Content uploaded by Branka Zei
Author content
All content in this area was uploaded by Branka Zei on Oct 13, 2014
Content may be subject to copyright.
Emotions, Qualia, and Consciousness, Ed: A. Kazniak. pp 278-284.
Istituto Italiano per gli studi filosofici
Series on Biophysics and Biocybernetics
Vol.10- Biocybernetics
The Role of Autonomic Balance in Experiencing Emotions
by
Branka Zei and Marc Archinard
University Hospitals of Geneva, Switzerland
Abstract
This research explores the role of the physiological component of emotional
arousal. The concept of autonomic balance is presented theoretically and
operationalized through measurement of heart rate variability (HRV). The role
of the latter is examined in its relation to emotional arousal, as reflected in both
subjective feeling and nonverbal vocal expression. Extraversion, as personality
trait, and state anxiety, are included in the experimental design. The results lend
support to the hypothesis that subjects with low HRV experience flattening of
emotional reactions mainly in vocal expression, but also in subjective feeling.
Implications of the findings are discussed in terms of the influence of HRV on
interoception, and emotional awareness.
Conceptual Issues
Emotions have been characterised as psycho-physiological phenomena that
include cognitions, visceral, humoral and immunological reactions, vocal and
other non verbal expressive displays as well as activation of behavioural
dispositions. The latter are supported by the autonomic nervous system (ANS).
So far the ANS component of emotional reactions has been mainly
addressed from the point of view of autonomic differentiation of emotions and
focused on sympathetic activation (for an extensive survey see Cacciopo, Klein,
Berntson, & Hatfield, 1993). However in such studies the question of basic non-
emotional autonomic responsivity, was not addressed. Autonomic responsivity
has more recently been conceptualised in terms of autonomic balance (Friedman
& Thayer, 1998) which is reflected in heart rate variability (HRV). The latter is
mainly driven by the parasympathetic division of the ANS. Porges, Doussard-
Roosevelt and Maiti (1994) pointed out that higher vagal tone (higher HRV) is
associated with normal emotional reactions. Poor HRV appears to be related to
a series of affective and cognitive disturbances (Friedman & Thayer 1996;
Klein, Cnaani, Harel, Braun, & Ben-Haim, 1995 ; Yeragani, Balon, & Pohl,
1990 ; Yeragani, Pohl, Srinivasan, Balon, Ramesh & Berchou, 1995; Eysenck,
1985).
In order to explore the neurological underpinnings of emotions, we considered it
meaningful to study the relationship between vagal tone as indicator of
autonomic balance, and emotional reactions as reflected in vocal arousal and
subjective feeling. We also assumed that the degree of awareness of the
subjective feeling could be linked to the degree of autonomic arousal
accompanying that state.
The influence of emotional autonomic activation on vocal behaviour was first
modelized by Williams and Stevens (1972) in terms of direct causal
relationships between dominantly sympathetic or dominantly parasympathetic
activation on the one hand and voice intensity, vocal cord vibration and timing
of speech on the other. Porges et al (1994) provided a precise description of the
link between vagal tone and vocal expression of emotion.
On the other hand, according to the vocal feedback hypothesis (Hatfield, Hsee,
Costello & Denney, 1995) whereby the degree of emotional experience is
affected by vocal feedback, we assumed that the subjective feeling of an
emotional state could be influenced by the degree of vocal arousal expressed
during that state. We defined vocal arousal as a set of speech characteristics
related to an emotional state.
Hypotheses:
1. The basic vagal tone (expressed in HRV index units) influences emotional
reactions in that the subjects with low HRV would display
a general flattening of vocal arousal and weaker vocal differentiation of
emotions
More specifically, vocal arousal being predicted as higher in anger than in
sadness (Scherer & Zei, 1988 ; Scherer, Banse, Wallbott, & Goldbeck, 1991;
Banse & Scherer, 1995), vocal differentiation of high versus low vocal arousal
states would be diminished in subjects with low vagal tone compared with those
with high vagal tone.
a flattening of subjective emotional feeling in that the subjects with lower
vagal tone would report lesser degree of subjective emotional feeling
compared with those having higher vagal tone.
2. The awareness of an emotional state would be positively correlated with the
degree of vocal arousal.
Experimental Design
Subjects:
40 diabetic patients (varying in age and duration of illness) were chosen because
they are likely to develop lesions of the ANS known as autonomic neuropathy.
One of the first symptoms of such lesions is reduced vagal tone. The latter is
reliably assessed through measurement of heart rate variability
Physiological data:
Two standard tests of the autonomic function (Vita G; Princi P. 1986) were
applied. They both measure the heart rate variability in two conditions: (1) Heart
rate difference in deep breathing and (2) Lying-to standing heart rate ratio. The
results were age adjusted and combined into a composite HRV index.
Psychological data: Speilberger state anxiety scale and Eysenck personality
inventory.
Induction of emotions and vocal data:
The subjects were asked to verbally recall their personal emotional experiences
of joy, anger, and sadness. At the end of each recall they were asked to
pronounce, on a mood congruent tone, the sentence “ALORS TU ACCEPTES
CETTE AFFAIRE” ("So you accept the deal"). The sentence was presented in
writing without punctuation so as not to suggest any tone of voice. The subjects
were then asked whether they had subjectively felt (and to what degree) the
emotion described during their recall. The results were coded on a scale: 0-3,
(ranging from "not at all" to "very much"). The subjects’ voices were recorded
on a DAT recorder . The distance of the microphone to the mouth was kept
constant.
Acoustical Analyses
One hundred and twenty samples of the standard sentence were acoustically
analysed.
Three categories of vocal arousal indicators were extracted:
1. Fundamental frequency (F0) of vocal cord vibrations computed from the
signal digitised at 44 kHz.
The following F0 parameters were extracted from the pitch curves and
expressed in Hz.
Mean, median, mode
Range between 5th - 95th percentile, 5th percentile
Maximum/minimum ratio, sd, coefficient of variation.
2. Acoustic energy computed from the raw signal values.
The following energy parameters were extracted from the amplitude envelopes
and expressed in pseudo-decibel units:
Maximum voiced energy
Mean voiced energy
Voiced energy range. The measurement was done at mid-point values of
vowel nuclei .
3. Speed of delivery expressed in the number of syllables uttered per second.
Prior to the measurement of the total signal length, all inter syntagmatic pauses
had been edited out. The speed of delivery thus corresponded more closely to
articulation speed. The latter was expected to be slower in sadness than in
anger.
All the acoustical analyses were done by means of a Macintosh platform
software “Signalyze” (Keller 1995).
Data transformations and creation of new variables:
In order to make the data directly comparable on a common scale and
eliminate sex related differences, z-scores were calculated for all vocal
parameters.
Autonomic tests results were age adjusted and normalised against external
reference values from healthy subjects (Vita & Princi, 1986). A cumulated
score on both tests was taken as HRV index for each patient.
On the basis of curve fitting and upon inspection of partial
correlations with HRV index (controlling for age, anxiety state and
extroversion), as well as linear multiple regression analyses three vocal
parameters appeared as significantly related to HRV index. These were:
F0Max/min ratio, voiced energy range and the rate of delivery. We then
calculated a summary score reflecting the overall degree of vocal arousal (Vocal
Arousal Index) for each condition (anger, joy, sadness). We justify cumulating
the three parameters into a composite score by the fact that while each of them
can vary independently, they often maintain trading relationships and appear as
a individual configuration representing the speaker’s personal way of signalling
affect (some speakers using mainly pitch parameters, others using mainly energy
parameters or speed of delivery or any combination of the three basic
dimensions of prosody).
As we expected the subjects with high HRV index to exhibit higher Vocal
Arousal Index in anger than in sadness, we then calculated the delta between the
vocal arousal index obtained in expressing anger and that obtained for sadness.
Each subject was thus characterised by his/her Vocal Arousal Delta Index
(dB+Max/MIN+rate) reflecting the degree of his/her vocal differentiation
between anger and sadness.
Statistical Analyses and Results
Vocal Arousal Data
We performed liner multiple regressions (stepwise method) with Vocal Arousal
Delta Index as dependent variable and HRV index, demographic and
psychlogical variables as independent variables.
The results of the regressions show a highly significant effect for HRV index (T
= 7.189;
P = .oooo) and a much lesser effect for state anxiety (T = -2.052; P .0470). The
HRV index alone explained 58% of data variance while the multiple R=.78782.
None of the other variables contributed significantly.
From the above results we conclude that vocal differentiation of emotions is
related above all to the HRV and marginally on the state anxiety.
Data on Self Reported Subjective Feelings
Seventy-five percent of subjects reported felt anger (mean = 1.7; sd = 1.24),
97.5 % reported felt joy (mean = 2.5; sd = .78) and 95 % of the subjects reported
felt sadness.
Mann-Whitney U tests with groups obtained by median split on HRV index
, showed significant differences in the degree of felt sadness (Z = - 3.3;
P=.0009), and anger (Z = - 2.4; P = .02). The groups with higher HRV reported
higher degree of subjective feeling for both sadness and anger than those with
lower HRV.
By contrast the correlations between Vocal Arousal Index and the degree of
subjective feeling (controlled for demographic and psychological variables) did
not reveal any significant correlation.
An unexpected finding concerned weeping episodes. Seventy seven percent
of subjects wept during the recall of sadness. The degree of weeping was coded
from 0-3 with: 0 = absence of visible weeping; 1 = noticeable tears in the eyes; 2
= tears running down the face; 3 = tears running down the face accompanied by
speaking difficulties. Correlations between the degree of crying and HRV index
(controlled for anxiety, extroversion, gender and age) were calculated. They
revealed highly significant correlations (r = .56, P = .000).
Discussion
Our hypothesis 1 was confirmed in that:
HRV (vagal tone) as indicator of autonomic balance was found to be related to
emotional arousal in that the subjects with lower HRV exhibited a flattening of
emotional reactions in two domains: vocal arousal and subjective emotional
feeling.
More specifically :
vocal differentiation between anger and sadness was smaller in subjects
with low HRV compared with those with higher HRV.
the degree of self reported subjective feeling was proportional to degree of
HRV.
As to the unexpected finding concerning the degree of weeping being
proportional to the HRV index, we had two complementary interpretations:
1. in neuropathic subjects the destruction of the parasympathetic nerves causes
diminished tearing.
2. emotional experience of sadness is lesser in subjects with low HRV.
Our hypothesis 2 was not confirmed in that the degree of subjective feeling
was not found to be related to the degree of vocal arousal.
Our results concerning the flattening of emotional reactions agree with
those of Andreasen and colleagues (Andreasen et al 1981) whose experiment
demonstrated that affective flattening is reflected in speech in a diminished
variance in both amplitude and fundamental frequency. The authors consider the
acoustic analysis of voice patterns as an objective means of evaluating flatness
of affect.
As to the results on the subjective feeling, it appears meaningful to consider
an explanation whereby higher HRV could enhance the interoception of one’s
own cardiac response to emotional stress and consequently a higher degree of
emotional awareness. Such a hypothesis would be in agreement with the
findings of Davis (Davis, M. R., Langer, A. W., Sutter, J. R., Gelling, P. D., &
Marlin, M.,1986) where the subjects with high heart rate variability displayed
more accurate perception of their own heart beat rates. In view of these findings
it appears meaningful to assume that the awareness of the strength of a
subjective emotional feeling covaries with the degree of autonomic arousal and
its proprioception. The latter thus appears to be related to HRV index as
indicator of basic non-emotional autonomic responsivity and/or autonomic
balance.
References:
Andreasen, N. C., Alpert, M., Martz, M. J. (1981) Acoustic Analysis.
Arch Gen Psychiatry, 38, 281-285.
Banse, R., & Scherer, K.R. (1996). Acoustic Profiles in Vocal Emotion
Expression. Journal of Personality and Social Psychology, 70, No 3, 614-636.
Bruchon-Schweitzer, M. (1983). C.D. Spielberger: Inventaire d'anxiété
état-trait, Forme Y. Consulting Psychologists Press Inc.
Cacioppo, J.T., Klein, D.J., Berntson, G.G., & Hatfield, E. (1993). The
Psychophysiology of Emotion. In: M. Lewis, & J. Haviland (Eds), Handbook of
Emotions (pp.119-141). New york: The Guilford. Press.
Davis, M. R., Langer, A. W., Sutter, J. R., Gelling, P. D., & Marlin, M.
(1986). Relative discriminability of heartbeat-contingent stimuli under three
procedures for assessing cardiac perception. Psychophysiology, 23, 76-81.
Eysenck, H.J. (1970). The structure of human personality. London: Methuen
Eysenck, H.J., & Eysenck M.J. (1985). Personality and individual differences.
New York: Plenum Press.
Friedman, B.H., & Thayer, J.F. (1996). Spectral characteristics of heart
period variability in shock avoidance and cold face stress in normal subjects.
Clin Autonom Res, 6, 147-152.
Friedman, B.H., & Thayer, J.F. (1998). Autonomic balance revisited:
Panic anxiety and heart rate variability. Journal of Psychosomatic Research, 44,
No 1, 133-151. Elsevier Science Publishers B.V.
Hatfield, E., Hsee, C. K., Costello, J., Schalekamp Weisman, M., Denney,
C. (1995). The impact of Vocal Feedback on Emotional Experience and
Expression. Journal of Social Behavior and Personality, Vol 10, 2, 293-312.
Klein, E., Cnaani, E., Harel, T., Braun, S., & Ben-Haim, S.A. (1995).
Altered heart rate variability in panic disorder patients. Biol Psychiatry 37, 18-
24.
Porges, S.W., Doussard-Roosevelt, J.A., & Maiti, A.K. (1994). Vagal tone
and the physiological regulation and emotion. Monographs of the Society for
Research in Child Development, 59, 167-186.
Scherer, K. R., & Banse, R. ; Wallbott, H.G., & Goldbeck, T. (1991).
Vocal cues in Emotion Encoding and Decoding. In A.M. Isen (Eds.), Motivation
and Emotion, 15, No 2, New York: Plenum Publishing Corporation.
Scherer, K.R., & Zei, B. (1988). Vocal Indicators of Affective disorders.
Psychotherapy and Psychosomatics, 49, 179-186.
Vita, G., Princi, P., Calabro, R., Toscano, A., Manna, L., & Messina, C.
(1986). Cardiovascular Reflex Tests. Journal of the Neurological Sciences, 75,
263-274. Elsevier: Science Publishers B.V. (Biomedical Division).
Williams, C. E., & Stevens, K.N. (1972). Emotions and speech: Some
acoustical correlates. J Acoust Soc Am, 52, 1238-1250
Yeragani, V.K., Pohl, R., Srinivasan K., Balon, R., Ramesh, C., &
Berchou, R. (1995). Effects of isoproterenol on heart rate variability patients
with panic disorder. Psychiatry Res, 56, 289-293.
Yeragani, V.K., Balon, R., & Pohl, R., (1990). Decreased R-R variance in
panic disorder patients. Acta Psychiatrica Scand, 81, 554-559.