On the Genesis of Emotional Experience

1
On the Genesis of Emotional Experience

• What kind of an emotion of fear would be left, if
  the feelings neither of quickened heart-beats nor
  of shallow breathing, neither of trembling lips
  nor of weakened limbs, neither of goose-flesh nor
  of visceral stirrings, were present, it is quite
  impossible to think.
•  
• -- William James (1884)

Alex Genov, Ph.D.
.
2
Overview of Research Program

• Studying the genesis of feelings and their role
  in behavior.
• Different levels of analysis
• physiological responses
• expressive behavior
• instrumental action
• subjective experience

3
The Logic Behind the Research

• Gist of Self-perception theory
• We identify our own mental states
  (e.g.,attitudes, feelings) in the same way
  outside observers would identify our mental
  states, namely by inferring them from perceptions
  of our behavior and/or the situation.

4
The Logic Behind the Research

• Basic experimental strategy
• In a well-disguised manner, the researcher
  manipulates a behavior presumed to express a
  feeling and then asks participants to report on
  their subjective experience.
• Random assignment to experimental and control
  groups, control for experimenter demand, etc.
  facilitates claims for causation.

5
The Logic Behind the Research

• Individual differences in the use of these two
  kinds of information
• Previous research by Laird and others has
  repeatedly demonstrated the following when asked
  to rate their emotional feelings after an
  experimental manipulation of their behavior, some
  people are influenced by cues from their behavior
  (personal cue responders) while others are
  influenced by cues from the situation
  (situational cue responders).

6
The Disguised Facial Expression Manipulation
(DFEM)

• Manipulation introduced as studying the effect
  of facial muscle activity on memory for abstract
  pictures.
• Participants asked to contract various facial
  muscles thus adopting a smile or a frown
  expression.
• At the same time they look at an abstract picture
  with either a happy or an angry title.
• The title always contradicts the valence of the
  expression.
• Participants report on their subjective
  experience using a visual analogue scale.
• Measuring feelings is introduced as controlling
  for the extraneous effects of random variations
  in mood.

7
(No Transcript)
8
Emotion Rating Scale (version 1) Please describe
how you are feeling now by making an X on the
part of the line that best describes how
strongly you feel each of these emotions Dont
feel at all Feel very
strongly Example SURPRISED
I-----------X----------------------------------I
Dont feel at all Feel very
strongly ANGRY I-------------------------
----------------------I SURPRISED
I-----------------------------------------------I
SAD I--------------------
---------------------------I SCARED
I-----------------------------------------------I
DISGUSTED I----------------------------
-------------------I HAPPY
I-----------------------------------------------I
ANXIOUS I-------------------------------
----------------I INTERESTED
I-----------------------------------------------I
9
DFEM

• Procedure Sequence
• smile 1 look at a picture called Rip-off for
  20 seconds
• fill out emotion rating scale
• frown 1 look at a picture called Spring for
  20 sec.
• fill out emotion rating scale
• smile 2 look at a picture called Betrayal
  for 20 sec.
• fill out emotion rating scale
• frown 2 look at a picture called Dancing for
  20 sec.
• fill out emotion rating scale

10
DFEM

• Response Definition
• (happy for smile 1) (happy for frown 1)
  (angry for frown 1) (angry for smile 1)
  score 1
• (happy for smile 2) (happy for frown 2)
  (angry for frown 2) (angry for smile 2)
  score 2
• If (scores 1 AND 2) gt 0 gt cue response
  personal
• If (scores 1 OR 2) lt or 0 gt cue response
  situational
• Positive scores on both pairs of trials indicate
  that participants report feelings consistent with
  their facial expressions across the two pairs of
  trials. Such participants are assigned to the
  personal cue group. All others were assigned
  to situational cue group.

11
DFEM

• These individual differences have been
  established most frequently in research on the
  effects of facial expressions and postures.
• In addition, subjects whose feelings are more
  affected by their facial expressions of emotion
• are more responsive to the manipulation of their
  posture
• are more empathetically correct in identifying
  the emotions of others when they imitate their
  expressions
• change their attitudes more in the induced
  compliance procedure
• conform less
• are more sensitive to pain and tolerate less of
  it in the cold pressor test, and so on.

12
Autonomic Activity in Emotions

• The Problem
• Most authorities agree that the autonomic
  nervous system plays some role in determining the
  subjective experience of emotion. They disagree,
  however, on the exact nature of this role.

13
Introduction

• View 1 Actual sympathetic activation
• is a necessary component of emotional experience
  and behavior, along with cognitions
• James -- a purely disembodied human emotion is
  a nonentity
• Schachter Singer the two-factor theory
• Zillmann excitation transfer

14
Introduction

• View 2 Actual sympathetic activation
• is not a necessary component of emotional
  experience and behavior, only cognitions matter
• Valins cognitive representation of SA and the
  false feedback paradigm

15
Introduction

• Both views supported by extensive empirical data
  gt A contradiction!

16
Introduction

• Individual differences to the rescue
• Schacters epinephrine injection manipulation
  measures responsiveness to personal cues, while
  the stooge manipulation measures responsiveness
  to situational cues.
• Valins false feedback manipulation can be seen
  as measuring responsiveness to situational cues.
• Therefore, Valins findings can be interpreted as
  supporting the cognitive component of Schachters
  theory, instead of contradicting the theory.
• Valins did not test adequately for the role of
  the physiological component of the two-factor
  theory, i. e. actual SA.
• Since, in general, groups of subjects have been
  compared, the effects of actual and believed SA
  may have occurred only in part of the group.
• The present study tested this possibility in a
  factorial experiment.

17
Hypotheses

• We expected
• In a single experiment to find evidence for the
  role of
• Actual SA and
• Cognitions about SA in emotional experience
• Demonstrate individual differences in the role of
  actual and believed SA
• Different effects for personal and situational
  cue responders
• An increase in actual SA (indexed by HR and SCL)
  will lead to an increase in self-reported
  feelings of fear for personal cue responders but
  not for situational cue responders.
• The belief of increased SA would lead to an
  increase in self-reported feelings of fear for
  both situational and personal cue responders.

18
Basic Elements of the Procedure and Cover Story

• Baseline physiological readings.
• The Disguised Facial Expression Manipulation.
• Four trials representing all combinations of high
  and low actual and believed SA. In each trial
  participants watched a film clip and described
  the co-occurring feelings.
• Measuring feelings of fear in response to four
  film clips.

• Baseline physiological readings.
• Measuring the effects of facial muscle activity
  on memory for abstract pictures.
• Measuring the effect of physical activity and
  film clips on various physiological indexes.
• Controlling for the extraneous effects of random
  changes in mood.

19
Baseline Physiological Readings

• One minute HR and SCL
• HR measured using a photoplethysmograph.
• SCL measured using constant voltage passed
  between two reusable silver-silver chloride
  sensors.
• Autonomic activity measured using J J I-330
  modules and Unicomp software from the American
  Biotech Co. A 486 IBM- compatible computer was
  programmed to process physiological data by
  averaging them into 10-second intervals.

20
Disguised Facial Expression Manipulation

• A smile trial

Rip-off
21
Belief of SA Manipulation metronome clicks

• During the initial instructions, we told
  participants the following One technical
  problem weve been having with the equipment is
  that the video input to the computer produces a
  sound artifact in the computer program, so your
  pulse rate will result in click from the hard
  drive. These clicks wont sound very much like
  heart rate because the computer averages the
  beats, so theyll sound a bit mechanical. Dont
  pay attention to the clicks. We have tried to
  get rid of them but couldnt figure out how.
• These instructions were part of the cover story
  designed to convince participants that the clicks
  they were going to hear (actually metronome
  clicks) corresponded to their own heart rate.

22
Actual SA Manipulation excitation transfer
counterbalancing

• High-Low-High-Low
• Peddle bike (HR 165 above baseline)
• Rest and fill out questionnaire
• Watch clip 1 (HR 140 above baseline)
• Fill our emotion rating scale 1
• Fill out questionnaires
• Watch clip 2 (HR at baseline)
• Fill out emotion rating scale 2
• Repeat

23
Actual SA Manipulation

• Low-High-Low-High
• Fill out questionnaires
• Watch clip 1 (HR at baseline)
• Fill out emotion rating scale 1
• Peddle bike (HR 165 above baseline)
• Rest and fill out questionnaire
• Watch clip 2 (HR 140 above baseline)
• Fill out emotion rating scale 2
• Repeat

24
Experimental Design
High A
Low A
Cue
Subj.
High B
Low B
High B
Low B
1 2 3 4 5 6
Personal
Situational
Actual SA A Believed SA B
25
Experimental Design
26
Results

• Funnel Questionnaire
• Based on their responses to the funnel
  questionnaire, 20 participants were excluded from
  the analyses because they had indicated awareness
  of the real purpose of the experiment, namely to
  study the effect of SA on emotional experience.
  A roughly equal number of personal (N 9) and
  situational (N 11) cue responders were
  excluded.

27
Results

• DFEM
• The DFEM resulted in classifying 31 participants
  in the Personal cue response category, and 39
  participants in the Situational cue response
  category. The distribution of type of cue
  response is consistent with that found in
  previous research (e.g., Duclos et al., 1989
  Duncan Laird, 1980). All the analyses were
  done with the 22 unaware Personal cue responders
  and the 28 unaware Situational cue responders.

28
Results

• Did the actual SA manipulation work? YES!
• Mean HR in the exercise condition was
  significantly higher than mean HR in the no
  exercise condition, t (48) 8.83, p lt .001
  (two-tailed).
• Mean SCL in the exercise condition was
  significantly higher that mean SCL in the no
  exercise condition, t (48) 2.73, p lt .01
  (two-tailed).

29
Results

• Did the belief of SA affect actual SA? NO!
• Mean HR in the high believed SA condition (M
  84.93, SD 16.72) did not differ from mean HR
  in the low believed SA condition (M 84.38, SD
  15.13), t (28) .44, p .66.

30
Results

• Subjective Experience of Fear
• An overall 4-way ANOVA was performed to test for
  main effects and interactions among two
  between-subjects factors, i.e. cue response and
  film clip order.
• No main effects were found for cue response, F
  (1, 42) .03, p .87, or for film clip order, F
  (3, 42) .31, p .82. There was no significant
  interaction as well.
• Film clip order was excluded from the analysis.

31
Results

• Subjective Experience of Fear
• The effects of the within-subjects factors, i.e.
  actual and believed SA, crossed with type of cue
  response, were tested using a 3-way ANOVA.
  Results indicated no significant main effects for
  any of the three variables.
• Cue response F (1, 48) .05, p .82.
• (Overall, personal cue responders did not differ
  from situational cue responders in their
  self-report of fear)
• Believed SA F (1, 48) .15, p .70
• Actual SA F (1, 48) .02, p .88.
• (Believed and actual SA, taken separately, did
  not result in differential fear ratings)

32
Results

• Subjective Experience of Fear
• However, there was a highly significant
• actual SA x believed SA x (cue response)
  interaction,
• F (1, 48) 7.63, p .008, ?² .14,
• observed power .77.

33
Results

• When participants were highly aroused, high
  believed SA had a discounting effect for personal
  cue responders and a augmenting effect for
  situational cue responders. In other words, when
  personal cue responders experienced high actual
  SA and heard fast clicks indicating high SA, they
  reported less fear than when they heard slower
  clicks indicating less SA. On the other hand,
  when situational cue responders experienced high
  actual SA and heard faster clicks, they reported
  more fear. This effect was not observed in the
  low actual SA condition.

34
Results

• To test for the significance of the effect of
  believed SA, two mixed 2-way ANOVAs were
  performed, one for the high and one for the low
  actual SA conditions. Type of cue response was
  the between-subjects factor and actual SA was the
  within-subjects factor. Within the high actual SA
  condition there was a significant (type of cue
  response) x (believed SA) interaction, F (1, 48)
  8.14, p .006, ?² .15. No significant
  effects were found within the low actual SA
  condition, F (1, 48) 1.58, p .21.

35
Results

• Correlation between actual HR and intensity of
  subjective experience

p lt .05 p .08, two-tailed
The difference between the two correlation
coefficients in the exercise condition is
significant, z 3.28, p .001.
36
Conclusions

• Valinss claim that cognitions about SA alone
  influence feelings was not supported.
• There was support for Schachters two-factor
  theory incorporating both autonomic and
  situational determinants.
• The Valins effect can be conceptualized as a
  response to situational cues and thus a sub set
  of the two-factor theory.
• There are individual differences in
  responsiveness to both types of cues
• Positive correlation between manipulated actual
  HR and feelings of fear for personal cue
  responders only (predicted).
• The situational cue of believed SA affected both
  types of cue responders (predicted) in different
  ways (not predicted)
• Valins-type effect for situational cue responders
  (predicted)
• Discounting effect for personal cue responders
  (not predicted)

37
Discussion

• The unexpected interaction, or the finding that
  bodily cues on some level affect both types of
  cue responders. Explanation?
• Foster et al.s (1998) two-stage theory of
  arousal and attraction applied to the present
  case
• Actual SA automatically affects feelings for
  everybody.
• At a later stage, individuals integrate
  additional information from the situation. Here
  this additional information was presented via the
  false HR feedback.
• Lairds Cue Response theory extended to SA
• The additional situational information affected
  both types of cue responders differently
• For situational cue responders the situational
  cue of believed fast HR augmented feelings of
  fear a Valins-type effect.
• For personal cue responders the same information
  lead to a discounting of their feelings of fear
  a misattribution effect consistent with previous
  research
• Nisbett Schachter, 1966
• Duncan Laird, 1988

38
So What?

• Implications of our results for
• Emotion theory
• Our model of emotional experience presents a
  promising way of integrating elements of James
  theory, Schachters theory, Valins position, and
  Lairds individual difference approach.
• Real life
• Combating anxiety with beta blockers would work
  better for personal cue responders.
• Placebo treatments and talk therapy for anxiety
  would work better for situational cue responders.
• Emotional side effects of beta blockers may be
  underestimated because they would be observed
  only for personal cue responders.

39
Future Directions

• Do personal cue responders simply produce more
  bodily cues (e.g., physiological reactions,
  expressive behavior, etc.)?
• Are they more sensitive to such cues?
• Are they more attentive to such cues?
• All of the above?
• The role of cognition
• the nature of situational cues
• factors affecting their strength
• mechanisms through which they affect emotional
  experience and behavior

40
Individual Differences in Pain Perception

• The purpose of this study was to apply
  self-perception theory to pain perception.
• A substantial number of studies have shown that
  people differ consistently across a wide variety
  of feelings and behaviors in responsiveness to
  their own personal, or bodily, activities.
• Similar differences seem likely to exist in
  responses to painful stimuli.
• This study investigated how individual
  differences in the use of personal vs.
  situational cues affects sensitivity and
  tolerance to pain.

41
Hypothesis

• People who are more responsive to personal cues
  will be more sensitive to pain cues, and hence in
  the cold pressor task would detect pain sooner
  and would be able to tolerate less of it.

42
Choice of Method

• Important to separate personal from
  situational cues
• The cold pressor
• Produces substantial increases autonomic
  responses (providing strong personal cues)
• Usually not expected to be very painful
  (providing a situational cue of low pain)

43
Procedure

• Baseline HR and SCL measures
• Cold pressor
• DFEM

44
Results

• Pain expectations
• In a pilot study, 19 people were asked how much
  they thought immersing their hand in ice-cold
  water would hurt, on a 4 point scale from "none"
  to very much". The modal response was the next
  to lowest, a little.
• In addition, analysis of the pain expectation of
  the main sample in the experiment (N 38), taken
  after the cold pressor test, revealed the same
  pattern. That is, 24 out of 38 participants
  retrospectively reported that, coming in to the
  experiment, they had expected the cold pressor
  test to hurt a little or less.

45
Results

• Pain expectations ?² .06

9
6
2
2
Not at all
A little
Quite a bit
Very much
46
Results

• Physiological reactions to the cold pressor test
  HR

After immersion in the ice-cold water, HR
increased significantly (M 89.3, SD 15.4
bpm) as compared to baseline (M 81.7, SD
14.7) for all participants, F (1, 36) 31.9,
plt.001, Eta Square .47. The HR x Cue interaction
was not significant indicating that there were no
differences between the personal and situational
groups F (1, 36) 1.01, p .32.
47
Results

• Physiological reactions to the cold pressor test
  SCL

After immersion in the ice cold water, SCL
increased significantly (M 9.3, SD 4.8
micromhos) as compared to baseline (M 7.8, SD
4.0) for all participants, F (1, 33) 27.4, p lt
.001, Eta Sq. .45. The SCL x Cue interaction
was not significant indicating that there were no
differences between the personal and situational
groups F (1, 33) .25, p .62.
48
Results

• Pain detection time (sec)

Personal cue responders reported detecting pain
significantly earlier (M 14.1 seconds, SD
7.4) than did the situational cue responders (M
32.5, SD 23.0), F (1,36) 8.22, p.007, Eta
Square .20.
49
Results

• Pain tolerance time (sec)

The personal cue group tolerated pain for a much
shorter time (M 52.5 seconds, SD 50.6) than
the situational group (M 97.9 seconds, SD
76.0), F (1,36)4.12, p.05, Eta Square.11
50
Conclusions

• Results of the present study indicate that
  participants who were more likely to feel happy
  when they adopted smiles and angry when they
  adopted frowns also reported feelings of pain
  earlier, and tolerated painful stimulation for a
  shorter length of time. These results are just
  what a self-perception theory analysis of pain
  experience predicted, and suggest that
  differences in pain experience may be related to
  more general differences in the processes which
  generate feelings.