a) to evaluate the trend of the BIS in the various stages of sleep in a group of children by means of a descriptive analysis on a limited series of cases; (b) to obtain information on the BIS trend in natural or endogenous situations in order to develop baseline data useful for future research in clinical settings such as sedation.

Back to Top | Article Outline


After obtaining the informed consent of their parents, and institutional committee approval from our Department, EEG and BIS were recorded during natural sleep in 15 children aged between 1.2 and 16.5 yr (mean age 8.2 yr; 8 boys and 7 girls), who were being followed for prior episodes of epilepsy and at the time of this evaluation were seizure-free for 2 yr. They had negative clinical findings and an EEG (waking and sleeping) with no significant anomalies, and they were receiving no neuroactive pharmacological treatment. The study was performed at the Neurophysiology Service of the University of Padova Pediatrics Department.

Before the assessment, the children were deprived of sleep (from 11 pm to 3 am on the previous night). When they arrived at the Department, they were accompanied to a soundproofed room (some of them asked to have a parent with them in the room).

Electrodes were applied to record the EEG (Galileo Planet 300 Plus Basic) and the BIS (Aspect A-2000). A conventional EEG polysomnograph was recorded using a 16-channel computerized system. The electrodes for the BIS were applied according to the standard layout, i.e., with 2 bilateral frontal electrodes coinciding with the right and left prefrontal cortex (F7 and F8) to provide a bipolar bifrontal signal and 1 earthing electrode between the corner of the eye and the hairline. The low- and high-frequency filters were set respectively at 0.5 and 30.0 Hz.

The quality of the signal was expressed by SQI (signal quality index) that provides a score from 0% to 100%. We only considered BIS values with an SQI >50%. None of the data were eliminated because of electromyograph artifacts.

Changes were simultaneously recorded, continuously for the EEG and at 1-min intervals for the BIS, in the following stages of sleep: while falling asleep, during the 4 stages of sleep, and on subsequent reawaking. We differentiated each stage of sleep according to the classification of Rechtschaffer and Kales (9). EEG and BIS recordings were obtained for at least one sleep cycle in each child, though it was not always complete with all sleep stages according to conventional EEG staging.

A descriptive analysis was performed on the resulting data.

Back to Top | Article Outline


For each stage considered (waking, sleeping stages I–IV, reawaking), we calculated the mean values, standard deviation and range (minimum and maximum), and the duration of the recording obtained in minutes (Table 1). The total recording time for the BIS was 621 min. Some of the first few minutes of the waking stage were not recorded in four subjects because of recording artifacts. Three subjects woke up (for 1–2 min) and then fell asleep again during the course of the recording. The average latency period between waking and light sleeping (stage I) was 16.22 min (range, 2–59). All the children reached stage II sleep; 3 children reached stage III and 1 reached stage IV (Fig. 1). Data analysis revealed considerable variations in the BIS during the various stages of sleep: these changes reflect the classic description of sleep architecture. As the stage of sleep became deeper, the BIS values progressively decreased (Table 1). On reawaking, the BIS gradually increased again to a mean of 87.67.

A correlation analysis performed on the data recorded by the BIS in the various stages of sleep was found significant (r = −0.713; P < 0.05): as sleep became deeper (from stage I to IV), there was a concomitant decrease in the BIS (Fig. 2).

To analyze the trend of the BIS in the reawaking stage, a comparison was drawn with the EEG values coinciding with wakefulness. It emerged that the BIS reaches the values >90 typical of a wakeful condition with a mean latency of 2.5 min (sd 1.6; range, 0–5) with respect to the EEG recording.

Back to Top | Article Outline


Although it was conducted on a limited number of subjects, our study nonetheless shows that the BIS is sensitive in reflecting the changes on the EEG trace that accompany the various stages of natural sleep. To be more specific, the BIS values decrease progressively as the stage of sleep becomes deeper. There are very few studies on this topic in the literature and they were all performed on very small samples (from 5 to 10) of adults (3–5). Sleigh et al. (3) found BIS values between 75 and 90 for sleep stages I and II and between 20 and 70 for stages III and IV. In addition to the limited number of subjects considered, an important drawback of this study was that the authors did not subdivide the sleep stages into four, but rather into the coarser gradation of light sleep and slow-wave sleep. Nieuwenhuijs et al. (4) found lower mean values for the BIS than Sleigh et al. in a similar study. Both these studies drew a distinction only between light and deep sleep. Similar to the method adopted more recently in the adult (5), our study distinguished each single stage of sleep.

It is worth noting that the BIS values recorded in deep sleep are comparable with those observed under deep sedation, a finding that confirms the nonspecificity of the BIS. This tool does not enable a distinction between the cause of the change in level of consciousness and the ability to wake up again (3). It is consequently impossible to use BIS values to distinguish “natural or endogenous” from “exogenous or pharmacological” sedation; such a discrimination has to be based on known clinical variables (reaction to touch or voice, eye movement, etc.) to avoid the risk of misinterpreting the subject’s real condition. A child’s reaction to a painful stimulus or surgical incision would be very different in the two above-mentioned situations, despite similar BIS values being recorded in the two cases. The same applies when a child’s degree of wakefulness needs to be assessed after a procedure conducted under sedation-analgesia; a low BIS may not necessarily indicate a persistent sedative effect of the medication administered, it might simply be a state of deep sleep resulting from the cessation of a stressful condition.

Finally, when children are woken, their BIS gradually increases until it reaches values similar to those recorded before they fell asleep. Observations in adults (4) have shown a gap, during the reawaking stage, between the subjects’ behavior and their slower gradual return to normal BIS values (4). This gap was also identifiable in our study, between the EEG values corresponding to a waking condition and the BIS values recorded, the latter demonstrating a prolonged latency. There are two main hypotheses to explain this: either this prolonged latency is caused by the BIS recording method (which averages the values recorded at intervals ranging between 5 seconds and 1 minute) or the slow change in the BIS values derives from a differential reactivation of the various cortical sites as the subject wakes up.

One of the drawbacks of our study is undeniably the limited number of subjects analyzed. For this reason, we were unable to evaluate the BIS differences according to children’ ages. Publications in the literature on adults are also of an exploratory nature, however, and have analyzed even smaller samples. Another shortcoming is that no data were recorded for the REM stage of sleep, so we are unable to say what changes the BIS may be able to record during this stage. The few works available in the literature on adults report BIS values very similar to those recorded in our study during the waking stage (3,4).

In conclusion, as in adults, the BIS recorded in children decreases considerably with deeper stages of natural sleep, in much the same way as in deep sedation and irrespective of the reason for the altered state of consciousness.

Back to Top | Article Outline


1. Vernon JM, Long E, Sebel PS, Manberg P. Prediction of movement using bispectral electroencephalographic analysis during propofol/alfentanil or isoflurane/alfentanil anesthesia. Anesth Analg 1995;80:780–5.
2. Sigl JC, Chamoun NC. An introduction to bispectral analysis for the electroencephalogram. J Clin Monit 1994;10:392–404.
3. Sleigh JW, Andrzejowski J, Steyn-Ross A, Steyn-Ross M. The bispectral index: a measure of depth of sleep? Anesth Analg 1999;88:659–61.
4. Nieuwenhuijs D, Coleman EL, Douglas NJ, et al. Bispectral index values and spectral edge frequency at different stages of physiologic sleep. Anesth Analg 2002;94:125–9.
5. Tung A, Lynch JP, Roizen MF. Use of the BIS monitor to detect onset of naturally occurring sleep. J Clin Monit 2002;17:37–42.
6. Denaman WT, Swanson EL, Rosow D, et al. Pediatric evaluation of the bispectral index (BIS) monitor and correlation of BIS with end-tidal sevoflurane concentration in infants and children. Anesth Analg 2000;90:872–7.
7. Bannister CF, Brosius KK, Sigi JC, Meyer BJ. The effect of BIS monitoring on anesthetic use and recovery in children anesthetized with sevoflurane in nitrous oxide. Anesth Analg 2001;92:877–81.
8. Berkenbosch JW, Fichter CR, Tobias JD. The correlation of the bispectral index monitor with clinical sedation score during mechanical ventilation in the pediatric intensive care unit. Anesth Analg 2002;94:506–11.
9. Rechtschaffer A, Kales A, eds. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects (NIM publication no. 204). Washington, DC: US Government Printing Office, 1968.
© 2005 International Anesthesia Research Society