The Effect of Mental Stress on the Non-Dipolar Components of the T Wave:
Modulation by Hypnosis
PETER TAGGART, MD, PETER SUTTON,PHD, CHRIS REDFERN,VELISLAV N. BATCHVAROV, MD, KATERINA HNATKOVA,PHD,
MAREK MALIK,PHD, URSULA JAMES, FBSCH, AND AVY JOSEPH,MSC
Objective:Mental or emotional stress-induced ventricular arrhythmias and sudden cardiac death are thought to be mediated by the
autonomic nervous system and ischemia. In the absence of ischemia, increased inhomogeneity of repolarization is thought to be
important. We tested the hypotheses that in the absence of ischemia, mental stress may modulate repolarization by changing
autonomic balance; and mental relaxation induced by hypnosis may offset the potentially adverse effects of stress on the cardiac
electrophysiology. Methods: Twelve healthy volunteers (6 male, age 18–35, mean 25 years) experienced a series of different
emotions intended to induce a wide range of autonomic response (42 test epochs) on two separate occasions, with and without
hypnosis, with continuous electrocardiogram recording. Low- (LF) and HF (high-frequency) heart rate variability was measured and
ventricular repolarization was assessed using the relative T-wave residua (proportion of nondipolar components of the T wave)
calculated for the T-onset – T peak (TWR-peak T), T peak –T end (TWR-end T), and the whole T wave (TWR). Results:
Emotionally induced changes in LF and LF/HF ratio correlated with changes in TWR, e.g., (R ! 0.51, p " .001; R ! 0.59,
p " .0001; and R ! 0.59, p " .0003, for LF/HF versus TWR, TWR-Peak T, and TWR-end T, respectively. Mental relaxation
induced by hypnosis increased LF power (1,205 ms
2
) versus 624 ms
2
, p " .003 for hypnotized versus nonhypnotized state), HF
power (1,619 ms
2
versus 572 ms
2
), p " .0004), and reduced LF/HF ratio (1.0 versus 1.5, p ! .052) and was associated with a
marked reduction in the changes in repolarization in response to emotion, e.g., 10.7 # 10
$6
versus 5.0 #10
$6
, p " .03 for TWR.
Conclusions: a) Mental stress in the absence of ischemia altered repolarization inhomogeneity via change in the autonomic balance.
b) Mental relaxation induced by hypnosis greatly reduced the effect of mental stress on repolarization. c) These findings may have
implications for arrhythmogenesis. Key words: mental stress, arrhythmia, electrophysiology, EKG, hypnosis.
EKG ! electrocardiogram; LF ! low frequency; HF ! high fre-
quency; LF/HF ! low frequency/high frequency ratio; TWR !
T-wave residua.
INTRODUCTION
I
t is well known that mental stress and the emotions of
everyday life can induce ventricular arrhythmias and sudden
cardiac death (1). Although in many cases the arrhythmia
induction is mediated by myocardial ischemia (2-5), mental
stress and emotions also can be arrhythmogenic without in-
ducing ischemia, even in coronary artery disease patients (6),
suggesting a nonischemic as well as an ischemic component.
It is generally believed that in such cases, the arrhythmia
occurrence is mediated by shifting sympathetic/parasympa-
thetic balance toward predominant sympathetic effects.
The exact mechanism by which sympathetic activity in-
duced by mental stress triggers or facilitates ventricular ar-
rhythmias in the absence of ischemia is not clear. One possible
mechanism is increased heterogeneity of ventricular repolar-
ization, which is an important predisposing factor for ventric-
ular re-entrant arrhythmias (7,8). It is known from studies in
animal models that sympathetic stimulation increases hetero-
geneity of ventricular repolarization (9). It is also known that
emotion may alter repolarization parameters that are corre-
lated with repolarization heterogeneity. For example, mental
stress enhances T-wave alternans in animals (10) and in hu-
mans susceptible to cardiac arrhythmias (11). However, no
systematic data in humans are available on the link between
autonomic balance in response to emotion and ventricular
repolarization.
In the present study, we induced a series of different
emotions intended to span a wide range of autonomic re-
sponses, i.e., from mainly sympathetic to parasympathetic
effects, on ventricular repolarization. The latter was quantified
by the nondipolar (i.e., not contained in the main cardiac
dipole, or heart vector) components of the electrocardiogram
(EKG) T wave, which are only weakly correlated with con-
ventional repolarization parameters such as QT interval and
QT dispersion (12,13), and which are increased in various
patient groups compared with healthy controls (12) and con-
tain independent prognostic power (14). Autonomic balance
was assessed using heart rate variability (HRV). In order to
minimize the possibility of inducing myocardial ischemia
during mental stress, only young healthy subjects with no
apparent heart disease were studied.
To further investigate the autonomic effects of mental
stress on ventricular repolarization, all mental stress tests were
also investigated under hypnosis, which is known to cause a
shift in sympatho-vagal balance toward overall vagal effects
(15,16).
METHODS
Subjects
Twelve healthy subjects aged 18 to 35 years (mean 25 years) were
recruited from a population of medical students, students of other disciplines,
friends and colleagues by notices, and by word of mouth. The subjects were
invited to participate in a study on “mental stress” and were blind to the
hypotheses. Equal numbers of males and females took part and all except one
From the Department of Cardiology, The Hatter Institute and Centre for
Cardiology, University College London Hospitals, London, UK (P.T., P.S.,
C.R.); The London College of Clinical Hypnosis (Medical), London, UK
(V.NB., U.J., A.J.); Department of Cardiac and Vascular Sciences, St
Georges’ Hospital Medical School, London, UK (K.H., M.M.).
Address correspondence and reprint requests to Dr. Peter Taggart, Depart-
ment of Cardiology, The Hatter Institute and Centre for Cardiology, Univer-
sity College London Hospitals, Grafton Way, London WC1E 6DB, UK.
E-mail: peter.taggart@uclh.org
Received for publication April 22, 2004; revision received November 23,
2004.
There are no financial associations that might pose a conflict of interest in
connection with the submitted article. This work was supported in part by a
grant from the British Heart Foundation (V.B., K.H.).
DOI: 10.1097/01.psy.0000160463.10583.88
376 Psychosomatic Medicine 67:376–383 (2005)
0033-3174/05/6703-0376
Copyright © 2005 by the American Psychosomatic Society

were white. The study was approved by the Hospital Ethics Committee, and
written informed consent was obtained from all participants.
Procedure
Each subject attended on two occasions at the same time of day. Two
ambulatory EKG recorders were fitted: one for analysis of HRV (Reynolds
Medical) and a second for analysis of T-wave repolarization parameters
(Marquette SEER-MC). The subjects were rested in the semirecumbent po-
sition, and a sequence of emotions was induced. Because different emotions
produce diverse sympathetic and parasympathetic effects (17-21), the test
protocols were designed to include a range of different emotions intended to
generate as wide a range of autonomic responses as possible. Each emotion
was either self-induced by recall of real life situations or induced by present-
ing pictures of emotive scenes. Six subjects (3M, 3F) experienced the emo-
tions of anger, happiness, the recollection of pain, and humor induced by
recall. Each emotional epoch lasted 5 minutes separated by 5 minutes rest
(Figure 1, protocol 1). The other 6 subjects experienced the emotions of
disgust, fear and a combination of fear with disgust (Figure 1, protocol 2).
Disgust was induced by presenting pictures depicting unpleasant scenes
accompanied by a graphic verbal description. Fear was then induced by recall
and maintained through a further presentation of pictures of unpleasant
scenes. These three epochs were consecutive, i.e., lasting a total of 15 minutes
with no intervening rest periods. Each subject underwent the same test
sequence on two occasions on separate days. On one of the two occasions, a
hypnotic state was induced by a medical hypnotist before the start of the
emotional test sequence. On the occasion when hypnosis was not performed
the subject was engaged in light conversation for 10 minutes before the
emotional test sequence (Figure 1).
Hypnosis
Hypnosis was induced by a standard technique (22) using an eye fixation
technique whereby the subject is asked to gaze at a spot of their choosing.
Direct and authoritarian suggestions for the body to grow heavy, tired, and
sleepy were given. Hypnosis was deepened with a 10 to 1 countdown
technique, and direct suggestions on hypnotic sleep were made. Communi-
cation with subjects under hypnosis was maintained through the installation of
an ideo-motor response, an unconsciously controlled finger movement to
obtain “yes” or “no” responses. The subjects were then awakened after the last
emotional epoch. Autonomic activity has been shown to vary with self-
reported estimates of stress (23). Self-rating scores were obtained for the
perceived intensity of the emotions experienced during the hypnotized pro-
tocol and the nonhypnotized protocol. The order of allocation to hypnosis first
or no hypnosis first was alternate for consecutive subjects. Hypnosis and the
stress protocols were administered by different people. As far as possible, the
intensity of the stress protocols was standardized for the hypnotized and
nonhypnotized occasions by the same duration and order of each emotional
epoch as interrogation by the same person. The participants were treated
identically on the two occasions apart from the induction of hypnosis before
the emotional tests and awakening afterward.
EKG Recordings and HRV
EKG was recorded continuously during the whole procedure using two
devices: three-channel ambulatory EKG-recorder (Delmar Reynolds Medical
Ltd.) for HRV measurements, and 12-lead digital EKG recorder (SEER MC,
GE Marquette, Milwaukee, WI) for repolarization measurements (T-wave
residua; TWR). HRV and repolarization measurements were analysed by
different people both blind to the identity of the subjects and the interventions.
R-R intervals for HRV power spectral density analysis were analysed using
the Pathfinder 700 system (Delmar Reynolds Medical Ltd.) computed for low
frequency band power (LF 0.04–0.15 Hz), high frequency band power (HF
0.15–0.4 Hz) and LF/HF ratio. Values were obtained for LF at 2 minute
intervals and for HF at 1 minute intervals (24). No clinically evident changes
in respiration occurred.
Analysis of Repolarization
The 12-lead digital recorder was programmed to acquire one 10-second
12-lead EKG every 10 seconds (i.e., continuously). From each lead of each
10-second 12-lead EKG recording the so-called median beat was constructed,
which, compared with the native EKG signal, has an improved signal-to-noise
ratio (25). Analysis of repolarization was performed on the series of median
beats. The nondipolar (i.e., not included in the main cardiac dipole, or heart
vector) components of the T wave (T wave residua, TWR) were calculated
using a previously reported method, which has been described in detail
(12,26). In brief, the method is based on singular value decomposition of the
8 independent leads of the standard 12-lead EKG (any 2 peripheral plus the
6 precordial leads) (12). Using this method, the EKG is reconstructed in
another system of 8 independent (i.e., orthogonal leads). In this system, the
first lead contains the maximum energy of the EKG single in one direction,
the second lead the maximum of the remaining energy in a direction perpen-
dicular to the first, etc. In this way, the energy of the first three orthogonal
leads corresponds to the heart vector, whereas the energy of the remaining
leads four to eight corresponds to nondipolar components. TWR quantify the
proportion of the EKG signal that cannot be explained by a single moving
dipole (the heart vector) during repolarization. They are considered to be due
to localized dipoles during the recovery phase, which are cancelled when
integrated into the overall dipole of the T wave (27). Thus, TWR represent
local variations in the action potential shape and duration that are lost
when the whole electrical activity of the heart is presented as a single
dipole (heart vector) as, for example, in orthogonal XYZ leads and
vectorcardiography. Therefore, their quantification is considered to pro-
vide a measure of the degree of localized depolarization heterogeneity
throughout ventricular myocardium Normally, TWR represent a very
small proportion of the whole ECG signal (12,28), and their increase
signifies increased depolarization heterogeneity.
TWR were calculated separately for the T
onset
–T
peak
(TWR-P), T
peak
–T
end
(TWR-E) as well as for the whole T wave (TWR). The reasons for this
were that because the nondipolar EKG content is not constant throughout
depolarization and repolarization (29), further information may be gained
from subdivisions of the T wave. In addition, many arrhythmias develop from
a reentrant circuit between endocardium and epicardium, i.e., transmurally.
Figure 1. Flowchart of the sequence of mental stress tasks. The shaded areas indicate rest periods. The times are indicated in minutes (See text).
EMOTION, AUTONOMIC RESPONSE AND REPOLARIZATION
377Psychosomatic Medicine 67:376–383 (2005)