ArticlePDF Available

Respiratory feedback in the generation of emotion


Abstract and Figures

This article reports two studies investigating the relationship between emotional feelings and respiration. In the first study, participants were asked to produce an emotion of either joy, anger, fear or sadness and to describe the breathing pattern that fit best with the generated emotion. Results revealed that breathing patterns reported during voluntary production of emotion were (a) comparable to those objectively recorded in psychophysiological experiments on emotion arousal, (b) consistently similar across individuals, and (c) clearly differentiated among joy, anger, fear, and sadness. A second study used breathing instructions based on Study 1's results to investigate the impact of the manipulation of respiration on emotional feeling state. A cover story was used so that participants could not guess the actual purpose of the study. This manipulation produced significant emotional feeling states that were differentiated according to the type of breathing pattern. The implications of these findings for emotion theories based on peripheral feedback and for emotion regulation are discussed.
Content may be subject to copyright.
Respiratory Feedback 1
Respiratory Feedback in the Generation of Emotion
Pierre Philippot, Gaëtane Chapelle
Université de Louvain, Louvain-la-Neuve, Belgique
Sylvie Blairy
Université du Québec à Montréal, Canada
Authors' notes
The studies presented in this paper have been made possible by a grant from the "Fonds
National de la Recherche Scientifique de Belgique" 1,5,041,94F.
The authors appreciate helpful comments of Robert S. Feldman, Ursula Hess, Arvid
Kappas and two anonymous reviewers on earlier drafts of this paper.
Correspondence regarding this paper should be addressed to Pierre Philippot who is at
Faculté de Psychologie, Université de Louvain, place du Cardinal Mercier, 10, B- 1348
Louvain-la-Neuve, Belgique. Electronic mail may be sent via Internet to
Respiratory Feedback 2
This article reports two studies investigating the relationship between emotional
feelings and respiration. In the first study, participants were asked to produce an
emotion of either joy, anger, fear or sadness and to describe the breathing pattern that fit
best with the generated emotion. Results revealed that breathing patterns reported
during voluntary production of emotion were (a) comparable to those objectively
recorded in psychophysiological experiments on emotion arousal, (b) consistently
similar across individuals and (c) clearly differentiated among joy, anger, fear and
sadness. A second study used breathing instructions based on Study 1’s results to
investigate the impact of the manipulation of respiration on emotional feeling state. A
cover story was used so that participants could not guess the actual purpose of the study.
This manipulation produced significant emotional feeling states that were differentiated
according to the type of breathing pattern. The implications of these findings for
emotion theories based on peripheral feedback and for emotion regulation are discussed.
Respiratory Feedback 3
Respiratory feedback in the generation of emotion
It is commonly agreed that emotion is best conceived of as a multi-component
process whose most central components include appraisal, facial expressions,
physiological responses, and subjective feeling states (i.e., Buck, 1985; Ekman, 1984;
Russell, 1991; Scherer, 1984). One of the oldest debates in emotion psychology
addresses the specification of the relations existing among these different components.
Historically, this debate can be traced back to William James' (1884) peripheral theory
of emotion, which stated that subjective feeling states were merely the
phenomenological result of body state. This position was vigorously counter-attacked
by Cannon (1927) who attempted to prove that body changes followed subjective
feeling states. As reviewed at the occasion of the centennial anniversary of William
James’s (1890) Principles of Psychology (Personality and Social Psychology Bulletin,
1990), this theoretical debate is far from being closed.
Presently, three main conceptions of the relationship between emotional feelings
and body states can be found in the literature. One conception--that we will label the
“undifferentiated arousal model”--states that autonomic responses increase as a function
of emotional intensity but that their pattern is undifferentiated across emotions
(Reisenzein, 1983; Schachter, 1964). At the functional level, the undifferentiated
arousal model predicts that the perception of emotional intensity can be influenced by
arousal intensity (i.e., that individuals’ perception of the intensity of their emotional
states is not only a function of their evaluation of the situation but also of the intensity
of their state of arousal). Research in this area has focused on the effect of the
manipulation of undifferentiated arousal on the intensity--but not on the quality--of
subjective feeling state. As reviewed by Reisenzein (1983) or Kirouac (1995), the
strong and consistent finding from this line of research is the intensification of the
emotional feeling state following exposure to an arousing stimulus, an effect known as
"activation transfer" (Zillmann, 1979, 1983). From this perspective, to influence
subjective feeling intensity, arousal must be (a) consciously perceived and (b)
subjectively attributed to the impact of the emotional situation.
Respiratory Feedback 4
Another conception, which we will label the “cognitive appraisal model,”
proposes that body changes in emotion are a function of cognitive appraisal (e.g.
Scherer, 1984; Pecchinanda & Smtih, 1996; Smith & Kirby, 2000) or of the direct
output from appraisal, action readiness (Frijda, 1986). More precisely, this model
suggests that the activation of a specific appraisal dimension would induce specific
body changes. For instance, novelty appraisal would induce a pause in breathing and a
deceleration followed by an acceleration of heart rate. The pattern of body changes with
a specific emotional state would be the sum of the changes induced by each appraisal
component. As different feeling states result from different appraisal patterns, they are
also characterized by different patterns of body changes. In the cognitive appraisal
model, the role played by body changes in the elicitation of feeling state is less
explicitly stated and definitely marginal, the central role being played by cognitive
A third conception, which we will label the “central network model” states that
emotions are centrally organized by neural or cognitive networks that connect the
different emotion components together. For some, these networks are innate neural
structures (e.g., Ekman, 1999; Izard, 1979; Tomkins, 1980); for others, they are
cognitive networks or schemata that develop as a function of individuals’ experiences
(e.g., Lang, 1979, 1984; Philippot & Schaefer, 2001; Teasdale, 1996). Though the
theories we gather under the central network model differ in many ways, they all share
similar features with respect to the patterning and function of body changes in emotion.
First, they all postulate that patterns of body changes are differentiated according to the
type of emotion experienced, even if cognitive network theories predict more
idiosyncratic patterns than theories postulating innate neural structures. Second, they
posit that the activation of the body state typical of an emotion elicits that emotion, a
process known as peripheral feedback. Third, they suggest that peripheral feedback
occurs automatically, at an implicit level (i.e., without awareness of the process;
Cacioppo, Berntson & Klein, 1992; Damasio, 1994; Teasdale, 1996). The implication
Respiratory Feedback 5
of these three postulates is that a specific emotion could be induced by manipulating
one’s body state, outside of this person’s awareness of the process.
At the empirical level, the central network model is supported by research
addressing the relation between facial expression and subjective feeling states, the
research area investigating the "facial feedback hypothesis." A wealth of evidence has
documented that manipulating facial expression affects feeling state (Laird, 1984;
Manstead, 1988; Matsumoto, 1987; McIntosh, 1996). The effect size of the so-called
facial feedback is generally small (around 13% of explained variance) but reliably
significant. The impact of facial muscle manipulation has been extended to
physiological changes, such as heart rate or skin temperature (Hess, Kappas, McHugo,
Lanzetta & Kleck, 1992; Kappas, 1989; Levenson, 1992; Levenson, Carstensen,
Friesen, and Ekman, 1991; Levenson, Ekman & Friesen, 1990). Further, Stepper and
Strack (1993) have documented that manipulating posture also has an impact on
subjective feeling states and affects later judgment of valenced material, extending
previous findings from Duclos, Laird, Shneider, Sexter, Stern, and VanLighten (1989)
showing that posture affects mood. Overall, there exists empirical evidence that
manipulating facial or postural muscles affects subjective feeling states, and possibly
physiological states, outside of individuals’ awareness of the process. Still, the
underlying mechanism of facial feedback is the object of a controversy (e.g., Izard,
1990; Laird, 1984; McIntosh, 1996) and the interpretation of the effect of facial
manipulation on physiological changes in terms of facial feedback has been questioned
(e.g. Boiten, 1996; Zajonc & McIntosh, 1992).
We propose that three questions must be addressed to further our understanding
of the relationship between body state and subjective feelings: First, are subjectively
differentiated feeling states characterized by different body states? Second, does body
state influence the intensity and/or the quality of subjective feelings? Third, does this
influence operate implicitly (i.e., outside of individuals’ awareness of the process), or
explicitly? This latter question is particularly important as explicit knowledge about
body changes may not relate to actual body changes (Rimé, Philippot & Cisamolo,
Respiratory Feedback 6
1990; Philippot & Rimé, 1997) and as several recent theories issued from the central
network model are based on a distinction between implicit and explicit processing (e.g.
Damasio, 1994; Teasdale, 1996; for a review see Philippot & Schaefer, 2001). The
undifferentiated arousal model postulates that emotional body states are undifferentiated
and have an explicit impact on feelings intensity but not quality. The central network
model posits that emotional body states are differentiated and influence the quality of
feeling states and that these effects can occur at an implicit level. The cognitive
appraisal model is vague and unspecific about whether and how changes in
physiological patterns might influence emotional feeling states.
At the empirical level, the first question--the peripheral differentiation of
emotion--is still the object of a controversy, with different reviews reaching different
conclusions (Levenson, 1992; Stemmler, 1992; Zajonc & McIntosh, 1992). With
respect to the second and third questions, the activation transfer research mentioned
above has shown that arousal can explicitly affect the intensity of feeling state,
providing that it is attributed to an emotional cause; the facial and postural feedback
research has demonstrated that muscular feedback can influence subjective feeling state
outside of individuals’ awareness of the process. Yet, despite its theoretical importance,
the possibility of a visceral--and not solely muscular--feedback on the quality of
emotional feelings has not been pursued at an experimental level. From this
perspective, physiological state should be manipulated, not in intensity but in quality, in
order to observe the impact of such manipulations on the nature rather than the intensity
of feeling states. Further, it should be established whether this effect occurs without
individuals explicitly using body state as a source of information to determine their
emotional feeling state. The present studies aim at exploring these neglected questions.
Specifically, they will examine the effects of respiration manipulation on
emotional feeling state. We choose to manipulate body state through respiration for
four reasons. First, like facial musculature, breathing is under both voluntary and
automatic control, allowing for the same types of manipulation as the ones commonly
used in facial feedback research. Second, clinical evidence has repeatedly confirmed
Respiratory Feedback 7
relations between breathing and at least one feeling state, anxiety (e.g. Beck & Scott,
1988). Moreover, these studies have also demonstrated the clinical efficacy of
respiration manipulation in reducing anxiety through breathing retraining. Third, a
recent literature review suggests that reliable respiratory differences might be found
between positive or negative feeling states, as well as tense or relaxed feeling states
(Boiten, Frijda & Wientjes, 1994). Fourth, while they can easily be achieved non
invasively, respiration changes affect many physiological responses, such as cardio-
vascular changes or skin conductance. They thereby constitute an easy but potent
avenue to manipulate the whole physiological state of the organism.
In order to induce specific emotions through respiratory manipulations, one has
to determine which are the respiratory patterns that correspond to these emotions. One
source of such information is constituted by the results of psychophysiological research
that observed respiratory changes during emotion induction. Boiten et al. (1994) have
reviewed the psychophysiological literature pertaining to respiratory changes occurring
during emotion. They note that there is very little empirical work on the topic and that
the available studies are fraught with methodological shortcomings. As a result, they
conclude that “it is difficult to draw specific and detailed conclusions concerning the
influence of emotion upon respiration” (Boiten et al., 1994, p. 119). Yet, they were able
to identify some consistency across studies that can be summarized as four types of
breathing related to emotional state. Fast and deep breathing was associated with
excitement, such as in anger, fear, or sometimes even joy. Rapid shallow breathing was
typical of tense anticipation, including concentration, fear, and panic. Slow and deep
breathing was most often observed in relaxed resting state. Finally, slow and shallow
breathing was associated with states of withdrawal and passiveness, such as depression
or calm happiness.
As can be seen from these descriptions, few of the four types of breathing
identified can be unequivocally associated with a specific emotional state. Two remarks
concerning these observations are of importance. First, that psychophysiological
research has not identified specific breathing patterns differentiating basic emotions
Respiratory Feedback 8
does not necessarily means that such patterns do not exist. This lack of definitive
evidence maybe due to under-sophisticated recording of breathing that most often
measures only frequency (see the discussion by Boiten et al. [1994] in this respect) or to
problems in inducing emotion in psychophysiological experiments (see the discussion
of this point by Stemmler [1989]). Second, although no unequivocal correspondence
between breathing patterns and emotions can be drawn, these descriptions suggest that
(a) fast and deep breathing might characterize anger, (b) rapid shallow breathing might
characterize fear, (c) slow breathing whether deep or shallow might characterize a state
of relaxed happiness, and (d) slow and shallow breathing might characterize sadness.
These propositions have to be considered cautiously, as, for instance, pattern (a) might
also characterize joyful or fearful excitement and pattern (d) might also characterize
calm happiness. In sum, although suggestive, the results of Boiten et al.’s review do not
provide an empirical basis precise enough to construct breathing instructions specific to
basic emotions.
Another source of information regarding emotional breathing patterns pertains to
previous studies investigating whether breathing manipulations induce differentiated
emotional feeling states. The only existing work as been conducted by Bloch,
Lemeignan, & Aguilera-T (1991) who propose that each of six emotions they qualify as
basic (joy, sadness, anger, fear, erotic love, and tenderness) is characterized by a
specific facial, postural and respiratory pattern they labeled “emotional effector
pattern.” The respiratory components of the patterns were derived from visual
inspection of polygraphic records of either actors expressing specific emotions or
participants relieving these emotions under deep hypnosis (Bloch [1994, personal
communication], the original report [Bloch & Santibanez, 1972] cannot be obtained).
The patterns obtained with this procedure differ in several respects from the
observations of Boiten et al (1994). Bloch’s et al. (1991) joy pattern, referring to
laughter, (quick and deep nasal inspiration, followed by oral expiration with small jolts)
is quite different from the pattern associated with calm happiness by Boiten, although it
presents some similarities with Boiten’s pattern of excitement, but for inspiration only.
Respiratory Feedback 9
Bloch’s et al. anger pattern (regular, quick, and deep nasal breathing) corresponds to our
prediction derived from Boiten et al. With respect to fear, Bloch et al. propose a quick
and shallow inspiration through the mouth, followed by a pause and a long expiration.
In contrast, our prediction based on Boiten et al. is that fear is characterized by shallow
and fast breathing. Finally, Bloch’s proposition for sadness (quick nasal inspiration
with jolts followed by a quick expiration through the mouth) does not correspond to our
prediction derived from Boiten et al. (i.e., slow and shallow breathing).
In their work, Bloch et al. (1991) have demonstrated that extensively training
actors to reproduce these emotional effector patterns results in the induction of the
corresponding emotion feeling state. Unfortunately, in these studies, participants were
explicitly (a) told that the aim of the training was to produce emotion through
respiratory, facial, and postural changes, and (b) informed of which emotional effector
pattern was intended to induce which emotion. Hence the effect of these manipulations
on feeling states may simply be the result of experimenter's demand. Further, breathing
was not manipulated independently from posture or facial expression, hence preventing
estimation of the specific impact of each source of peripheral feedback.
In order to confront the divergence between Bloch’s patterns and those
suggested by Boiten, and given the fact that Bloch et al. (1991) report positive results,
we examined in a preliminary study whether the respiratory component of their
emotional effector patterns is in itself sufficient to induce a specific emotional feeling
state. We replicated the study of Bloch et al. (1991) for four emotions (joy, anger, fear,
and sadness) with two major changes. First, in order to avoid experimental demands,
we used a procedure in which participants were oblivious to the fact that the actual topic
of the study was emotion or emotion induction. Second, in order to disentangle the
effect of facial and postural feedback from those of a possible respiratory feedback, we
only manipulated breathing, keeping facial expression and posture constant. The results
revealed that participants tended to report the target emotion in the joy and anger
breathing condition1, F(9,153) = 1.97, p<.10.
Respiratory Feedback 10
In sum, the findings of Bloch et al. (1991) are only partially replicated in the
breathing conditions of joy and anger. There are three possibilities accounting for these
weak results. First, respiratory feedback may have no effects on emotional feeling
states. Second, it is possible that respiratory feedback alone is not a sufficient condition
to induce emotional feeling state; It may additionally require the corresponding facial
and postural pattern. Along this line, Bloch (personal communication, July 16th, 1994)
argues that not only the whole respiratory, facial and postural pattern needs to be
activated, but also that no emotion can be induced if the “correct “ pattern is not exactly
reproduced. (This notion of correctness has also been debated in the context of the
facial feedback hypothesis, see for instance, McIntosh [1996].) Finally, a third
possibility is that the respiratory instructions used by Bloch et al. (1991) are not the
most appropriate to induce discrete emotions. Indeed, it is unclear how Bloch et al.’s
respiratory patterns were originally established (i.a., no statistical analysis are reported
and the original report [Bloch & Santibanez, 1972] cannot be obtained). Further, the
breathing patterns for joy attempts to mimic laughter, while the one for sadness attempts
to mimic crying. While expressive emotional components such as laughter or tears tend
to be associated with joy and sadness respectively, this does not necessitate that they
determine the breathing patterns associated with these emotions.
Study 1
Study 1 was designed to investigate whether different, more precise and accurate
breathing instructions than those used by Bloch et al. (1991) could be established. It
consisted in explicitly asking participants to generate emotional states and to identify
and report the corresponding breathing patterns. These subjective reports were to be
compared with the results of Boiten’s et al. (1994) review of the studies that
investigated objective respiratory parameters. Our expectations were that (a) the
information obtained from subjective reports would offer more details and a greater
differentiation among emotions than the information issued from Boiten’s et al. review,
and (b) the subjective reports would be concordant with the breathing patterns derived
from Boiten et al., the latter thus validating to some extent the former. The rationale
Respiratory Feedback 11
was that if, and only if, predictions (a) and (b) were met, Study 1’s results could provide
breathing patterns potentially able to induce specific emotions. However, this
possibility would be void if either of the two predictions were not met.
Participants were invited to produce four emotional feeling states (of joy, anger,
fear, and sadness), following a procedure adapted from the one described by Hess et al.
(1992). When participants felt that they had reached the desired state, they were invited
to describe their breathing in a questionnaire investigating several respiratory
Participants and procedure
Eleven female and 12 male students volunteered to take part in the study. They
were aged between 18 and 29 years (mean age = 23.8) and they participated individually
in the experiment. The experimenter told them that the purpose of the study was to
investigate how emotions could be expressed via respiratory patterns. They were
simply instructed to produce an emotion--either joy, anger, fear, or sadness, in a random
order--by modifying their respiration. They were also encouraged to maximize the
intensity of their emotions and they were told that they could help themselves with
personal memories or fantasies. Participants performed the experimental trials standing
up alone in a laboratory room. The experimenter was in an adjacent room and contact
was maintained with an interphone system. When participants judged themselves to
have reached their best production of the target feeling state, participants reported in a
questionnaire the characteristics of the specific respiratory pattern they had performed to
express the emotion and, on a 7-point scale, the degree to which they felt they were
successful in producing a breathing pattern corresponding to that emotion.
As a manipulation check, participants were also asked to report on the French
version of the Differential Emotion Scale (Philippot, 1993) the intensity of the emotion
feeling states they had experienced during the trial (from 0: no emotion at all to 6: the
most intense emotion possible). This scale included the following items: concentrated,
joyful, sad, angry, afraid, anxious, disgusted, scornful, surprised, ashamed, guilty and
Respiratory Feedback 12
happy. Only the six relevant items were retained for the data analysis (joyful, sad,
angry, afraid, anxious, and happy).
Respiration questionnaire
Based on a pre-test study, a questionnaire investigating several respiratory
parameters was constructed. Participants were asked to describe their inspiration and
their expiration separately on five items: Was their respiration diaphragmatic, thoracic
or both? Did they breath through their nose, their mouth or both? Did the frequency
change (from “-3” = much slower to “3” = much faster), did the amplitude change (from
“-3” = much more shallow to “3” = much deeper), and did they pause (from “0” = not at
all to “4” = a lot)? Additional questions were asked for the whole respiratory pattern:
Were there sighs, tremors, or tensions in the thorax (from “0” = not at all to “4” = a lot),
and did the regularity of the respiration change (from “-3” = much more irregular to “3”
= much more regular)?
Manipulation check
On average, participants reported that they felt successful in producing
emotional breathing patterns (mean success = 3.74 with “0” = unsuccessful trial, “3” =
rather successful trial, and “6” = perfect trial). Yet, a MANOVA with emotion
condition as a within-subject factor revealed that reported success varied according to
emotion, F(3,20) = 4.99, p < .01. Post-hoc analyses indicated that the joy respiratory
pattern (mean = 4.48) was easier to produce than patterns of sadness, fear, and anger
(respective means = 3.61, 3.56, 3.30).
Not only did participants indicate that they were successful in producing
breathing patterns subjectively related to the target emotion, but they also reported
feeling the corresponding subjective state. Indeed, a 4 X 6 MANOVA with emotion
condition and emotion item of the DES as within-subject factors and sex as a between-
subjects factor revealed main effects of emotion and emotion item, respectively, F(3,63)
= 3.99, p < .02, F(5,105) = 6.93, p < .0001, that were qualified by an emotion X
emotion item interaction , F(15,315) = 34.01, p < .0001. The pattern of the results and
Respiratory Feedback 13
the post-hoc analyses represented in Table 1 clearly demonstrate that the manipulation
induced specific emotional feeling states of a significant intensity.
Insert about here Table 1
Respiratory patterns
The central question of the present study was whether people can report
respiratory patterns that differentiate among each basic emotional feeling state. Thus, to
investigate the effect of emotion condition on the different parameters of inspiration and
expiration, 2 X 4 MANOVAs were computed with inspiration-expiration and emotion
as within-subjects factors. For the frequency and amplitude parameters, only the effect
of emotion was significant, F(3,20) = 29.22, p < .0001, and F(3,20) = 16.54, p < .0001,
respectively. For the pause parameters, only the interaction between inspiration-
expiration and emotion reached significance, F(3,20) = 3.70, p < .03. Post-hoc analyses
detailed these effects. As shown in Table 2, respiratory frequency increased for anger
and fear, decreased for joy and did not change from baseline level for sadness.
Respiratory amplitude increased dramatically in joy and, although to a lesser extent, in
anger. For fear and sadness, amplitude remained at baseline levels. For pauses, post-
hoc analyses revealed that the interaction was accounted for by the fact that while
people reported more pauses after expiration in joy, F(1,22) = 4.80, p <.04, they
reported less pauses after inspiration in fear, F(1,22) = 3.61, p <.07.
Insert about here Table 2
The effects of emotion on regularity, sighs, tremors, and thoracic tension were
examined with a single factor (emotion) MANOVA. It appeared that all these
parameters were significantly modulated by the type of emotion produced, F(3,20) =
4.09, p < .02, for sighs; F(3,20) = 25.30, p < .0001, for tremors; F(3,20) = 13.06, p <
.0001, for regularity; and F(3,20) = 45.89, p < .0001, for thoracic tension. Post-hoc
Respiratory Feedback 14
analyses specified these effects. As indicated in Table 2, sighing is specifically
associated with sadness. In joy, respiration is more regular and presents much less
thoracic tension than in anger and fear. Sadness falls in between this pattern and is
characterized by tremors, which are totally absent in joy and moderately present in
anger and fear.
Finally the impact of emotion on whether the respiration was oral or nasal, and
whether it was diaphragmatic or thoracic was examined using separate χ2 for
inspiration and expiration in each breathing condition. As displayed in Table 2, a
majority of participants judged the respiration to be nasal for joy and sadness,
respectively, χ 2 = 40.29, p < .001, χ 2 = 29.82, p < .001, for inspiration and expiration
in joy, and χ 2 = 34.77, p < .001, χ 2 = 25.39, p < .001, for inspiration and expiration in
sadness. Respiration also tended to be nasal in anger, although to a lesser extent, χ 2 =
5.82, p < .10, for inspiration and χ 2 = 10.76, p < .01, for expiration. No significant
trend appeared for fear. As regards the diaphragmatic or thoracic quality of the
respiration, participants reported that expiration was predominantly diaphragmatic in
anger and thoracic in fear, χ 2 = 6.61, p < .05.
Study 1 yielded three important findings: First, respiratory patterns that are
differentiated among basic emotions were established on the basis of subjective reports;
Second, these subjective patterns are congruent with the objective patterns reviewed by
Boiten et al. (1994); Third, the explicit manipulation of respiration combined with
imagery induced significant and specific emotional feeling states.
Regarding the first finding, the consistency of naive participants in their
association between type of breathing pattern and specific emotion is remarkable.
Previous research has already shown that people report experiencing different body
sensation profiles for different emotions (Lyman & Waters, 1986; Rimé et al., 1990;
Philippot & Rimé, 1997). Yet these studies had all considered a rather global
perception of body changes (e.g. respiratory changes were measured on a single
“respiratory change” item) and no precise body changes had been explored as
Respiratory Feedback 15
specifically as in the present study. Thus, previous findings can be extended to note that
people experience a very fine differentiation of body state during emotion, not only for
the body as a whole but also for very specific changes, at least including breathing
changes. In addition, these body sensations are quite homogenous across individuals
and differentiated across emotions. These observations are contradictory to the
undifferentiated arousal model notion of diffused perception of undifferentiated arousal
inherited from Schachter’s (1964) theory. They are congruent with cognitive appraisal
models and central network models such as the Somatovisceral Afference Model of
Emotion (SAME) proposed by Cacioppo et al. (1992).
With respect to the second finding, Study 1 participants’ reports do not
contradict the findings of Boiten et al. (1994). In the present study, joy is associated
with regular, moderately deep and slow breathing through the nose and with minimal
thoracic tension, tremors, and sighs. The breathing tends to be diaphragmatic or both
thoracic and diaphragmatic. This pattern is parallel to the slow and deep breathing
Boiten et al. (1994) observed in a relaxed resting state. Yet, these authors report that
calm happiness (as well as depression) is associated with slow but shallow breathing,
whereas excited joy (as well as anger or fear) is associated with fast and deep breathing.
Bloch’s et al. (1994) joy pattern (quick and deep nasal inspiration, followed by oral
expiration with small jolts) is different from the pattern associated with joy by the
participants of the present study as well as from the three patterns associated with
positive states by Boiten et al. As mentioned above, Bloch’s et al. joy breathing pattern
attempts to imitate laughter and might not be typical of joy.
For anger, participants in the present study reported a rather fast, irregular and
deep nasal breathing with marked thoracic tension, minimal sighs, and some tremors.
The expiration was diaphragmatic. This pattern corresponds to the fast and deep
breathing Boiten et al (1994) associated with excitement, including angry excitation. It
also parallels to some degree Bloch’s et al. anger pattern (regular, quick, and deep nasal
breathing), except that our participants reported irregular rather than regular breathing.
Respiratory Feedback 16
With respect to fear, our participants reported fast, irregular, rather shallow
breathing, with much thoracic tension, some tremors, and minimal sighs. More thoracic
breathing was reported for fear than for any other emotions. This pattern corresponds
very well to the rapid, shallow breathing associated with tense anticipation by Boiten et
al. It has also basic features in common with Bloch’s et al. fear pattern. Yet, the latter
has additional specifications not reported by our participants : for Bloch et al. the
respiration has to be oral and there must sometimes be a long expiration.
Finally, for sadness, our participants reported nasal breathing with average
amplitude and frequency, marked with sighs and tremors as well as some thoracic
tension and irregularity. Of the four types of breathing proposed by Boiten et al., the
present pattern is closest to the slow and shallow breathing associated with state of
withdrawal and passiveness. It shares some similarities with Bloch’s et al. sadness
pattern (inspiration with brief jolts through the nose and expiration at one time through
the mouth), specifically, normal frequency and amplitude, but also marked
dissimilarities, including oral expiration, jolts in the inspiration and expiration in one
time through the mouth for Bloch et al. As mentioned above, Bloch’s et al. sadness
pattern attempts to imitate crying and might not be specific to sadness.
In summary, as predicted, the emotional breathing patterns reported by the
participants of the present study are characterized by a clear and detailed differentiation
among emotions. Moreover, they are congruent with the results of Boiten’s et al.
(1994) meta-analysis. This suggests that, in their attempts to produce emotional states
by manipulating their respiration, our participants have relied on breathing patterns that
are similar to observations of psychophysiological studies investigating respiratory
changes during emotion induction. As the conditions of clear differentiation among the
four emotions investigated and congruence with Boiten’s et al meta-analysis are met,
the data of the present study can provide a valid basis for the construction of different
sets of breathing instructions that would be specific to the discrete emotions of joy,
anger, fear, and sadness.
Respiratory Feedback 17
The third finding of the present study is precisely related to emotion induction.
Indeed, the analysis of the emotional feeling state questionnaire revealed that specific
and rather intense emotions have resulted from the explicit instruction to produce
emotion by manipulating respiration. This observation is in line with the report of Hess
et al. (1992) that people have the ability to produce rather intense and specific emotions
“on demand.” Future research should examine whether the instruction to alter one’s
breathing adds a specific contribution to voluntary production of emotion. Of course,
effects of experimenter demand can certainly not be completely discounted, although,
during debriefing, participants reported that they had experienced genuine emotions.
Similarly, the relative influence of other strategies, such as relying on personal
memories, cannot be assessed in the present experiment.
Still, the present findings suggest that an explicit manipulation of respiration
might be sufficient to induce a specific emotional feeling state. However, to test this
assertion, the effects on feeling states resulting from the manipulation of respiration
needs to be observed in a context free of experimental demand and in which other facets
of emotion responses are kept constant.
Study 2
Study 2 examined whether specific emotional states could be induced by
manipulating participants’ breathing patterns with instructions based on Study 1's
results. In addition, Study 2 investigated whether this effect could occur implicitly, this
is, without participants explicitly using breathing changes to infer their emotional
feeling state. Participants were told that they were participating in a health psychology
experiment aimed at examining the impact of breathing style on cardio-vascular
characteristics. After a training session, they performed the four breathing patterns that,
unknown to them, were characteristic of joy, anger, fear, and sadness. Their feeling
state was recorded by disguised items hidden in a questionnaire supposedly addressing
the body symptoms induced by the breathing patterns.
Participants and procedure
Respiratory Feedback 18
Twenty one female and 5 male students aged between 17 and 23 years (mean
age: 19.2) volunteered for the study. They participated individually in the experiment
which consisted of two sessions of 45 minutes separated by a minimum of one night
and a maximum of 48 hours. Participants were trained to perform the procedure during
the first session and the actual data collection took place during the second session.
During the first session, the experimenter explained the cover story. Participants
were told that the study had been designed to investigate the effects of breathing on
cardio-vascular changes and on physical feelings. Participants were told how
respiration could influence the cardio-vascular system at a functional and at a
mechanical level. They were told that the hypothesis was that these effects could also
influence subjective physical sensations. Then, the experimenter explained the
procedure. The experiment consisted of four trials. Each trial was preceded by a short
relaxation period during which participants had to close their eyes, breathe smoothly,
relax every muscle, and visualize an imaginary circle inflating and deflating at the
rhythm of their respiration. After relaxation, participants were to perform a respiratory
pattern for two minutes and, immediately after, to complete a questionnaire on physical
sensations. The experimenter explained that various respiratory and cardio-vascular
measurements would be taken during the breathing exercises. He showed the
transducers (a respiratory belt and the FinaPress sensor of the Ohmeda 2300 blood
pressure monitor2) and explained how this equipment operated.
Once the procedure was explained, the experimenter gave the breathing
instructions, showed how to perform them and gave feedback to the participant about
his or her performance. After having ascertained that the participant understood the
breathing instructions, the experimenter affixed the transducers and went to the adjacent
technical room. Communication with the participant was maintained throughout the
experiment via an intercom system. After calibration of the physiological
measurements, the rehearsal of the procedure began. The experimenter gave the
relaxation instructions, then reminded the participant of the breathing instructions, had
the participant perform them for two minutes, and asked to the participant to fill in the
Respiratory Feedback 19
questionnaire. During the breathing trial, the experimenter could monitor on a
computer screen the respiratory movements of the participant and check whether the
instructions were correctly followed.
When they arrived for the second session, participants were reminded of the
procedure. Then the experimenter affixed the respiratory belt and the FinaPress sensor
and went to the technical room. The four trials were performed in a random order.
Respiration was recorded during relaxation and trial periods. Finally, participants were
debriefed and the actual purpose of the experiment was explained. They were
specifically asked whether they suspected that the experimenter attempted to modify
their emotional state by manipulating their respiration. No participants reported any
suspicion about the real purpose of the experiment, about the fact that it concerned
emotion induction, or about the fact that the questionnaire measured their emotional
feeling state. Thus, if an effect on feeling state is observed, it can be considered as
occurring outside of the participants’ awareness of the process, i.e. the awareness of a
relationship between breathing and feeling state.
The questionnaire consisted in 22 items comprising different sensations. Items
were "vertigo," "nausea," "paresthesia," "lump in the throat," "headache," "impression
of unreality," "stomach sensations," "feeling cold, shivering," "feeling hot," "racing
heart," "muscular tension," "perspiration," "goose flesh," "blushing," "weak knees," and
"general activation." Mixed among these items, four scales indexed emotional feeling
states: "feelings of fear, anxiety,” for fear; "feelings of sadness, depression," for
sadness; "positive feelings, good spirit," for happiness; and "feelings of aggressivity,
aggravation," for anger. Each item had to be rated by marking a check on a 10
centimeter line, anchored 0% to 100%. Participants were instructed that 0% reflected
no such sensation at all, while 100% was the strongest sensation they could imagine
feeling for this item. The dependent measures consisted of millimeters from the zero-
point on each scale.
Physiological measures
Respiratory Feedback 20
Respiration was recorded by an elastic tube strapped around the participant's
chest. A sound of 575 Hz emitted at one end of the tube is received at the other end.
The phase of the sound received varies according to the length of the tube which is itself
determined by the respiratory movements of the ribcage. A coupler monitors these
phase changes and outputs a signal varying in tension as a function of tube length (1.2
cm/V). Technical aspects of this system are described in van Rossum (1988). The
signal of the coupler was sampled at a frequency of 10Hz by a Computer-based
Oscillograph and Data Acquisition System (CODAS) of Dataq Instruments. Codas,
which consists of a combination of hardware and software, allows continuous data
through-put to hard disk while maintaining a real time display directly on the host
computer's monitor. In addition, after the acquisition, data can be displayed on the
monitor for artifacts inspection.
Respiration indices were derived from the raw signal of the strain-gauge (Boiten,
1993). The computer program used to that effect (Philippot & Philippot, 1991) outputs
for each respiratory cycle: its length, amplitude, ratio of inspiration and expiration
times, number of pauses and their length, and number of hampers.
Breathing instructions
The breathing instructions were derived from the results of Study 1:
Joy: "Breathe and exhale slowly and deeply through the nose; your breathing is
very regular and your ribcage relaxed."
Anger: "Breathe and exhale quickly through the nose; slightly deeper than
regular breathing amplitude. Your breathing is slightly irregular with some tremors and
your ribcage is very tense."
Fear: “Breathe and exhale quickly from the top of your ribcage; with a normal
amplitude. Your breathing is slightly irregular with some tremors and your ribcage very
Sadness: “Breathe and exhale through the nose with a normal amplitude and
pace. Your ribcage is slightly tense, and there are some sighs in your expiration.”
Results and discussion
Respiratory Feedback 21
First, analyses were conducted to ascertain that participants' actual breathing
patterns differed across conditions. MANOVAs with breathing condition as a within
subject factor were computed on the differences scores (mean during the trial minus
mean during relaxation) for the indices of frequency, amplitude and ratio of inspiration
and expiration times. As can be seen in Table 3, the effect of breathing condition was
clearly significant for each index. Post-hoc analyses using the Bonferroni procedure
revealed that participants followed the instructions (see subscripts in Table 3).
Respiration time was longest during the joy condition, slightly shorter for the sadness
condition and much shorter in the anger and fear conditions, with fear respiration being
slightly faster than anger respiration. The amplitude increased in the joy condition,
remained at baseline levels for the anger and sadness conditions, and was shorter during
fear. Finally, the ratio of inspiration and expiration times increased for anger, fear, and
sadness but stayed at baseline level in joy.
Insert Table 3 about here
Second, the impact of breathing condition on emotional feeling state was
examined. A MANOVA with breathing condition and feeling scale as within-subject
factors was computed on the measures of the four feelings. A significant effect of
breathing condition indicated that, overall, some breathing patterns induced more
intense feeling state than others, F(3,23) = 9.02, p<.0004. Similarly, some feeling states
tended to be reported as more intense than others, as indicated by a significant effect of
feeling scale, F(3,23) = 5.89, p<.004. Of direct interest for our hypothesis, a significant
interaction indicated that feeling state varied according to breathing conditions, F(9,17)
= 8.73, p<.0001; This effect accounts for 40% of the variance.
Post-hoc analyses specified the impact of breathing condition on feeling state
(see Table 4). MANOVAs with feeling scale as within-subject factor were computed
for each breathing condition. The effect of feeling scale was significant for each
condition, indicating that each breathing condition induced a differentiated feeling state,
Respiratory Feedback 22
F(3,22) = 13.32, p<.0001, for joy; F(3,22) = 7.20, p<.001, for anger; F(3,22) = 5.71,
p<.004, for fear; and F(3,22) = 4.10, p<.02, for sadness, respectively. These effects
were specified with paired t-test using the Bonferroni procedure. As can be seen in
Table 4, the joy breathing condition induced significantly more positive feeling than any
other condition and more than any other feeling within this condition. The same is true
for the feeling of anger in the anger breathing condition. It should be noted that this
breathing pattern also induced feelings of fear and anxiety, although to a lesser degree
than anger feelings. The fear breathing condition induced feelings of anger and of
fear/anxiety at a similar intensity level. Yet, the feelings of fear/anxiety induced in this
condition are not more intense than those induced by any other conditions. Finally, the
sadness breathing condition induced to a comparable extent positive feelings and
feeling of sadness. It should be noted that it is in this condition that feelings of sadness
were the most intensely reported, as the three other breathing conditions induced no
feeling of sadness at all.
Insert Table 4 about here
In sum, it appears that the joy and anger breathing conditions successfully
induced the target feeling state. The fear and sadness breathing conditions induced a
mixed pattern of fear/anxiety and anger for the former and of positive state and sadness
for the latter. These blends in pattern could be explained in two different ways. One
possibility is that these breathing conditions indeed induced a blended emotional feeling
state. Another possibility is that some individuals responded to these manipulations
with a given feeling state, while others responded with another feeling state. For
instance, some participants may have felt joy while performing the sadness breathing
task, while other participants may have felt sad.
To decide between these alternatives, correlations were computed between
feeling scales. In the fear condition, anger and fear were positively correlated (r(27) =
.58, p < .002), indicating that the fear breathing pattern did indeed induce a blended
Respiratory Feedback 23
emotional feeling state. In contrast, in the sadness condition, positive state and sadness
were negatively correlated (r(27) = -0.32, p < .10). Thus, it seems that different
individuals reacted with different feeling states to the sadness breathing pattern. A
possibility is that, given the similarity in breathing instructions between the sadness and
joy breathing conditions, some participants performed a breathing pattern closer to the
joy breathing patterns, while other performed a “purer” sadness pattern. If this were
true, based on data presented in Table 3, "happy responders" in the sadness breathing
condition should evidence longer respiration time and amplitude, and smaller ratio of
inspiration/expiration time than "sad responders". Pauses parameters, however, should
not discriminate between these two groups. To test these hypotheses, respiratory
parameters were compared with t-tests between "sad and happy responders".
Participants who reported more happiness than sadness in the sadness breathing
condition were classified as "happy responders". If the opposite was true, they were
classified as "sad responders". Six participants who reported as much happiness as
sadness were eliminated (generally, these participants reported no sadness together with
no happiness at all). There were no differences between groups for the respiration time,
amplitude, or pauses parameters. However, as predicted, "sad responders" were
characterized by longer inspiration / expiration time ratio (M = .23, SD = .09) than
happy responders (M = .10, SD = .16), t(18) = 2.20, p < .03. Thus, it seems that one
objective respiratory parameter distinguishes between happy and sad responders. This
finding suggests that the quality of the feeling state observed results directly from the
breathing pattern performed rather than from any other factors. Future research might
attempt to better control the breathing patterns performed by using a bio-feedback
General discussion
Study 1 has indicated that people experience respiratory changes that are
subjectively differentiated across different types of emotions. Study 2 has documented
that differentiated emotional feeling states were induced by respiration manipulations
without participants’ awareness of the process. The intensity of the feeling states
Respiratory Feedback 24
induced in Study 2 was not trivial: Mean ratings of joy, anger and fear were of 54, 55,
and 47 on a scale in which 100 indicated the strongest intensity that participants could
imagine feeling. The amount of variance accounted for by this effect (40%) is larger
than the one accounted for by facial feedback (13%, in Matsumoto [1987]). To our
knowledge, this is the first demonstration that the alteration of respiration is sufficient
to induce emotion. It extends to visceral feedback the effects of body feedback on
emotional states established for facial expression (e.g. Matsumoto, 1987) and posture
(e.g. Stepper & Strack, 1993). These observations support the notion that body
feedback plays a role in the determination of the quality of emotional feeling state and
that this effect can occur without awareness of the process.
Taken together, the present results are totally congruent with the central network
perspective described in the introduction. They are not congruent with the
undifferentiated arousal model that postulates that emotion is characterized by a state of
undifferentiated arousal that uniquely influences the intensity of feelings, provided that
the individual is aware of the arousal and consciously attributes it to an emotional
cause. Indeed, in Study 2, although individuals were aware of their body changes, they
did not consciously relate them to an emotional state, as confirmed by a thorough
debriefing on the matter. Thus, the present results indicate that body changes might
influence feeling states independently of one’s awareness of the process. Still, it is
uncertain whether the awareness of the body state is necessary or not for peripheral
feedback to occur. A convincing demonstration against this specific question would be
to demonstrate the influence of respiratory changes on feeling states with participants
who were aware of neither the process nor the respiratory changes. However, for
practical reasons, such a demonstration, might be very difficult to realize.
Considering more specifically our results, it appears that joy, anger, and sadness-
-provided the execution of the proper breathing pattern--were successfully induced with
instructions derived from the observations of Study 1. However, mixed results were
observed for fear, which was not differentiated from anger. This observation raises the
question of whether breathing manipulation affects feeling state by activating discrete
Respiratory Feedback 25
emotions (e.g. Ekman, 1984; Levenson et al., 1990) or by moving it along dimensions
of pleasantness and activation (e.g. Feldman Barrett & Russell, 1998). Indeed, of the
four states induced, only fear and anger were in the same quadrant of unpleasant, high
arousal state. Anger was successfully induced, though accompanied by some fear,
whereas fear was not distinguished from anger.
There are at least four different interpretations that could account for this
observation. The first interpretation is that respiratory feedback is capable of inducing
discrete emotions and the fear breathing instructions derived from Study 1 were
incomplete or inadequate. This possibility can only be examined by further
psychophysiological research on respiration during emotion. A second interpretation is
that, while body feedback as a whole is capable of inducing discrete emotions, breathing
alone would not be sufficient to induce differentiated states of anger and fear because
additional feedback from other body functions is necessary. According to the SAME
model proposed by Cacioppo et al. (1992), one source of peripheral feedback might not
be enough to produce a discrete somatovisceral pattern that specifically refers to a
specific emotion. A third possibility is that peripheral feedback is not capable of such
fine distinctions, the latter requiring more cognitive appraisal processes. Finally, as
suggested above, it may be that feeling states are organized dimensionally (Feldman &
Russell, 1998) and that Study 2 results simply reflect this reality. Future research is
needed to decide among these possibilities. To test the second possibility, we are
presently planning studies in which facial, postural, and respiratory feedback will be
manipulated independently. By providing ambiguous and unambiguous somatovisceral
patterns (emotionally incongruent or congruent feedback from face, posture or
respiration), such manipulations allow for testing the Somatovisceral Afference Model
of Emotion (SAME) proposed by Cacioppo et al. (1992).
From a clinical perspective, our results suggest relations between anger-hostility
and fear-anxiety, as induced by rapid breathing. Indeed, it is remarkable that the fast
and deep breathing normally expected to induce more hyperventilation (Beck & Scott,
1988; Huey & West, 1983)--and consequently, more anxiety--than the fast and shallow
Respiratory Feedback 26
breathing actually induced more anger than anxiety. The fast and shallow breathing
induced as much anger as anxiety. These observations suggest that hyperventilation
might be as strongly related to anger and hostility as to fear and anxiety. This is
congruent with the observation that people who panic, for whom hyperventilation is
functional, score higher on hostility (Dadds, Gaffney, Kenardy, Oei, et al., 1993). The
anxiety produced by hyperventilation might thus originate in a hostile coping attitude in
challenging situations.
A final comment concerns the regulation of emotion. Previous research has
shown that attempts to regulate emotion by the suppression of its expression resulted in
an increase in physiological responding (Gross, 1998; Manstead, 1991). Other
researchers have observed juste the opposite (Kappas, McHugo, & Lanzetta, 1989).
Thus, attempts to regulate emotion in one physiological system (facial muscles) resulted
in increased manifestation in other body channels (visceral arousal) in some studies and
in decreased manifestation in other studies. Also relevant to this question, a wealth of
clinical evidence has shown that feelings of anxiety can be alleviated by specific
breathing exercises (Lum, 1981). It is therefore unclear whether the control of one body
channel necessarily results in increased manifestations in other channels. It might be
that the direction of the effect depends upon the body channel and the type of control
considered. The findings of the present studies encourage future research to examine
the regulatory effects of specific breathing instructions in people exposed to emotional
situations by independently manipulating breathing instructions and emotional
In sum, the present studies have shown an implicit influence of respiratory
feedback on the induction of emotional feeling state. They thus offer further support to
those theories of emotion stating that the quality of emotional feelings are, at least in
part, modulated by body feedback, without necessity of individual’s awareness of the
relationship between body changes and feeling state. It remains to be established
whether respiratory feedback induces discrete emotions or whether it moves the feeling
state along pleasantness and arousal dimensions. Finally, we propose that the
Respiratory Feedback 27
respiratory feedback effect constitutes a rich avenue for future research in emotion
Respiratory Feedback 28
Beck, J.G. & Scott, S.K. (1988). Physiological and symptom responses to
hyperventilation: A comparison of frequent and infrequent panickers. Journal of
Psychopathology and Behavioral Assessment, 10, 117-127.
Bloch (1994, July 16th). personal communication.
Bloch, S., Lemeignan, M., & Aguilera-T, N. (1991). Specific respiratory patterns
distinguish among human basic emotion. International Journal of Psychophysiology,11,
Bloch, S., & Santibanez-H, G. (1972). Simposio Latino-Americano de
psicobiologia del aprentizaje (pp. 170-184). In S. Bloch & R. Anerios (Eds.)
Psicobiologia del aprentizaje. Santiago, Chile: Universidad de Chile.
Boiten, F.A. (1994). Component analysis of task-related respiratory patterns.
International Journal of Psychophysiology, 15, 91-104.
Boiten, F.A. (1996). Autonomic response patterns during voluntary facial action.
Psychophysiology, 33, 123-131.
Boiten, F.A., Frijda, N.H., Wientjes, C.J.E. (1994). Emotions and respiratory
patterns: review and critical analysis. International Journal of Psychophysiology, 17,
Buck, R. (1985). Prime theory: An integrated view of motivation and emotion.
Psychological Review, 92, 389-413.
Cacioppo, J.T., Berntson, G.G. & Klein, D.J. (1992). What is an emotion? The
role of somatovisceral afference, with special emphasis on somatovisceral « illusions ».
In M.S. Clark (Ed.), Review of Personality and Social Psychology, 14, 68-98.
Cannon, W.B. (1927). The James-Lange theory of emotions: A critical
examination and an alternative theory. American Journal of Psychology, 39, 106-124.
Dadds, M.R., Rosenthal-Gaffney, L., Kenardy, J., Oei, T.P.S., et Evans, L.
(1992). An exploration of the relationship between expression of hostility and the
anxiety disorders. Journal of Psychiatric Research.
Respiratory Feedback 29
Damasio, A.R. (1994) Descartes' error: emotion, reason and the human brain.
New York: Grosset/Putnam Books.
Duclos, S.D., Laird, J.D., Schneider, E., Sexter, M., Stern, L. & VanLighten, O.
(1989). Emotion-specific effects of facial expressions and postures on emotional
experience. Journal of Personality and Social Psychology, 57, 100-108.
Ekman, P. (1984). Expression and the nature of emotion. In K. Scherer and P.
Ekman (Eds.), Approaches to emotion (pp. 319-343). Hillsdale, N.J.: Erlbaum.
Ekman, P. (1999). Basic emotions. In T. Dalgleish & M.J. Power, (Eds).
Handbook of cognition and emotion. (pp. 45-60). Chichester, England: John Wiley &
Feldman Barrett, L., Russell, J.A. (1998) Independance and Bipolarity in the
Structure of Affect. Journal of Personality and Social Psychology, 74, 967-984.
Frijda, N.H. (1986). The emotions. Cambridge, UK: Cambridge University
Gross J.J. (1998) Antecedent- and response-focused emotion regulation:
divergent consequences for experience, expression, and physiology. Journal of
Personality and Social Psychology, 74, 224-237.
Hess, U., Kappas, A., McHugo, G.J., Lanzetta, J.T. & Kleck, R.E. (1992). The
facilitative effect of facial expression on the self-generation of emotion. International
Journal of Psychophysiology, 12, 251-265.
Huey, S.R., & West, S.G. (1983). Hyperventilation: its relation to symptom
experience and to anxiety. Journal of Abnormal Psychology, 92, 422-432.
Izard, C.E. (1979). Emotion in personality and psychopathology. New York:
Plenum Press.
Izard, C.E. (1990). Facial expressions and the regulation of emotions. Journal of
Personality and Social Psychology, 58, 487-498.
James, W. (1884). What is an emotion? Mind, 9, 188-205.
James, W. (1890). Principles of psychology (2 Vols.). New York: Holt.
Respiratory Feedback 30
Kappas, A. (1989). Control of emotion. Unpublished Ph.D. dissertation,
Darmouth College, USA.
Kappas, A., McHugo, & G.J., Lanzetta, J.T. (1989, October). Componential
control and the modulation of emotional experience. Paper presented at the Twenty-
Ninth Annual Meeting of the Society for Psychophysiological Research, New Orleans.
Kirouac, G. (1995). Les émotions. Québec: Presses de l'Université du Québec.
Laird, J. (1984). The real role of facial response in the experience of emotion: a
reply to Tourangeau and Ellsworth, and others. Journal of Personality and Social
Psychology, 47, 909-917.
Lang, P.J. (1979). A bio-informational theory of emotional imagery.
Psychophysiology, 16, 495-512.
Lang, P.J. (1984). Cognition in emotion: Concept and action. In C. Izard, J.
Kagan, & R. Zajonc (Eds.), Emotion, Cognition, and Behavior. New York: Cambridge
University Press.
Lazarus, R.S. (1991). Progress on a cognitive-motivational-relational theory of
emotion. American Psychologist, 46, 819-834.
Levenson, R.W. (1992). Autonomic nervous system differences among
emotions. Psychological Science, 3, 23-27.
Levenson, R.W., Carstensen, L.L., Friesen, W.V., & Ekman, P. (1991).
Emotion, physiology, and expression in old age. Psychology and Aging, 6, 28-35.
Levenson, R.W., Ekman, P. & Friesen, W.V. (1990). Voluntary facial action
generates emotion-specific autonomic nervous system activity. Psychophysiology, 27,
Lum, L.C. (1981). Hyperventilation and anxiety state. Journal of the Royal
Society of Medicine, 74, 1-4.
Lyman, B. & Waters, J.C.E. (1986). The experiential loci and sensory qualities
of various emotions. Motivation and Emotion, 10, 25-37.
Respiratory Feedback 31
Manstead, A.S.R. (1988). The role of facial movement in emotion. In H.L.
Wagner (Ed.), Social psychophysiology and emotion: Theory and clinical application.
Manstead, A.S.R. (1991). Expressiveness as an individual difference. In R.S.
Feldman and B. Rimé (Eds.), Fundamentals of Nonverbal Behavior (pp. 285-328).
Cambridge: Cambridge University Press.
Matsumoto, D. (1987). The role of facial responses in the experience of
emotion: More methodological problems and a meta-analysis. Journal of Personality
and Social Psychology, 52, 769-774.
McIntosh, D.N. (1996). Facial feedback hypotheses: evidence, implications, and
directions. Motivation and Emotion, 20, 121-147.
Pecchinenda, A. & Smith, C.A. (1996). The affective significance of skin
conductance activity during a difficult problem-solving Cognition-and-Emotion, 10,
Personality and Social Psychology Bulletin (1990). Special issue: Centennial
celebration of the Principles of Psychology of William James. Personality and Social
Psychology Bulletin, 16, 597-783.
Philippot, P. (1993). Inducing and assessing differentiated emotion-feeling states
in the laboratory. Cognition & Emotion, 7, 171-193.
Philippot, M. & Philippot, P. (1991). Thorax: A program of respiratory signal
analysis [computer program]. Louvain-la-Neuve, Belgium: University of Louvain,
Experimental & Social Psychology Unit.
Philippot, P. & Rimé, B. (1997). The perception of bodily sensation during
emotion: A cross-cultural perspective. Polish Bulletin of Psychology, 28, 175-188.
Philippot, P. & Schaefer, A. (2001). Emotion and memory. In T.J. Mayne and
G.A. Bonano (Eds.), Emotion: current issues and future directions. New York: Guilford
Reisenzein, R. (1983). The Schachter theory of emotion: Two decades later.
Psychological Bulletin, 94, 239-264.
Respiratory Feedback 32
Rimé, B., Philippot, P., & Cisamolo, D. (1990). Social schemata of peripheral
changes in emotion. Journal of Personality and Social Psychology, 59, 38-49.
Russell, J. (1991). In defence of a prototype approach to emotion concepts.
Journal of Personality and Social Psychology, 60, 37-47.
Schachter, S. (1964). The interaction of cognitive and physiological
determinants of emotional state. In L. Berkowitz (Ed.), Advances in Experimental
Social Psychology: Vol. 1. (pp. 49-79). New York: Academic Press.
Scherer, K.R. (1984). On the nature and function of emotion: A component
process approach. In K. Scherer and P. Ekman (Eds.), Approaches to emotion (pp. 293-
317). Hillsdale, N.J.: Erlbaum.
Smith, C.A. & Kirby, L.D. (2000). Consequences require antecedents: Toward a
process model of emotion elicitation. In J.P. Forgas (Ed). Feeling and thinking: The role
of affect in social cognition. New York: Cambridge University Press.
Stemmler, G. (1989). Autonomic differentiation of emotion revisited:
Convergent and discriminant validation. Psychophysiology, 26, 617-632.
Stemmler, G. (1992). The vagueness of specificity: Models of peripheral
physiological emotion specificity in emotion theories and their experimental
discriminability. Journal of Psychophysiology, 6, 17-28.
Stepper, S. & Strack, F. (1993). Proprioceptive determinants of affective and
nonaffective feelings. Journal of Personality and Social Psychology, 64, 211-220.
Teasdale, J.D. (1996). Clinical relevant theory: Integrating clinical insight with
cognitive science. In P. Salkovskis (Ed.), Frontiers of cognitive therapy. New York:
Tomkins, S.S. (1980). Affect as amplification: some modification in theory. In
R. Plutchik & H. Kellerman (Eds.), Theories of emotion (pp. 141-164). New york:
Academic Press.
van Rossum, E. (1988). Ademhalingsmeetsystee (ReportNo. 5). Amsterdam,
Holland: Vrij Universiteit, Subfakulteit Psychologie, Electronische Werkplaats.
Respiratory Feedback 33
Zajonc, R.B. & McIntosh, D.N. (1992). Emotions research - Some promising
questions and some questionable promises. Psychological Science, 3, 70-74.
Zillmann, D. (1979). Excitation transfer in hostility and aggression. In D.
Zillman (Ed.), Hostility and aggression. Hillsdale, N.J.: Erlbaum.
Zillmann, D. (1983). Transfer of excitation in emotion behavior. In J.T.
Cacioppo & R.E. Petty (Eds.), Social psychophysiology (pp. 215-240). New York:
Respiratory Feedback 34
1It should be noted that only 20 participants took part in this study and that, although it
used a within-subject design its statistical power is consequently weak.
2This apparatus only served to convince participants of the cover story. Cardiovascular
data were thus not recorded. Given the strong impact of breathing on cardiovascular
parameters, such data would have been useless in the present context.
Respiratory Feedback 35
Table 1.
Emotional Feeling State as a Function of Emotion Condition.
Feeling Emotion Condition
State Joy Anger Fear Sadness
Joyful Mean 3.09b 1.48c 1.34d 1.17c
SD 1.62 0.89 0.71 0.49
Sad Mean 1.09c 1.61c 1.69cd 4.13a
SD 0.28 1.15 1.18 2.09
Angry Mean 1.04c 4.39a 1.91cd 2.13b
SD 0.20 1.97 1.44 1.68
Afraid Mean 1.00c 1.52c 4.08b 1.43c
SD 0.00 0.89 2.25 0.78
Anxious Mean 1.13c 3.04b 3.91b 2.43b
SD 0.34 1.96 1.95 1.99
Happy Mean 4.30a 1.17c 1.34d 1.13c
SD 1.79 0.49 0.77 0.45
Note. Means with different subscripts differ at least at the 0.05 level of significance
according to t-test using Bonferroni’s correction.
Only emotional feeling states items relevant to the emotion conditions are presented in
this Table.
Respiratory Feedback 36
Table 2.
Respiration Parameters Means Values and Standard Deviation as a Function of
Respiration Emotion
Parameter Joy Anger Fear Sadness
Frequency Mean -1.85 c 1.04 a 1.45 a -0.35 b
SD 1.25 1.22 1.36 1.61
Amplitude Mean 2.07 a 0.91 b -0.22 c 0.22 bc
SD 0.82 1.27 1.91 1.95
Regularity Mean 1.69 a -0.83 c -0.91 c -0.61 b
SD 1.26 1.59 1.93 1.56
Sighs Mean 1.26 b 0.96 b 0.87 b 2.39 a
SD 1.35 1.43 1.46 1.53
Tremors Mean 0.04 c 1.39 b 1.48 b 2.65 a
SD 0.21 1.23 1.40 1.59
Thoraxic Tension Mean 0.13 c 2.43 a 2.65 a 1.52 b
SD 0.34 1.44 1.03 1.53
Note. Means with different subscripts differ at least at the 0.01 level of significance
according to t-test using Bonferroni’s correction.
Respiratory Feedback 37
Table 3
Respiratory Parameters as a Function of Breathing Condition.
Respiratory Breathing Condition
Parameter Joy Anger Fear Sadness F(3,23) p
Time Mean4.13a -6.41c -7.26d -1.26b 84.14 .0001
SD 4.85 2.48 2.50 2.70
Amplitude Mean0.50a 0.04b -0.20c 0.15b 20.98 .0001
SD 0.45 0.34 0.26 0.29
Ti/Te Mean0.05b 0.14a 0.32a 0.17a 6.90 .002
SD 0.17 0.24 0.38 0.14
Pause Length Mean-0.44a -1.20b -1.39b 0.25a 6.22 .002
SD 1.55 1.75 1.69 2.35
Pause Number Mean-0.15 0.52 0.22 0.67 1.12 .35
SD 0.99 2.15 1.67 1.84
Note. Means with different subscripts differ at least at the 0.01 level of significance
according to t-test using Bonferroni’s correction.
Respiratory Feedback 38
Table 4.
Emotion Feeling States as a Function of Breathing Condition.
Emotion Breathing Condition
Feeling State Joy Anger Fear Sadness
Positive state Mean 54a I 5c III 6c II 23b I
SD 33 8 12 25
Anger Mean 1c II 55a I 47b I 7c II
SD 2 37 32 12
Anxiety, Fear Mean 1b II 40a II 39a I 8b II
SD 2 35 34 13
Sadness Means 5b II 12b III 13b II 21a I
SD 13 16 26 27
Note. Means with different subscripts differ at least at the 0.01 lsmithevel of
significance according to t-test using Bonferroni’s correction. Alphabetic subscripts
indicate a comparison between breathing conditions for a given feeling state; Roman
figures subscripts indicate a comparison between feeling state for a given breathing
... Bien que cette théorie soit en totale rupture avec les conceptions classiques de l'époque, elle va trouver un écho dans le siècle qui suit. Elle sera notamment partiellement validée par les travaux de James Douglas Laird (1974) [Lai74] qui traite de l'impact des expressions faciales simulées dans le ressenti des émotions, ou encore par les travaux de Sabine Stepper et Fritz Strack (1993) sur l'impact de la posture [SS93] ou encore par les travaux de Pierre Philippot (2002) [PCB02] sur l'impact de la respiration. ...
... Bien que cette théorie soit en totale rupture avec les conceptions classiques de l'époque, elle va trouver un écho dans le siècle qui suit. Elle sera notamment partiellement validée par les travaux de James Douglas Laird (1974) [Lai74] qui traite de l'impact des expressions faciales simulées dans le ressenti des émotions, ou encore par les travaux de Sabine Stepper et Fritz Strack (1993) sur l'impact de la posture [SS93] ou encore par les travaux de Pierre Philippot (2002) [PCB02] sur l'impact de la respiration. ...
... On voit sur ce tableau la description de la joie par exemple. Pour Philippot [PCB02], elle se caractérise par une situation peu soudaine et assez prévisible, qui suscite un fort plaisir intrinsèque, qui est une situation plutôt aidante (donc elle a une forte relation avec les attentes du sujet), sur laquelle on a plus ou moins de contrôle et qui est plus ou moins compatible avec les normes. ...
Les centres d'appels reçoivent tous les jours des milliers de coups de téléphone permettant de faire le lien entre des clients et des conseillers. Ainsi, de nombreuses informations peuvent être extraites de ces conversations, dont l'aspect émotionnel.Cette thèse CIFRE a été réalisée en collaboration avec l’entreprise Allo-Media qui est spécialisée dans l'analyse automatique de conversations téléphoniques de centre d'appels. Concrètement, elle met en place des relevés d'information sur différents aspects de la conversation en indexant ces informations pour permettre un traitement automatique des données. L’entreprise cherche à enrichir ses annotations avec une solution innovante permettant de rajouter un aspect émotionnel en adéquation avec le contexte de la relation clientèle afin d'alerter sur les points saillants de la conversation.Cette thèse tente donc de répondre à plusieurs problématiques : (i) tout d'abord la définition de l'émotion de satisfaction et de frustration dans la parole, (ii) la mise en place d'une reconnaissance automatique de ces émotions de façon continue tout au long de la conversation et (iii) des méthodes d'évaluation de ces systèmes automatiques.Les contributions de cette thèse sont : (i) la construction d’un corpus à partir de données réelles, annoté de façon continue en satisfaction et frustration, (ii) la mise en place de différentes stratégies pour construire un système de reconnaissance automatique utilisant des réseaux de neurones profonds en nous comparant à l'état de l'art, (iii) l’exploration de la dissociation des aspects acoustique et linguistique des conversations afin d'améliorer nos systèmes de reconnaissance et enfin (iv) la mise en place d’une évaluation nuancée de ces systèmes.
... Therefore, research and development on vital sign monitoring systems that include respiratory rate measurement can be usefully utilized to prevent physical health risks by detecting deterioration in patients' heart conditions in advance [12,13]. Respiration is influenced not only by physical health, but also by psychological conditions, including anxiety, depression, anger, and stress, and is closely related to changes in levels of consciousness caused by drugs and sleep [14][15][16]. Therefore, respiration is a very important indicator that can be applied to observe changes in human physical and psychological health and other conditions, and can be widely used in medicine, as well as various other fields. ...
... The guidelines were used to induce three types of respiratory patterns to include various respiratory patterns in the data, and images and respiratory signals were photographed for 260 s per the guidelines. Guideline #1 induced breathing at 40,35,30,25,20,15,10,15,20,25,30,35, and 40 bpm for 20 s, and Guideline #2 induced breathing in the form of rapid change in bpm speed to reflect rapid changes in the data. Guideline #3 induced breathing that included pauses during breathing to reflect a stationary state in the learning data. ...
... The guidelines were used to induce three types of respiratory patterns to include various respiratory patterns in the data, and images and respiratory signals were photographed for 260 s per the guidelines. Guideline #1 induced breathing at 40,35,30,25,20,15,10,15,20,25,30,35, and 40 bpm for 20 s, and Guideline #2 induced breathing in the form of rapid change in bpm speed to reflect rapid changes in the data. Guideline #3 induced breathing that included pauses during breathing to reflect a stationary state in the learning data. ...
Full-text available
Human respiration reflects meaningful information, such as one’s health and psychological state. Rates of respiration are an important indicator in medicine because they are directly related to life, death, and the onset of a serious disease. In this study, we propose a noncontact method to measure respiration. Our proposed approach uses a standard RGB camera and does not require any special equipment. Measurement is performed automatically by detecting body landmarks to identify regions of interest (RoIs). We adopt a learning model trained to measure motion and respiration by analyzing movement from RoI images for high robustness to background noise. We collected a remote respiration measurement dataset to train the proposed method and compared its measurement performance with that of representative existing methods. Experimentally, the proposed method showed a performance similar to that of existing methods in a stable environment with restricted motion. However, its performance was significantly improved compared to existing methods owing to its robustness to motion noise. In an environment with partial occlusion and small body movement, the error of the existing methods was 4–8 bpm, whereas the error of our proposed method was around 0.1 bpm. In addition, by measuring the time required to perform each step of the respiration measurement process, we confirmed that the proposed method can be implemented in real time at over 30 FPS using only a standard CPU. Since the proposed approach shows state-of-the-art accuracy with the error of 0.1 bpm in the wild, it can be expanded to various applications, such as medicine, home healthcare, emotional marketing, forensic investigation, and fitness in future research.
... Additionally, accurate emotional measurement is a key and difficult point of driving emotion research. Measurement based on electroencephalogram (EEG) [24], electrocardiogram (ECG) [25], galvanic skin response (GSR) [26], respiration (RSP) [27], facial expression [28], or voice signal [29] are popular emotion identification methods. Among these methods, the ones based on the physiological indexes (EEG, ECG, GSR, and RSP) are acknowledged as the more effective [24][25][26][27]30]. ...
... Measurement based on electroencephalogram (EEG) [24], electrocardiogram (ECG) [25], galvanic skin response (GSR) [26], respiration (RSP) [27], facial expression [28], or voice signal [29] are popular emotion identification methods. Among these methods, the ones based on the physiological indexes (EEG, ECG, GSR, and RSP) are acknowledged as the more effective [24][25][26][27]30]. While eight kinds of driving emotions were investigated in this paper, it seems impossible to find a measurement that effectively works for all of them with the mentioned methods. ...
Full-text available
Emotion is an implicit psychological characteristic that changes over time. When it accumulates to a certain extent, it will be accompanied by certain external manifestations. Drivers with different traits have different emotional performance, which leads to different effects from different driver traits on the driver’s emotional activation efficacy. In this study, we thoroughly explore the effects of different genders, age, driving competence, driving anger tendency, driving safety attitude and stress state on driver’s emotional activation efficacy. This paper selects 74 young and middle-aged drivers with an age distribution between 20 and 41 years old. The eight most typical driving emotions (anger, surprise, fear, anxiety, helplessness, contempt, ease and pleasure) were screened through questionnaires. An experimental framework for the emotional stimulation and measurement of eight driving emotions was designed based on multiple emotional stimulation methods and PAD emotional model. The effect of emotional activation on drivers of different genders, age, driving competence, driving anger tendency, driving safety attitude and stress state was explored in depth. The results show that gender, age, driving safety attitude, driving anger tendency, stress state, etc., all have different degrees of influence upon the activation efficacy of emotion. The research results reveal the rules for the generation of different driving emotions to a certain extent and provide a theoretical basis for further exploring the cognitive and behavioral characteristics of drivers with different emotions.
... At the cognitive level, the hypocapnia induced by involuntary hyperventilation was found to be related to cognitive deficits (Ley, 1999) and decreased cerebral blood flow velocity (Debreczeni et al., 2009). At the emotional level, hyperventilation is related to anger, anxiety, and fear, and these emotional responses may trigger hostile coping responses in hostile situations, and should therefore better be prevented (Philippot et al., 2002). Finally, given that voluntary hyperventilation may interfere with the neuronal activity-driven regulation of cerebral circulation (Debreczeni et al., 2009;Ley, 1999), future research must better understand its impact on the cognitive factors influencing the different kinds of physical sport performance. ...
Breathing techniques are predicted to affect specific physical and psychological states, such as relaxation or activation, that might benefit physical sport performance (PSP). Techniques include slow-paced breathing (SPB), fast-paced breathing (FBP), voluntary hyperventilation (VH), breath-holding (BH), and alternate- and uni-nostril breathing. A systematic literature search of six electronic databases was conducted in April 2022. Participants included were athletes and exercisers. In total, 37 studies were eligible for inclusion in the systematic review, and 36 were included in the five meta-analyses. Random effects meta-analyses for each breathing technique were computed separately for short-term and longer-term interventions. Results showed that SPB and BH were related to improved PSP, with large and small effect sizes for longer-term interventions, respectively. In short-term interventions, SPB, BH, and VH were unrelated to PSP. There was some evidence of publication bias for SPB and BH longer-term interventions, and 41% of the studies were coded as having a high risk of bias. Due to an insufficient number of studies, meta-analyses were not computed for other breathing techniques. Based on the heterogeneity observed in the findings, further research is required to investigate potential moderators and develop standardised breathing technique protocols that might help optimise PSP outcomes.
... This covert sensing will reduce the distraction and nervousness of the participant and, thus, decrease the systematic bias. Furthermore, our sensor can offer critical information for studying RSA, voluntary respiration manipulation, and their effect on cardiovascular [55] and skin conductance changes [56]. This area of research has also been associated with beneficial effects on mental and physical health [57]. ...
Full-text available
This work presents a study on users’ attention detection with reference to a relaxed inattentive state using an over-the-clothes radio-frequency (RF) sensor. This sensor couples strongly to the internal heart, lung, and diaphragm motion based on the RF near-field coherent sensing principle, without requiring a tension chest belt or skin-contact electrocardiogram. We use cardiac and respiratory features to distinguish attention-engaging vigilance tasks from a relaxed, inattentive baseline state. We demonstrate high-quality vitals from the RF sensor compared to the reference electrocardiogram and respiratory tension belts, as well as similar performance for attention detection, while improving user comfort. Furthermore, we observed a higher vigilance-attention detection accuracy using respiratory features rather than heartbeat features. A high influence of the user’s baseline emotional and arousal levels on the learning model was noted; thus, individual models with personalized prediction were designed for the 20 participants, leading to an average accuracy of 83.2% over unseen test data with a high sensitivity and specificity of 85.0% and 79.8%, respectively
... Breathing interventions, particularly those in which breathing frequency is manipulated, offer promising, low cost, unobtrusive approaches to beneficially impact the autonomic nervous system as well as perceived emotional states (Schuman and Killian, 2019). Slow breathing frequency has been associated with reductions in perceived negative affect (Ma et al., 2017), while fast breathing is associated with increases in perceived negative affect, including discrete emotions such as anger and fear (Philippot et al., 2002;Ali et al., 2018). However, even "negative" emotional experiences can be of benefit in performance contexts, eliciting increases in speed and power that may be ideal for performance of many sport tasks. ...
Full-text available
Breathing interventions have been shown to improve sport performance. Although evidence exists to support the role of perceived arousal as a critical underlying mechanism of breathing interventions, methodological differences in the literature preclude clear understanding of potential contributing factors to the effectiveness of such interventions. Under neutral contexts, we have demonstrated attention, dyspnea, and hindrance may need to be considered as mediators of how breathing frequency affects motor performance. We sought to extend our previous findings to determine how breathing frequency affects motor performance under varying emotional conditions. Participants ( N = 35, Mage = 21.68, SD = 2.96; 20 females) performed slow, normal, and fast metronome-paced breathing while viewing pleasant and unpleasant stimuli prior to executing a pinch grip task. Performance was assessed via reaction time (RT), variability (V) and error (AE). Assessment of indices of perceived arousal included measuring heart rate variability (HRV) and visual analog scale responses. Visual analog scales were also used to assess attention, dyspnea, and hindrance. Repeated measures ANOVAs showed slow breathing increased RT and HRV compared to normal and fast breathing under emotional conditions (all p ’s < 0.05). Hierarchical multiple regression models revealed that decreased breathing frequency predicted increases in RT ( β = −0.25, p < 0.05) under pleasant conditions, while predicting increases in HRV for unpleasant conditions ( β = −0.45, p < 0.001). Increases in dyspnea ( β = 0.29, p < 0.05) and hindrance ( β = 0.35, p < 0.01) predicted increases in RT under pleasant conditions, while only increases in hindrance predicted increases in RT under unpleasant conditions ( β = 0.41, p < 0.01). Decreases in breathing frequency predicted increases in HRV under unpleasant conditions ( β = −0.45, p < 0.001). Overall, our findings suggest under varying emotional contexts breathing frequency differentially affects movement, potentially mediated by factors other than perceived arousal. In addition, these results inform the use of breath regulation as an antecedent emotion regulation strategy.
... Nevertheless, the non-physiological signal response can be concealed and may not directly reveal the inherent state of the human emotions [48,64,66]. Inner state of the human emotions is revealed by physiological signals, such as electroencephalogram (EEG), electrocardiogram (ECG), electrooculogram (EoG), electromyogram (EMG), galvanic skin response (GSR), respiration rate (RSP), and eye-gaze tracker [8,39,61]. ...
Full-text available
Electroencephalogram (EEG)-based emotion recognition models are gaining interest as they show the intrinsic state of human. A wide range of features are extracted from the scalp EEG recorded using a different set of electrodes across the brain regions. However, there are no standard set of features accepted amongst researchers for emotion recognition. As a result, new researchers in the field use all features reported in the literature which leads to the curse of dimensionality problem and performance degradation due to high correlation within the feature set. Thus, the primary objective of this work is to improve the performance of the emotion recognition model by using an optimal feature set. This research article proposes differential-evolution-based feature selection (DEFS) algorithm to obtain an optimal feature set for effective subject-independent emotion recognition. The optimal feature set obtained from the DEFS algorithm is used to train the SVM classifier. A wide range of experiments are conducted to analyze the performance of our proposed model using a publicly available EEG-based emotion recognition dataset. The proposed model has been compared with several state-of-the-art feature selection and optimization algorithms. The results are analyzed in the aspects of classification performance, fitness value optimization and computational time. In addition, to assure the subject-independent behavior of the proposed model, subject-wise performance has been analyzed. The proposed DEFS-SVM emotion recognition model has got the classification accuracies of 73.60, 74.23, 71.88 and 71.80% to detect valence arousal, valence, dominance, and liking emotional states, respectively. The experimental results assured that our proposed model outperforms all other algorithms in all aspects. Also, the proposed feature selection algorithm is suitable for any EEG-based emotion recognition model to optimize the feature set.
Demystifying Emotions provides a comprehensive typology of emotion theories in psychology (evolutionary, network, appraisal, goal-directed, psychological constructionist, and social) and philosophy (feeling, judgmental, quasi-judgmental, perceptual, embodied, and motivational) in a systematic manner with the help of tools from philosophy of science, allowing scholars in both fields to understand the commonalities and differences between these theories. Agnes Moors also proposes her own novel, skeptical theory of emotions, called the goal-directed theory, based on the central idea that all kinds of behaviors and feelings are grounded in goal-striving. Whereas most scholars of emotion do not call the notion of emotion itself into question, this review engages in a critical examination of its scientific legitimacy. This book will appeal to readers in psychology, philosophy, and related disciplines who want to gain a deeper understanding of the controversies at play in the emotion domain.
Significant developments within the past few years have made possible the publication of this rather large volume focusing on specific emotions of human experience, such as interest, joy, anger, distress, fear, shame, shyness, and guilt. The relevant events include new evidence on the relationship of emotions to cognitive processes and to personality traits and defense mechanisms. They also include discoveries relating to the biological foundations of emotions and theory regarding their significance in human evolution. Finally, there have been important findings on the role of emotions and emotion expressions in social relations, pain, grief, and psychopathology. These developments are elaborated in the pages of this volume. The contributors represent the disciplines of clinical, social, and experi­ mental psychology, psychiatry, and psychoanalysis. The contributions show important common themes that cut across disciplines, but they also reflect some differences that invite further thought and research. Above all, they add to our knowledge of human emotions and to our ability to understand and resolve human problems. The Department of Psychology of the University of Delaware has pro­ vided an excellent intellectual climate for work on a volume that ranges across several specialities and disciplines. Conversations with colleagues in the offices and hallways of Wolf Hall have provided answers to many questions. They also yielded some questions that compelled me to seek greater clarification of an issue.
Four essential modifications in my affect theory are presented. First, an ambiguity in the concept of affect as amplification has been revised so that the affect is now considered to be an analogic amplifier in much the same manner as pain is an analogic amplifier of the injury it amplifies. Second, I now view the skin of the face as more essential than its musculature in providing the feedback which we experience as motivating. It is shown also that the skin in general is a powerful motivational organ in sex, pain, and sleep. Third, I now view innate affect as essentially suppressed and backed up in the adult, exacting a price whose cost is yet to be precisely determined. The mechanism by which this is achieved is through suppression of vocalization of affect. Fourth, I now view the affect as amplifying not only its activation but also the response to the affect, because it coexists with and thereby imprints its form on whatever response follows it.
18 likely hyperventilators and 16 unlikely hyperventilators, balanced with regard to sex, were selected from an undergraduate population (N = 385) using a screening questionnaire of hyperventilation-related symptoms experienced in life. Likely hyperventilators reported a higher mean intensity of hyperventilation-related symptoms than unlikely hyperventilators after overbreathing room air but not after overbreathing carbon dioxide enriched air. Mixed results were obtained in comparing the 2 groups in their reports of dummy symptoms. The likely hyperventilators' normal breathing also showed higher volume per minute and rate than that of the unlikely hyperventilators. Results support a relationship between chronic hyperventilation and symptom experience in a nonclinical population. (41 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved).
A theory of emotional imagery is described which conceives the image in the brain to be a conceptual network, controlling specific somatovisceral patterns, and constituting a prototype for overt behavioral expression. Evidence for the hypothesis that differentiated efferent activity is associated with type and content of imaginal activity is considered. Recent work in cognitive psychology is described, which treats both the generation of sensory imagery and text comprehension and storage as examples of the processing of propositional information. A similar propositional analysis is applied to emotional imagery as it is employed in the therapeutic context. Experiments prompted by this view show that the conceptual structure of the image and its associated efferent outflow can be modified directly through instructions and through shaping of reports of image experience. The implications of the theory for psychopathology are considered, as well as its relevance to therapeutic behavior change.