See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/262339524

Effect of Muslim Prayer (Salat) on α

Electroencephalography and Its Relationship
with Autonomic Nervous Syst....

Article in Journal of alternative and complementary medicine (New York, N.Y.) · May 2014
DOI: 10.1089/acm.2013.0426 · Source: PubMed

CITATIONS READS

9 467

4 authors:

Hazem Doufesh Fatimah Ibrahim

Al-Quds University University of Malaya
6 PUBLICATIONS 30 CITATIONS 188 PUBLICATIONS 1,040 CITATIONS

SEE PROFILE SEE PROFILE

Noor Azina Ismail Wan Azman Wan Ahmad

University of Malaya University of Malaya
118 PUBLICATIONS 455 CITATIONS 104 PUBLICATIONS 637 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Wireless power transfer for implantable medical devices View project

The Malaysian Elders Longitudinal Research (MELoR) View project

All content following this page was uploaded by Hazem Doufesh on 04 July 2015.

The user has requested enhancement of the downloaded file.

THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINE
Volume 20, Number 7, 2014, pp. 558–562
ª Mary Ann Liebert, Inc.
DOI: 10.1089/acm.2013.0426

Effect of Muslim Prayer (Salat)

on a Electroencephalography and Its Relationship
with Autonomic Nervous System Activity

Hazem Doufesh, MSc,1,2,3 Fatimah Ibrahim, PhD,1,2 Noor Azina Ismail, PhD,2,4
and Wan Azman Wan Ahmad, MRCP 2,5

Abstract

Objectives: This study investigated the effect of Muslim prayer (salat) on the a relative power (RPa) of
electroencephalography (EEG) and autonomic nervous activity and the relationship between them by using
spectral analysis of EEG and heart rate variability (HRV).
Methods: Thirty healthy Muslim men participated in the study. Their electrocardiograms and EEGs were
continuously recorded before, during, and after salat practice with a computer-based data acquisition system
(MP150, BIOPAC Systems Inc., Camino Goleta, California). Power spectral analysis was conducted to extract
the RPa and HRV components.
Results: During salat, a significant increase ( p < .05) was observed in the mean RPa in the occipital and parietal
regions and in the normalized unit of high-frequency (nuHF) power of HRV (as a parasympathetic index).
Meanwhile, the normalized unit of low-frequency (nuLF) power and LF/HF of HRV (as sympathetic indices)
decreased according to HRV analyses. RPa showed a significant positive correlation in the occipital and parietal
electrodes with nuHF and significant negative correlations with nuLF and LF/HF.
Conclusions: During salat, parasympathetic activity increased and sympathetic activity decreased. Therefore,
regular salat practices may help promote relaxation, minimize anxiety, and reduce cardiovascular risk.

Introduction Raichur and associates described the effect of meditation

training on the variation in respiration.9

I slamic prayer, commonly represented by the Arabic

term salat, is a form of meditation,1 and it is obligatory for
Muslims to perform the prayers five times daily at specific
Power spectral analysis is commonly applied to electro-
cardiography (ECG) signals to assess HRV as an indicator
of the autonomic nervous system.10,11 Many cardiovascular
prescribed times of the day. It is a religious physical activity disorder states are hypothesized to be associated with typical
that involves various Quran recitations and the performance variations in HRV.12 The cardiovascular system is usually
of specific postural positions, namely standing, bowing, controlled by autonomic regulation through the activity of the
prostration, and sitting.2 sympathetic and parasympathetic branches of the ANS.13 The
Various meditation forms can influence not only the auto- sympathetic branch, in a simplified sense, is responsible for
nomic nervous system (ANS),3,4 but also the central ner- stimulating activities associated with the fight-or-flight re-
vous system (CNS).5,6 For instance, Arambula and colleagues sponse, and the parasympathetic branch is responsible for the
showed an increase in a and h electroencephalography (EEG) calming-down response.14 The three common frequency
activities during Kundalini yoga meditation, and Peressutti and bands in the HRV spectrum are the very-low-frequency
colleagues showed variations in heart rate variability (HRV) (VLF) component (0.001–0.04 Hz), the low-frequency (LF)
and respiratory rate during meditation.7 Lee and colleagues component (0.04–0.15 Hz), and the high-frequency (HF)
found that heart rate, respiratory rate, and systolic blood component (0.15–0.4 Hz). VLF mainly reflects thermoregu-
pressure significantly decrease during Qi-training,8 and latory cycles, LF is usually considered a marker of mixed

1
Department of Biomedical Engineering and 2Centre for Innovation in Medical Engineering, Faculty of Engineering, University of
Malaya, Kuala Lumpur, Malaysia.
3
Department of Electronics Engineering, Faculty of Engineering, Al-Quds University, Jerusalem, Palestine.
4
Department of Applied Statistics, Faculty of Economics and Administration, University of Malaya, Kuala Lumpur, Malaysia.
5
Department of Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia.

558
MUSLIM PRAYER AND EFFECTS ON EEG 559

sympathetic-parasympathetic nervous activities, and HF Materials and Methods

reflects parasympathetic (vagal) nervous activities.11,12,15 In Participants
addition, the ratio of LF/HF represents the sympathovagal
balance, which is essential for good health.12,16 Recently, This study recruited 30 healthy Muslim men aged 20–35
HRV information has been obtained from individuals before years. The participants had no neurologic or psychological
and during meditation to understand ANS response induced disorder, and they were asked not to take any heavy meals or
by the meditative state.15,17 Meditation affects ANS by in- to do physical activity at least 4 hours before measurements
creasing and decreasing parasympathetic and sympathetic were taken.
activities, respectively. EEG can also be investigated with
power spectral analysis. The five spectral frequency bands Procedure
are d (0.5–4 Hz), h (4–8 Hz), a (8–13 Hz), b (13–30 Hz), and The experimental procedure for each study participant
c (30–70 Hz), which correspond to the classification of brain was divided into three sessions: prebaseline, Duha salat
waves. The a wave is one of the most dominant brain waves practice, and postbaseline. Before the first session (pre-
in meditation state. The a wave activity can be measured in baseline) recording, the participants lay on a bed in a quiet,
all regions of the brain. However, the highest a wave am- semi-darkened room and were asked to relax in the supine
plitude was observed in the occipital and parietal regions.18 position for 20 minutes so that they could adapt to the ex-
The increasing of a band frequency in meditation was perimental conditions. EEG and ECG signals were simul-
hypothesized to be promoted by changes in ANS, which taneously recorded. The data were collected with both eyes
induce relaxation response in humans.19,20 The generation open for 2 minutes, followed by both eyes closed for another
of a waves is generally associated with stimulation of 2 minutes and both eyes open for 2 minutes. Then, the
parasympathetic activity and reduction of the sympathetic participants were allowed to rest for 5 minutes.
activity of ANS.21 High levels of a activities were correlated In the Duha salat session, the participants were asked
with low levels of anxiety and feelings of calm and positive to perform salat. EEG and ECG signals were collected
affect.22,23 throughout the performance of each of the different postural
Many studies describe the relationship between the CNS positions of salat. Figure 1 shows a complete single cycle of
and ANS during meditation. For example, Takahashi and prayer movements. Participants were reminded to keep their
colleagues performed aspectral analysis for EEG band fre- eyes opened during the salat session. In the postbaseline
quencies and HRV components during Zen meditation and session, data were recorded, and the procedure was similar
discovered that the a EEG power was negatively correlated to that used for the prebaseline data recording.
with normalized unit of LF (nuLF) power as well as in LF/
HF and that h EEG power was positively correlated with Data acquisition and signal processing
normalized unit of HF (nuHF) power.21 Travis demonstrated
reduced breath rate, increased respiratory sinus arrhythmia During the experiment, EEG and ECG signals were
amplitudes, and increased EEG a amplitude compared with continuously recorded with a computer-based data acquisi-
EEG and autonomic patterns during Transcendental Medi- tion system (MP150; BIOPAC Systems Inc., Camino Go-
tation.24 Tang and colleagues reported a positive correlation leta, California). To avoid any artifacts due to physical
between h and HF component of HRV during short-term movements, only four static positions (standing, bowing,
meditation.17 No studies have assessed Muslim prayer sitting, and prostrating) were analyzed; the signals in be-
(salat), and limited information is available on the func- tween movements were excluded.
tionality of ANS in the interaction between EEG and HRV EEG was recorded with an AgCl electrode cap, with
during salat. This study aims to evaluate the possible cor- electrodes positioned on the participant’s head with the use
relations between the spectral power of the a band frequency of the standard 10–20 system. On the basis of previous
of EEG and HF or LF bands of HRV during salat and to studies, electrodes were placed at O1, O2, P3, P4, C3, C4,
elucidate the physiologic mechanisms between salat and the F3, and F4 and referenced to the linked ear lobe electrode
CNS and ANS. during recording. Electrode impedances were brought below

FIG. 1. Complete cycle of salat postures and movements.

560 DOUFESH ET AL.

5 K O. Unipolar recording technique was used to record the Table 1. a Relative Power During
signals. The signals were sampled at a rate of 1000 samples/s Pre- and Postbaseline and Salat Practice
and amplified with BIOPAC EEG100C amplifiers. As a pre-
liminary step to estimate power spectral density, all signals Mean RPa (SD) (lv2/Hz)
Scalp
were band-pass filtered between 1.0 and 100 Hz with Acq- position Prebaseline During salat Postbaseline
Knowledge 4.0 software (BIOPAC Systems Inc.). Then, a
Matlab program (MATLAB R2010a, MathWorks, Natick, O1 23.06 (8.09) 29.13 (8.59) 27.02 (9.87)
Massachusetts) was written to estimate the power spectral O2 22.40 (6.89) 27.42 (6.78) 25.36 (8.68)
density of the EEG signals. This program is based on Welch’s P3 24.03 (7.43) 28.80 (6.93) 26.72 (7.35)
averaged periodogram method with a Hanning window that P4 23.45 (5.49) 28.15 (6.83) 25.23 (8.22)
has a 1024-point fast Fourier transformation and 512-point C3 15.86 (3.57) 18.29 (4.42) 16.92 (3.03)
C4 15.18 (2,81) 17.66 (5.13) 16.47 (2.60)
overlapping for each segment length of the signal recorded. F3 10.56 (2.88) 12.00 (4.03) 11.59 (1.47)
The resulting values were normalized into the a relative power F4 10.04 (2.45) 11.66 (3.04) 10.58 (1.70)
(RPa) according to the following equation:
R fh RPa, a relative power; SD, standard deviation.
fl Sx ( f )df
RPa ¼ R f max · 100
0 Sx ( f )df
(F[2,87] = 3.39; p = .038), P3 (F[2,87] = 3.27; p = .043), P4
where fmax = 95 Hz, fl = 8 Hz, fh = 13 Hz.25 (F[2,87] = 3.50; p = .034), C3 (F[2,87] = 3.22; p = .045), C4
ECG signal was obtained with three electrodes attached to (F[2,87] = 3.36; p = .039), F3 (F[2,87] = 1.87; p = .159), and F4
the participant’s chest in a standard lead II configuration. (F[2,87] = 3.36; p = .039). Furthermore, post hoc analysis
The sampling rate of the ECG was 1000 samples/s, and the showed that there were significant differences between Salat
signal was amplified with a BIOPAC ECG100C differential and prebaseline for all electrode positions except F3, but no
amplifier. Then, the signal was band-pass filtered between significant difference between salat and postbaseline condition.
0.5 and 35 Hz before signal analysis. A Hanning window Table 2 presents the changes in HRV components during,
with a 256-point fast Fourier transformation and 128-point before, and after salat practices. The results showed that LF,
overlapping was used for Welch’s method to evaluate the HF, and nuHF significantly increased whereas nuLF and LF/
power spectral density in HRV. HRV was calculated from a HF significantly decreased during salat practice. ANOVA test
series of 5-minute epochs of ECG signal according to results showed that the differences in means of all HRV pa-
guidelines. Spectral HRV components were evaluated and rameters between prebaseline, Duha salat practice, and post-
obtained in absolute values of power (ms2) based on their baseline were significant at the 5% level ( p < .05), where VLF
frequency to one of the following three bands: VLF, LF, and (F [2,87] = 5.80; p = .004), LF (F [2,87] = 4.23; p = .018), HF
HF. The HF and LF components of HRV were conven- (F[2,87] = 9.52; p = .000), total power (F[2,87] = 7.13; p = .002),
tionally observed in normalized units (nuHF and nuLF). The nuLF (F[2,87] = 4.82; p = .010), nuHF (F[2,87] = 4.82; p = .010),
LF/HF ratio, an estimate of the balance between sympa- LF/HF (F[2,87] = 3.72; p = .028), and heart rate (F[2,87] = 5.86;
thetic and parasympathetic activities, was also calculated p = .004). In addition, post hoc analysis showed significant dif-
from the absolute power of both frequency components. ferences between salat and prebaseline but no significant differ-
ence between salat and postbaseline condition.
Statistical analysis Table 3 shows the correlation coefficients between the
RPa and HRV parameters. The results indicate that RPa was
Experimental data were analyzed with SPSS software,
significantly positively correlated with nuHF and signifi-
version 17 (SPSS Inc., Chicago, Illinois). Analysis of vari-
cantly negatively correlated with nuLF and LF/HF in the
ance (ANOVA) was used to test the changes in the means of
occipital and parietal regions of the brain during salat
the RPa and HRV variables during salat and pre- and
practice. The pre- and postbaseline conditions showed no
postbaseline. Additional comparisons were also conducted
significant correlations between RPa and HRV components.
with the post hoc test. The Pearson product-moment corre-
lation coefficient was obtained to determine the correlation
Discussion
between the HRV frequency power and RPa of the EEG
signals and also between HRV and respiration. A p-value This study investigated the EEG and ECG signals of 30
less than .05 was considered to represent a statistically young, healthy Muslim men. It aimed to explain the effect
significant difference. and the possible relationships among the relative power
spectra of the a band frequency of EEG and ANS activities
represented by the frequency bands of HRV during salat.
Results
The results (Table 1) indicated that RPa was significantly
EEG and ECG signals from all participants were analyzed. higher ( p < .05) during salat than at pre- and postbaseline. A
The average prayer duration performed in this study was 5.23 notable increase in a wave activity was observed at the
(standard deviation, 1.46) min. Table 1 shows the means and occipital and parietal regions of both brain hemispheres. The
standard deviations of RPa during pre- and postbaseline and production of awaves is normally promoted by the para-
salat practice. ANOVA tests showed that the means of RPa sympathetic nervous system with suppression of the sym-
were significantly higher ( p < .05) during salat than those of pathetic system.21 These findings strongly suggest that the
the two baselines before and after salat in most of the eight high levels of a activity during salat are associated with
electrode positions, where O1 (F[2,87] = 3.60; p = .031), O2 increased relaxation, reduced tension, sustained focus, and a
MUSLIM PRAYER AND EFFECTS ON EEG 561

Table 2. Heart Rate Variability Parameters During Pre- and Postbaseline and Salat Practice
HRV parameter Prebaseline During salat Postbaseline
2
VLF power (ms ) 217.93 (155.22) 409.37 (287.24) 310.39 (188.27)
LF power (ms2) 909.91 (296.82) 1190.70 (464.67) 1072.92 (345.13)
HF power (ms2) 538.64 (192.57) 800.15 (268.31) 683.11 (230.46)
Total HRV power (ms2) 1666.49 (568.57) 2400.26 (949.47) 2066.47 (711.99)
nuLF 62.91 (3.92) 59.23 (5.62) 61.11 (4.00)
nuHF 37.08 (3.92) 40.76 (5.62) 38.88 (4.00)
LF/HF 1.72 (0.29) 1.50 (0.35) 1.60 (0.30)
Values are mean (standard deviation).
HRV, heart rate variability; VLF, very low frequency; LF, low frequency; HF, high frequency; LF/HF, ratio of LF to HF; nuLF,
normalized unit of LF; nuHF, normalized unit of HF.

balanced condition of the human mind and body.5,6,26 The significant correlations were found with RPa in nuLF and
results were also in accordance with previous laboratory LF/HF (as indices of sympathetic tone) (Table 3). The sig-
results that demonstrated increased RPa during prostrate nificant correlations between the power spectral components
position, particularly during salat positions.27 of HRV and RPa reflect the sympathovagal balance, sug-
The novelty of this study is the marked increase in gen- gesting that parasympathetic and sympathetic nervous ac-
eral HRV components power during salat practice. Both the tivities increase and decrease during salat, respectively.
LF and HF components significantly increased at p < .05, as Furthermore, these results present an important point in the
well as the nuHF power when compared with pre- and interpretation of the physiologic mechanisms between the
postbaseline. There was a relatively stronger parasympathetic CNS and ANS among salat. These findings also agreed with
mobilization. The significant reduction in nuLF substantiates those of a recent study on meditation.21 The aforementioned
the reduced sympathetic activates because the LF band power study reported that the percentage change in a EEG power
of the HRV is mainly related to sympathetic modulation was positively correlated with nuHF and negatively corre-
when expressed in normalized units,12 and the parasympa- lated with nuLF and LF/HF.
thetic (vagal) modulation is a major contributor to HF band In conclusion, in light of these results, the increased EEG
power. The ratio of LF/HF represents the sympathovagal occipital and parietal RPa during salat suggests that salat
balance.28 However, an LF/HF ratio of 1.72 and 1.60 at pre- produces positive changes in brain function and human
baseline and postbaseline, respectively, versus 1.50 during well-being. These changes are associated with an increase in
salat practice indicates a relatively stronger parasympathetic the parasympathetic component and decrease sympathetic
mobilization. These findings were consistent with a previous component in the ANS. This combination of high para-
study that reported an increase in nuHF (as a parasympathetic sympathetic activity in nuHF power, low sympathetic ac-
index) and a decrease in nuLF (as a sympathetic index) tivity in nuLF power, and increase in the EEG RPa during
during meditation.3,4,15,21 salat practice suggest that the interaction between the cen-
During salat, positive significant correlations were found tral nervous system and ANS during salat promotes relax-
between RPa in the occipital and parietal regions and that in ation and minimizes anxiety for individuals who regularly
nuHF (as an index of parasympathetic tone), and negative practice salat.

Table 3. Correlation Coefficients Between a Relative Power

and Heart Rate Variability During Pre- and Postbaseline and Salat
HRV Parameters HRV Parameters
Variable Scalp position nuLF nuHF LF/HF Scalp position nuLF nuHF LF/HF
Prebaseline O1 0.153 - 0.153 0.111 O2 - 0.087 0.087 - 0.090
P3 - 0.170 0.170 - 0.142 P4 - 0. 120 0.120 - 0.103
C3 - 0.080 0.113 - 0.113 C4 - 0.159 0.159 - 0.143
F3 - 0.072 0.072 - 0.036 F4 - 0.024 0.024 - 0.017
During salat O1 - 0.514** 0.514** - 0.546** O2 - 0.626** 0.626** - 0.689**
P3 - 0.392* 0.392* - 0.479** P4 - 0.406* 0.406* - 0.477**
C3 - 0.281 0.281 - 0.318 C4 - 0.314 0.314 - 0.365*
F3 - 0.202 0.202 - 0.283 F4 - 0.224 0.224 - 0.329
Postbaseline O1 - 0.286 0.286 - 0.293 O2 - 0.275 0.275 - 0.279
P3 - 0.233 0.233 - 0.252 P4 - 0.222 0.222 - 0.239
C3 - 0.237 0.237 - 0.253 C4 - 0.262 0.262 - 0.283
F3 - 0.164 0.164 - 0.133 F4 - 0.151 0.151 - 0.173
*Correlation is significant at the 0.05 level (two-tailed).
**Correlation is significant at the 0.01 level (two-tailed).
 562 DOUFESH ET AL.

Acknowledgment .com/mind-bodymedicine.net/parasympatheticpathways.com/
afmbfinsmbk.html, accessed April 22, 2014.
This research was supported and funded by the Prime
15. Nesvold A, Fagerland MW, Davanger S, et al. Increased
Minister’s Department, Malaysia (project no. 66-02-03- heart rate variability during nondirective meditation. Eur J
0061/H-00000-3703), and University of Malaya, through a Prev Cardiol 2012;19:773–780.
postgraduate grant (PS107-2010A). 16. Malliani A, Lombardi F, Pagani M. Power spectrum anal-
ysis of heart rate variability: a tool to explore neural reg-
Author Disclosure Statement ulatory mechanisms. Br Heart J 1994;71:1–2.
No competing financial interests exist. 17. Tang YY, Ma Y, Fan Y, et al. Central and autonomic
nervous system interaction is altered by short-term medi-
References tation. Proc Natl Acad Sci U S A 2009;106:8865–8870.
18. Pfurtscheller G, Neuper C, Mohl W. Event-related desyn-
1. Alwasiti HH, Aris I, Jantan A. EEG activity in Muslim prayer: chronization (ERD) during visual processing. Int J Psy-
a pilot study. Maejo Int J Sci Technol 2010;4:496–511. chophysiol 1994;16:147–153.
2. Yucel S. The Effects of Prayer on Muslim Patients’ Well- 19. Delmonte MM. Physiological responses during medita-
Being [microform]. Boston: Boston University School of tion and rest. Appl Psychophysiol Biofeedback 1984;9:
Theology, 2007. 181–200.
3. Wu SD, Lo PC. Inward-attention meditation increases 20. Ivanovski B, Malhi GS. The psychological and neuro-
parasympathetic activity: a study 1based on heart rate physiological concomitants of mindfulness forms of med-
variability. Biomed Res 2008;29:245–250. itation. Acta Neuropsychiatrica 2007;19:76–91.
4. Sarang P, Telles S. Effects of two yoga based relaxation 21. Takahashi T, Murata T, Hamada T, et al. Changes in EEG
techniques on heart rate variability (HRV). Int J Stress and autonomic nervous activity during meditation and their
Manage 2006;13:460–475. association with personality traits. Int J Psychophysiol 2005;
5. Chang KM, Lo PC. F-VEP and Alpha-suppressed EEG- 55:199–207.
physiological evidence of inner-light perception during Zen 22. Brown BB. Recognition of aspects of consciousness
meditation. Biomed Eng Appl Basis Communicat 2006; through association with EEG alpha activity represented by
18:1–7. a light signal. Psychophysiology 1970;6:442–452.
6. Kasamatsu A, Hirai T. An electroencephalographic study 23. Hardt J, Kamiya J. Anxiety change through electroen-
on the zen meditation (Zazen). Psychiatry Clin Neurosci cephalographic alpha feedback seen only in high anxiety
1966;20:315–336. subjects. Science 1978;201:79–81.
7. Peressutti C, Martı́n-González JM, M. Garcı́a-Manso J, 24. Travis F. Autonomic and EEG patterns distinguish trans-
et al. Heart rate dynamics in different levels of Zen medi- cending from other experiences during Transcendental
tation. Int J Cardiol 2010;145:142–146. Meditation practice. Int J Psychophysiol 2001;42:1–9.
8. Lee MS, Kim BG, Huh HJ, et al. Effect of Qi-training on 25. Amodio P, Orsato R, Marchetti P, et al. Electroence-
blood pressure, heart rate and respiration rate. Clin Physiol phalographic analysis for the assessment of hepatic encepha-
2000;20:173–176. lopathy:comparison ofnon-parametric and parametric spectral
9. Raichur RN, Kulkarni SB, Rahul RR, et al. Effect of estimation techniques. Clin Neurophysiol 2009;39:107–15.
meditation training on pulmonary function tests. Recent 26. Lee SH, Ahn SC, Lee YJ, et al. Effectiveness of a medi-
Res Sci Technol 2010;2:16. tation-based stress management program as an adjunct to
10. Akselrod S, Gordon D, Ubel F, et al. Power spectrum analysis pharmacotherapy in patients with anxiety disorder. J Psy-
of heart rate fluctuation: a quantitative probe of beat-to-beat chosom Res 2007;62:189–195.
cardiovascular control. Science 1981;213:220–222. 27. Doufesh H, Faisal T, Lim K-S, et al. EEG spectral analysis
11. Pomeranz B, Macaulay RJ, Caudill MA, et al. Assessment on Muslim prayers. Appl Psychophysiol Biofeedback 2012;
of autonomic function in humans by heart rate spectral 37:11–18.
analysis. Am J Physiol 1985;248:H151–153. 28. Malliani A, Pagani M, Lombardi F, et al. Cardiovascular
12. Task Force of the European Society of Cardiology and the neural regulation explored in the frequency domain. Cir-
North American Society of Pacing and Electrophysiology. culation 1991;84:482–92.
Heart rate variability: standards of measurement, physio-
logical interpretation and clinical use. Eur Heart J 1996; Address correspondence to:
93:1043–1065. Fatimah Ibrahim, PhD
13. Aubert AE, Seps B, Beckers F. Heart rate variability in Centre for Innovation in Medical Engineering
athletes. Sports Med 2003;33:889–919. Department of Biomedical Engineering
14. Humphreys RB, Eagan KP. Autonomic foundations of Faculty of Engineering, University of Malaya
consciousness and mind-body medicine. In Consciousness Kuala Lumpur 50603
Research Abstracts from ‘‘Toward a science of conscious- Malaysia
ness: Tucson 2000.’’ Thorverton, UK: Journal of Con-
sciousness Studies, 2000 (in press). http://www.mbcring E-mail: fatimah@um.edu.my

View publication stats