ArticlePDF Available

The Heart of the Story: Peripheral Physiology During Narrative Exposure Predicts Charitable Giving.

Authors:
Biological
Psychology
105
(2015)
138–143
Contents
lists
available
at
ScienceDirect
Biological
Psychology
jo
u
r
n
al
homep
age:
www.elsevier.com/locate/biopsycho
The
heart
of
the
story:
Peripheral
physiology
during
narrative
exposure
predicts
charitable
giving
Jorge
A.
Barrazaa,,
Veronika
Alexandera,
Laura
E.
Beavina,
Elizabeth
T.
Terrisa,
Paul
J.
Zaka,b,∗∗
aCenter
for
Neuroeconomics
Studies,
School
of
Social
Science,
Policy,
and
Evaluation,
Claremont
Graduate
University,
United
States
bDepartment
of
Neurology,
Loma
Linda
University
Medical
Center,
United
States
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
5
March
2014
Accepted
13
January
2015
Available
online
21
January
2015
Keywords:
Autonomic
physiology
Hormones
Emotion
Narrative
Heart
rate
variability
Charity
Influence
a
b
s
t
r
a
c
t
Emotionally
laden
narratives
are
often
used
as
persuasive
appeals
by
charitable
organizations.
Physio-
logical
responses
to
a
narrative
may
explain
why
some
people
respond
to
an
appeal
while
others
do
not.
In
this
study
we
tested
whether
autonomic
and
hormonal
activity
during
a
narrative
predict
subsequent
narrative
influence
via
charitable
giving.
Participants
viewed
a
brief
story
of
a
father’s
experience
with
his
2-year-old
son
who
has
terminal
cancer.
After
the
story,
participants
were
presented
with
an
opportunity
to
donate
some
of
their
study
earnings
to
a
related
charity.
Measures
derived
from
cardiac
and
electro-
dermal
activity,
including
HF-HRV,
significantly
predicted
donor
status.
Time-series
GARCH
models
of
physiology
during
the
narrative
further
differentiated
donors
from
non-donors.
Moreover,
cardiac
activ-
ity
and
experienced
concern
were
found
to
covary
from
moment-to-moment
across
the
narrative.
Our
findings
indicate
that
the
physiological
response
to
a
stimulus,
herein
a
narrative,
can
predict
influence
as
indexed
by
stimulus-related
behavior.
©
2015
Elsevier
B.V.
All
rights
reserved.
1.
Introduction
Can
bodily
states
predict
costly
behavior?
The
brain
exerts
control
on
the
body
via
neural
(autonomic)
and
hormonal
(neu-
roendocrine)
systems
(Janig,
2003).
Likewise,
these
systems
relay
information
about
bodily
states
(the
“internal
environment”)
back
to
the
brain.
Neural
states
as
people
are
processing
information
can
be
observed
without
intruding
on
the
experience
of
process
itself
(Falk
et
al.,
2010),
and
have
been
associated
with
objective
influence
outcomes
(Falk,
Berkman,
&
Lieberman,
2012).
In
this
research
we
examine
how
reactivity
in
these
peripheral
systems
can
predict
whether
someone
will
behaviorally
respond
to
a
related
stimulus.
Recent
work
has
associated
the
neuroactive
hormones
adreno-
corticotropin
hormone
(ACTH)
and
oxytocin
(OT)
with
cognitive
(attention)
and
affective
engagement
(empathic
concern)
while
viewing
public
service
announcements
(Lin,
Grewal,
Morin,
Johnson,
&
Zak,
2013).1ACTH
has
long
been
affiliated
with
Corresponding
author.
Tel.:
+1
9099678658.
∗∗ Corresponding
author
at:
Center
for
Neuroeconomics
Studies,
School
of
Social
Science,
Policy,
and
Evaluation,
Claremont
Graduate
University,
United
States.
E-mail
address:
jorge.barraza@cgu.edu
(J.A.
Barraza).
1Unlike
with
Lin
et
al.
(2013),
we
were
unable
to
include
oxytocin
in
our
analysis
as
we
encountered
a
substantial
proportion
of
missing
data
due
to
the
assay
process.
attention
toward
environmental
stimuli
(e.g.,
Born,
Fehm,
&
Voigt,
1986).
Other
steroidal
hormones
are
linked
to
social
behav-
iors.
For
instance,
cortisol
is
hypothesized
to
motivate
action
in
response
to
the
factors
in
the
environment
(see
Dickerson
&
Kemeny,
2004),
including
social
stimuli
(Rahe,
Rubin,
&
Gunderson,
1972).
Testosterone
has
been
shown
respond
to
social
challenges
(Bos,
Panksepp,
Bluthe,
&
van
Honk,
2012)
and
in
the
absence
of
social
threats
increases
prosocial
behavior
(Boksem
et
al.,
2013).
An
extensive
research
suggests
that
both
sympathetic
and
parasympathetic
systems
are
indicative
of
attention
and
affective
engagement.
People
are
more
likely
to
attend
to
stimuli
elicit-
ing
sympathetic
arousal
(see
Boucsein,
2012;
Kensinger,
2004;
MacLeod
&
Mathews,
2004).
Activity
in
both
sympathetic
and
parasympathetic
systems,
via
electrodermal
and
cardiac
activ-
ity,
has
been
shown
to
occur
in
response
to
emotional
stories
(Eisenberg,
Fabes
et
al.,
1988;
Eisenberg,
Schaller
et
al.,
1988;
Eisenberg
et
al.,
1991).
A
key
component
of
the
parasympathetic
nervous
system,
the
vagus
nerve,
is
proposed
to
be
central
to
the
mammalian
“social-engagement
system”
(Porges,
2007).
Whereas
resting
vagal
activity
is
associated
with
affective
experiences,
The
remaining
data
had
such
large
between-
and
within-subject
variation
that
they
were
not
included
in
the
analyses.
http://dx.doi.org/10.1016/j.biopsycho.2015.01.008
0301-0511/©
2015
Elsevier
B.V.
All
rights
reserved.
J.A.
Barraza
et
al.
/
Biological
Psychology
105
(2015)
138–143
139
notably
empathic
concern
(e.g.,
Oveis
et
al.,
2009),
changes
in
vagal
activity
(reactivity)
are
used
as
situational
indicators
of
vagal
con-
trol
(Beauchaine,
2001).
2.
The
present
research
The
present
research
examines
if
reactivity
in
autonomic
and
neuroendocrine
systems
predict
whether
someone
will
act
in
response
to
a
narrative.
As
our
stimulus,
we
selected
a
100-second
narrative.
Narratives
can
serve
as
vehicles
for
transmitting
influ-
ence
by
conveying
a
desired
way
to
feel,
think,
or
act
(Gerrig,
1993).
Narratives
promote
attitude
congruence
(story-consistent
beliefs;
e.g.,
Appel
&
Richter,
2010;
Busselle
&
Bilandzic,
2009;
Green,
2004;
Green
&
Brock,
2000),
a
positive
evaluation
of
information
within
the
narrative
(Escalas,
2004;
Paharia,
Keinan,
Avery,
&
Schor,
2011),
and
identification
with
fictional
groups
in
a
story
(Gabriel
&
Young,
2011).
Narratives
are
successful
at
motivating
costly
behavior.
For
instance,
character-based
appeals
are
found
to
be
a
more
effective
tool
for
eliciting
donations
than
an
information-based
rhetorical
appeal
(Small
&
Loewenstein,
2003).
A
narrative
from
a
chari-
table
organization
was
selected
as
it
provides
a
straightforward
behavioral
outcome
measure:
a
monetary
donation.
Moreover,
a
charity
narrative
permits
us
to
make
explicit
predictions
about
the
psychological
and
physiological
processes
involved
in
narra-
tive
influence.
We
evaluated
whether
cardiac
vagal
control,
heart
rate
(which
reflects
both
sympathetic
and
vagal
influences),
and
electrodermal
activity
as
people
experienced
an
influential
narra-
tive
would
differ
between
the
responders
and
non-responders
to
a
subsequent
donation
appeal.
Furthermore,
we
examined
several
candidate
hormones
hypothesized
to
be
associated
with
attention
to
the
narrative.
3.
Method
3.1.
Participants
and
procedure
We
recruited
163
participants
(68
females)
from
Claremont
colleges
and
the
surrounding
community
through
mass
e-mails,
posted
fliers,
and
an
existing
online
recruitment
pool
(ages
18–52,
M
=
20.91,
SD
=
5.20).
The
general
sample
size
was
determined
assuming
a
medium
effect
size
prior
to
start
of
data
collection.
Par-
ticipant
earnings
varied
with
the
number
of
correctly
answered
post-narrative
questions
and
charitable
donations
made;
maximum
possible
earnings
were
$40.
Study
sessions
were
conducted
at
the
Center
for
Neuroeconomics
Studies
at
Clare-
mont
Graduate
University
in
Claremont,
CA.
Claremont
Graduate
University’s
Institutional
Review
Board
and
the
U.S.
Army
Medical
Research
and
Materiel
Com-
mand’s
Office
of
Research
Protections,
Human
Research
Protection
Office
approved
this
study.
Prior
to
consent,
participants
were
informed
that
the
purpose
of
the
study
was
to
investigate
what
happens
in
your
body
when
you
are
exposed
to
emotional
stories.
The
consent
form
further
informed
participants
that
they
would
see
one
of
several
stories
selected
by
the
researchers,
though
all
participants
viewed
the
same
story.
After
obtaining
written
informed
consent,
12
mL
of
blood
was
drawn
by
a
qualified
phlebotomist
from
an
antecubital
vein
to
establish
basal
hormone
levels
and
par-
ticipants
were
fitted
with
autonomic
physiology
sensors.
Participants
completed
a
questionnaire
that
included
demographic
items
and
a
number
of
state
and
trait
mea-
sures.
Once
finished,
participants
were
seated
privately
in
a
dimly
lit
room
in
front
of
a
15 MacbookPro©laptop
(Apple,
Inc.)
equipped
with
headphones.
All
proceeding
tasks,
including
the
donation
task,
were
presented
in
MATLAB©(Mathworks,
Inc.),
using
the
Psychophysics
Toolbox
extensions
(Brainard,
1997).
After
a
5-min
baseline
acquisition
period
for
autonomic
nervous
system
(ANS)
measures,
participants
watched
a
100-s
video
obtained
with
permission
from
St.
Jude’s
Children’s
Research
Hospital
of
a
father
who
has
a
2-year-old
son
who
is
dying
of
brain
cancer
(used
previously
in
Barraza
&
Zak,
2009).
Peripheral
nervous
system
activity
was
recorded
throughout
the
stimulus.
Post-stimulus,
participants
were
asked
to
rate
their
emotions
using
12
adjectives
previously
used
to
assess
empathic
concern
and
personal
distress
(Batson
et
al.,
1997),
emotions
also
believed
to
be
important
in
narrative
experience
(Mar,
Oatley,
Djikic,
&
Mullin,
2011).
Immediately
after
these
ratings,
participants
received
another
12
mL
blood
draw
in
an
adjacent
room.
Participants
returned
to
their
seats
and
were
asked
to
answer
five
questions
related
to
the
narrative,
earning
$5
for
each
correct
answer.
These
earnings
were
added
to
the
$15
base
participation
payment.
The
earnings
task
was
designed
so
that
participants
earned
money
in
the
study
based
on
effort
rather
than
receiving
a
windfall.
Questions
were
made
to
be
simple
such
that
a
large
majority
of
participants
answered
all
questions
correctly.
Participants
were
next
informed
that
the
preceding
story
was
produced
by
St.
Jude’s
Children’s
Research
Hospital
and
were
given
a
brief
description
regarding
their
activities.
The
option
to
donate
none,
some,
or
all
of
their
participation
earnings
to
St.
Jude’s
was
next
presented
to
participants
in
private
and
with
a
reminder
of
their
anonymity.
After
the
donation
decision,
participants
were
privately
paid
their
earnings
and
dismissed.
There
was
no
deception
of
any
kind
in
this
study
and
donated
money
was
sent
to
St.
Jude’s
at
the
conclusion
of
the
study.
3.2.
Self-report
measures
We
employed
the
Ten-Item
Personality
Inventory
(TIPI;
Gosling,
Rentfrow,
&
Swann,
2003),
to
assess
broad
personality
dimensions
(extraversion,
agreeableness,
conscientiousness,
neuroticism,
openness).
Item
scores
ranged
from
1
“strongly
disagree”
to
7
“strongly
agree”.
Each
subscale
consists
of
two
items;
scale
scores
were
computed
by
averaging
the
respective
item
scores.
The
four
subscales
in
the
Interpersonal
Reactivity
Index
(IRI;
Davis,
1983)
were
used
to
measure
empathic
personality
dimensions
(empathic
concern,
personal
distress,
perspective-taking,
fantasy).
Item
scores
ranged
from
1
“does
not
describe
me
well”
to
7
“describes
me
very
well.”
Subscales
were
computed
by
averaging
the
seven
items
per
subscale.
State
negative
and
positive
affect
was
assessed
using
the
Positive
And
Negative
Affect
Schedule
(PANAS,
Watson
et
al.,
1988).
Item
scores
ranged
from
1
“not
at
all”
to
5
“extremely.”
Positive
affect
and
negative
affect
subscales
were
computed
by
averaging
the
ten
items
per
subscale.
3.3.
Autonomic
measures
Cardiac
(sampling
rate
1
kHz)
and
electrodermal
activity
(sampling
rate
250
Hz)
were
collected
using
a
Biopac
MP150
data
acquisition
system
and
BioNomadix®
transmitters
and
recorded
with
AcqKnowledge®software
version
4.2
(Biopac
Inc.,
Goleta,
CA).
To
measure
cardiac
activity,
participants
were
fitted
with
three
dispos-
able
Ag–AgCl
electrocardiogram
(ECG)
electrodes
using
a
Lead(III)
configuration.
To
measure
skin
conductance,
two
disposable
Ag–AgCl
electrodermal
(EDA)
electrodes
were
placed
on
participants’
distal
phalanx
surfaces
of
the
middle
and
index
fin-
gers
of
their
non-dominant
hand.
Before
placement
of
EDA
electrodes,
participants
washed
hands
with
non-detergent
bar
soap.
Following
data
collection,
the
data
were
manually
inspected
in
AcqKnowledge®
software
version
4.2
(Biopac
Inc.,
Goleta,
CA).
Skin
conductance
waveforms
were
visually
inspected
for
brief
periods
of
signal
loss,
and
data
drop-offs
shorter
than
1
s
in
length
were
replaced
with
averages
from
adjacent
parts
of
the
waveform.
Addi-
tionally,
waveform
noise
due
to
experimenter-observed
movement
was
smoothed
using
mean-value
replacement
from
adjacent
parts
of
the
waveform.
Next,
a
10-Hz
low-pass
filter
was
applied
to
the
waveform
to
remove
high-frequency
noise
(Norris,
Larsen,
&
Cacioppo,
2007),
and
a
square
root
transformation
was
applied
to
adjust
for
skew
inherent
to
skin
conductance
data
(Dawson,
Schell,
&
Filion,
1989;
Figner
&
Murphy,
2001).
After
transformations,
average
skin
conductance
level
(SCL)
was
extracted
for
the
final
2
min
of
the
baseline
and
for
the
100
s
time-span
of
the
narra-
tive.
These
values
were
used
to
calculate
percent
change
in
SCL
from
baseline
to
the
narrative.
For
time
series
analyses,
1
s
segments
of
SCL
were
taken
from
baseline
and
narrative
stimulus.
Non-specific
skin
conductance
responses
(NS-SCRs)
were
iden-
tified
using
a
threshold
of
0.01
S,
and
NS-SCR
counts
were
taken
for
baseline,
and
narrative.
Following
extraction
of
NS-SCR
counts,
these
values
were
used
to
calculate
rate
of
NS-SCRs/min
for
baseline,
narrative,
and
the
three
narrative
segments.
Cardiac
data
from
23
participants
were
excluded
due
to
problems
with
data
collection,
thus
leaving
a
total
of
141
participants
for
further
analysis.
ECG
artifacts
were
manually
removed
from
the
data.
Data
were
further
passed
through
the
band-
pass
finite
impulse
response
(FIR)
filter,
to
remove
both
high-
and
low-frequency
noise,
and
then
smoothed.
R-R
intervals
were
identified
and
extracted
from
Biopac
and
imported
into
Kubios
software
(http://kubios.uef.fi)
for
derivation
of
heart
rate
variability
(HRV)
measures,
including
the
high
frequency
(HF)
component
as
the
measure
of
vagal
control.
Linear
trend
components
were
removed
from
the
data
prior
to
HRV
analysis.
The
HF
power
was
extracted
from
0.12
to
0.4
Hz
band
and
then
log-transformed
as
suggested
by
Lewis,
Furman,
McCool,
and
Porges
(2012).
3.4.
Hormone
measures
Three
hormones
were
assessed
at
baseline
and
immediately
after
narrative
exposure:
adrenocorticotropin
hormone
(ACTH),
cortisol
(CORT),
and
testosterone
(T).
Sessions
were
run
in
the
afternoon
when
diurnal
variations
in
CORT
are
relatively
stable.2Two
8-mL,
EDTA
(ethylenediaminetetraacetic
acid)
whole-blood
tubes
and
one
8-mL,
serum-separator
tube
were
drawn
while
maintaining
a
sterile
field
and
using
a
Vacutainer
butterfly
needle
(BD,
Franklin
Lakes,
NJ,
USA)
at
baseline
and
post-
stimulus.
Following
the
draw,
whole-blood
tubes
were
rocked
to
facilitate
mixing
2Though
each
hormone
follows
a
different
time
course
(e.g.,
de
Wied,
1990;
Dickerson
&
Kemeny,
2004;
Rowe
et
al.,
1974),
we
collected
blood
for
assay
within
1–5
min
of
the
narrative
stimulus
conclusion.
The
collection
point
was
selected
given
the
rapidity
of
changes
in
both
oxytocin
(Fabian
et
al.,
1969)
and
ACTH
(de
Wied,
1990).
140
J.A.
Barraza
et
al.
/
Biological
Psychology
105
(2015)
138–143
and
prevent
coagulation,
and
immediately
placed
onto
ice.
Within
15
min
of
the
draw,
plasma
tubes
were
transferred
from
the
ice
to
centrifuge
at
1500
rpm
for
12
min
at
4C.
Serum
tubes
were
also
rocked
following
the
draw,
and
they
were
placed
at
room
temperature
for
30
min.
Serum
tubes
were
then
transferred
to
the
centrifuge,
where
they
were
spun
at
2300
rpm
for
10
min.
Plasma
and
serum
were
removed
from
the
tubes
with
disposable
pipettes
and
placed
into
2-mL
microtubes
with
screw
caps.
These
tubes
were
immediately
placed
on
dry
ice
and
stored
at
80 C
until
assay.
Four
hormones
were
assayed
using
either
radioimmunoassay
(RIA)
or
enzyme
immunoassay
(EIA)
kits.
Adrenocorticotropin
hormone
(ACTH)
was
assayed
from
plasma
using
two
RIA
kits
produced
by
DiaSorin,
Inc.
(Stillwater,
MN,
USA).
The
inter-
and
intra-assay
coefficients
of
variation
for
the
first
kit
were
15.40%
at
38.70
pg/mL
and
8.63%
at
16.03
pg/mL
(10
replicates),
and
for
the
second
kit
they
were
9.83%
at
111.87
pg/mL
and
2.94%
at
87.77
pg/mL
(10
replicates).
Cortisol
was
assayed
from
serum
using
an
RIA
kit
produced
by
Diagnostic
Systems
Laboratories
(Webster,
TX,
USA).
This
assay
was
performed
using
a
LC–MS
method
developed
by
the
Biomarkers
Core
Laboratory.
Samples
were
treated
with
the
internal
standard
d4-cortisol
pro-
vided
by
CDN
Isotopes
(Pointe-Claire,
Quebec,
Canada).
Testosterone
was
assayed
from
plasma
using
two
EIA
kits
produced
by
ALPCO,
Inc.
(Salem,
NH,
USA).
The
inter-
and
intra-assay
coefficients
of
variation
for
the
first
kit
were
4.73%
at
1.19
ng/mL
and
10.66%
at
1.08
ng/mL,
and
for
the
second
kit
they
were
9.07%
at
3.83
ng/mL
and
8.89%
at
3.48
ng/mL.
After
acetonitrile
extraction,
OT
was
assayed
from
plasma
using
an
RIA
kit
produced
by
Bachem,
Inc.
(Torrance,
CA,
USA).
The
inter-
and
intra-
assay
coefficients
of
variations
for
OT
were
4.58%
and
4%
at
4.69
pg/mL,
respectively.
ACTH,
cortisol,
and
testosterone
were
assayed
at
the
Endocrine
Core
Laboratory
of
the
Yerkes
National
Primate
Research
Center
at
Emory
University
(Atlanta,
GA).
Oxy-
tocin
(OT)
was
assayed
at
the
Reproductive
Endocrine
Research
Laboratory
at
the
University
of
Southern
California
(USC,
Los
Angeles,
CA).
Due
to
the
high
number
of
values
falling
outside
of
the
typical
range
seen
in
the
literature
(see
McCullough,
Churchland,
&
Mendez,
2013)
and
a
high
number
of
values
falling
below
detectable
range
(1
pg/mL),
we
concluded
that
OT
values
were
not
reliable
to
be
included
in
the
analysis.
4.
Results
4.1.
Donations,
personality,
and
post
narrative
affect
Overall,
52%
percent
of
participants
made
donations
(average
donation
$6.94,
SD
=
$6.99).
There
were
no
gender
differences
in
the
decision
to
donate
or
the
amount
donated
(ps
>
0.10).
Donors
rated
themselves
higher
on
the
five-factor
agreeableness
dimension
(M
=
4.94,
SD
=
1.13)
than
non-donors
(M
=
4.53,
SD
=
1.26;
p
=
0.032,
d
=
0.35).
Differences
were
also
found
in
trait
measures
of
empa-
thy,
with
donors
scoring
higher
on
empathic
concern
(donors
M
=
5.36,
SD
=
0.85;
non-donors
M
=
4.88,
SD
=
1.11;
p
=
0.002,
d
=
0.49)
and
perspective-taking
(donors
M
=
5.09,
SD
=
0.82;
non-
donors
M
=
4.72,
SD
=
1.00;
p
=
0.013,
d
=
0.41).
Donors
also
reported
greater
affect
in
response
the
narrative.
After
the
narrative,
donors
reported
greater
concern
(donor
M
=
5.80,
SD
=
1.27;
non-donors
M
=
4.52,
SD
=
1.43;
p
=
0.011,
d
=
0.96),
and
distress
(donor
M
=
5.76,
SD
=
1.06;
non-donor
M
=
5.33,
SD
=
1.18;
p
=
0.02,
d
=
0.39)
than
non-
donors.
About
half
of
participants
earned
the
full
amount
of
40
dollars
USD
(53%;
mean
earnings
=
37.53,
SD
=
2.75).
Donors
and
non-donors
did
not
significantly
differ
in
their
earnings
(donors
M
=
37.62,
SD
=
2.74;
non-donors
M
=
37.43,
SD
=
2.78;
p
=
0.67).
4.2.
Narrative
physiology
Mixed
model
analysis
of
variance
was
used
in
order
to
examine
differences
in
the
physiology
of
donors
and
non-donors
during
narrative
exposure,
with
age
entered
as
a
covariate
(see
Fig.
1).
For
cardiac
measures,
main
effects
show
the
narrative
accelerated
heart
rate,
significantly
decreasing
R-R
interval
across
groups,
F(1,
136)
=
12.9,
p
<
0.001,
2=
0.09,
and
decreasing
HF-HRV,
F(1,
122)
=
8.5,
p
<
0.01,
2=
0.07.
There
was
no
significant
interaction
for
donation
status
(donor/non-donor).
For
electrodermal
measures,
main
effects
results
reveal
the
narrative
significantly
increased
average
skin
conductance
level
across
groups,
F(1,
147)
=
89.99,
p
<
0.001,
2=
0.38,
and
increased
skin
conductance
responses,
F(1,
145)
=
92.73,
p
<
0.001,
2=
0.39.
Interactions
indicate
that,
com-
pared
to
non-donors,
donors
had
higher
sympathetic
activation
Table
1
Logistic
regression
model
predicting
donations
with
delta
change
in
physiology.
Model
and
predictor
ˇ
SE
(ˇ)
Wald
statistic
p
Odds
ratio
Model
1
RR
interval
0.58
0.25
5.41
0.020
1.90
HF-HRV
0.01
0.01
4.24
0.039
1.01
NS-SCR
0.29
0.14
4.66
0.031
1.34
SCL
2.94
0.32
1.14
0.286
0.05
Model
2
RR
interval 0.59 0.26 5.25 0.022 1.79
HF-HRV
0.01
0.01
3.99
0.046
1.01
NS-SCR
0.34
0.15
5.24
0.022
1.41
SCL
3.64
2.88
1.60
0.206
0.03
ACTH
0.01
0.01
0.31
0.578
1.00
Cortisol
0.03
0.36
2.97
0.085
0.97
Testosterone
0.11 0.22 0.02 0.881 0.89
in
both
SCL,
F(1,
147)
=
4.90,
p
<
0.05,
2=
03,
and
NS-SCR,
F(1,
145)
=
12.86,
p
<
0.001,
2=
0.08,
during
narrative
exposure,
but
not
at
baseline.
Across
all
hormone
measures
(cortisol:
pre-narrative
M
=
15.83,
SD
=
7.94,
post-narrative
M
=
13.14,
SD
=
6.95;
ACTH:
pre-narrative
M
=
38.79,
SD
=
22.76,
post-narrative
M
=
42.12,
SD
=
27.89;
testosterone:
pre-narrative
M
=
4.10,
SD
=
4.02,
post-
narrative
M
=
4.09,
SD
=
4.08),
the
only
significant
effect
was
a
decline
in
cortisol
from
baseline
to
narrative,
F(1,
147)
=
19.01,
p
<
0.001,
2=
0.12.
The
average
change
in
cortisol
did
not
have
a
sig-
nificant
difference
between
donor
and
non-donor
groups
(p
>
0.10).
4.3.
Predicting
donations
The
decision
to
donate
to
the
narrative-aligned
charity
was
associated
with
baseline-corrected
autonomic
and
hormonal
mea-
sures
in
a
logistic
regression
(Table
1).
Given
that
most
autonomic
measures
had
a
significant
change
from
baseline
during
narrative
stimulus,
we
entered
autonomic
variables
in
the
first
step
(model
1).
Hormone
measures
were
added
to
the
second
step
to
examine
if
there
was
added
variance
explained.
As
expected,
HF-HRV
sig-
nificantly
predicted
the
decision
to
donate,
odds
ratio
(OR)
=
1.01,
p
=
0.046.
In
addition,
heart
rate
(R-R
interval),
OR
=
1.79,
p
=
0.022,
and
skin
conductance
responses
(NS-CSR),
OR
=
1.41,
p
=
0.022,
were
predictive
of
the
decision
to
donate
within
the
same
model.
None
of
the
endocrine
measures
significantly
predicted
the
deci-
sion
to
donate
(p
>
0.05).
The
results
remained
significant
when
controlling
for
agreeableness
(ˇ
=
0.26,
p
=
0.11),
empathic
concern
(ˇ
=
0.52,
p
=
0.02),
perspective-taking
(ˇ
=
0.63,
p
=
0.01),
gender
(ˇ
=
0.62,
p
=
0.16),
or
age
(ˇ
=
0.02,
p
=
0.65).
4.4.
Experienced
affect
Given
that
the
R-R
interval
was
the
strongest
physiologic
pre-
dictor
of
the
decision
to
donate
within
the
regression
model,
we
set
to
explore
the
relationship
between
heart
rate
and
narrative
experience
further.
Participants
from
a
separate
study
(N
=
45;
age
M
=
24.47,
SD
=
5.89;
63.3%
female)
viewed
the
story
in
5-s
seg-
ments,
providing
a
rating
for
how
much
concern
they
felt
at
every
segment
(i.e.,
“how
much
concern
do
you
feel”).
The
item
was
rated
on
Likert-type
scale
ranging
from
1
(“did
not
feel
this
way
at
all”)
to
7
(felt
this
way
very
much”).
Mean
R-R
interval
levels
and
con-
cern
were
strongly
correlated
from
segment
to
segment
(r
=
0.68,
p
=
0.001;
see
Fig.
2).
Concern
reported
after
the
narrative
was
pos-
itively
associated
with
the
decision
to
donate
(r
=
0.20,
p
=
0.005)
and
the
amount
donated
(r
=
0.19,
p
=
0.009).
Experienced
narrative
distress
was
not
associated
with
donation
behavior
(ps
>
0.10).
J.A.
Barraza
et
al.
/
Biological
Psychology
105
(2015)
138–143
141
Fig.
1.
Mean
physiology
at
baseline
and
during
narrative
exposure
for
donors
(YResp)
and
non-donors
(NResp).
Fig.
2.
Mean
standardized
RR-interval
(with
a
5-s
lag)
and
concern
scores
across
narrative.
4.5.
Time
series
analysis
In
order
to
further
examine
the
physiological
differences
between
donors
and
non-donors,
we
examined
the
cardiac
(R-
R
interval)
physiologic
time
series
averaged
for
each
group
(e.g.,
Bollerslev,
1986;
Greene,
2012;
Hamilton,
1994).
The
data
were
baseline-corrected
and
interpolated
into
1-s
epochs
to
reduce
noise.
We
estimated
both
traditional
(autoregressive
integrated
moving
average,
ARIMA)
and
more
recent
(generalized
autoregres-
sive
conditional
heteroskedasticity,
GARCH)
time
series
models
until
a
best-fit
model
was
identified
(see
Fig.
3).
We
also
tested
for
a
Fig.
3.
Time
series
of
the
RR-interval
for
donors
and
non-donors
across
baseline
and
narrative
video
(baseline
mean
corrected
by
percent
from
1
to
+1).
structural
break
on
narrative
onset
in
order
to
test
if
cardiac
activ-
ity
immediately
following
narrative
presentation
differed
between
donors
and
non-donors.
The
entire
time
series
was
tested
for
stationarity
using
a
Dickey–Fuller
test,
which
showed
the
data
were
stationary
(p
=
0.02
and
p
<
0.001,
respectively
rejects
the
null
hypothesis
of
a
unit
root).
The
best
fitting
model
for
the
donor
group
was
an
ARIMA(1,7),
ARCH(1)
EGARCH(1,2),
with
a
significant
structural
break
at
stim-
ulus
onset.
The
AIC
fit
measure
for
this
model
was
1618,
with
all
coefficients
significant
(p
<
0.01)
and
no
significant
autocorrela-
tions.
The
best
fitting
model
for
the
non-donor
group
was
similar,
an
142
J.A.
Barraza
et
al.
/
Biological
Psychology
105
(2015)
138–143
ARIMA(1,2),
ARCH(1)
EGARCH(1,2,3,4).
This
model’s
AIC
value
was
1506
and
all
estimated
coefficients
were
significant
(p
0.08).
The
time
series
models
show
that
cardiac
activity
in
both
donors
and
non-donors
has
autoregressive
feedback
and
volatility
cluster-
ing.
The
differences
components
of
the
best-fit
time
series
models
are
not
interpretable.
Nevertheless,
these
models
show
that
phys-
iologic
arousal
at
stimulus
onset
differed
between
donors
and
non-donors.
5.
Discussion
The
present
research
examined
the
connection
between
auto-
nomic
and
hormonal
systems
and
behavioral
responses
to
a
persuasive
narrative.
Both
sympathetic
and
parasympathetic
reac-
tivity
during
narrative
exposure
significantly
and
independently
predicted
charitable
giving.
These
findings
persisted
when
con-
trolling
for
personality
traits.
Importantly,
as
shown
by
modeling
the
cardiac
time
series,
autonomic
measures
differed
significantly
across
donors
and
non-donors
within
the
narrative
itself,
indicating
different
reactions
to
particular
elements
of
the
narrative.
Studies
have
reported
heart
rate
acceleration
during
exposure
to
stimuli
that
elicit
positive
affect
(Lang,
Greenwald,
Bradley,
&
Hamm,
1993
and
Bradley
&
Lang,
2000).
One
might
expect
that
donors
would
experience
greater
concern
and
thus
show
increased
cardiac
activity
compared
to
non-donors,
especially
since
empathic
concern
is
classified
as
a
positive
emotion
(Condon
&
Barrett,
2013;
Goetz,
Keltner,
&
Simon-Thomas,
2010).
Empathic
concern,
how-
ever,
is
associated
with
heart
rate
deceleration
(Eisenberg,
Fabes
et
al.,
1988;
Eisenberg,
Schaller
et
al.,
1988).
Heart
rate
decelera-
tion
is
also
observed
during
evocative
films
for
children
who
were
more
willing
to
help
bring
homework
or
donate
some
of
their
par-
ticipation
earnings
to
a
child
in
need
(Eisenberg
et
al.,
1989).
In
our
study,
while
heart
rate
accelerated
relative
to
baseline
for
our
sam-
ple,
heart
rate
appeared
to
decelerate
as
the
narrative
progressed,
as
indexed
by
an
increase
in
the
R-R
interval.
Moreover
we
found
that
across
the
narrative
the
R-R
interval
was
positively
correlated
with
ratings
of
concern
from
an
independent
sample.
While
vagal
control
appeared
to
decline
significantly
during
the
narrative
for
donors
and
non-donors,
we
found
that
vagal
control
significantly
predicted
donor
status.
Prior
research
suggests
that
higher
resting
vagal
activity
is
associated
with
positive
emotions
(Kok
&
Fredrickson,
2010;
Oveis
et
al.,
2009)
and
perceptions
of
helpfulness
by
others
(Eisenberg
et
al.,
1996).
It
is
important
to
note
that
our
behavioral
outcome
was
a
positive
social
behavior
(charita-
ble
giving),
rather
than
tonic
positive
emotions
as
previous
studies.
However,
resting
vagal
activity,
which
can
be
interpreted
as
trait-
like,
was
not
associated
with
our
outcome
measure.
Whereas
tonic
vagal
activity
may
be
associated
with
dispositions
toward
emotion-
ality
(e.g.,
coping,
emotional
regulation,
see
Appelhans
&
Luecken,
2006),
phasic
vagal
control
may
be
a
better
indicator
of
respond-
ing
to
a
specific
stimulus
(e.g.,
Friedman,
Stephens,
&
Thayer,
2014;
Stephens,
Christie,
&
Friedman,
2010).
Electrodermal
activity
significantly
increased
during
narrative
exposure,
and
this
increase
was
more
pronounced
in
the
donor
group.
Moreover,
our
results
show
that
SCR
was
significantly
asso-
ciated
with
donor
status,
but
not
SCL.
Both
of
these
measures
were
significantly
and
positively
correlated
in
this
study
(r
=
0.32)
consis-
tent
with
the
literature
(reported
correlations
range
from
r
=
0.44
to
r
=
0.75;
Boucsein,
2012).
However,
there
is
evidence
that
SCL
and
SCR
are
not
identical
in
their
relation
to
stimuli.
For
instance,
SCRs
may
reflect
the
general
presence
of
highly
arousing,
negatively
tuned
cognitive
activity
while
SCL
may
indicate
general
arousal
(e.g.,
Nikula,
1991).
There
is
some
indication
that
SCRs
are
better
indicators
of
anticipatory
responses
than
SCLs
(e.g.,
Phillips,
Evans,
&
Fearn,
1986).
Our
regression
model
indicates
that
the
differences
in
SCL
between
donors
and
non-donors
during
the
stimulus
may
be
due
to
phasic
SCR
activity.
Endocrine
measures
(basal
or
reactive)
did
not
appear
to
be
associated
with
behavioral
responses.
We
were
unable
to
repli-
cate
the
significant
increase
in
ACTH
after
an
influential
message
reported
in
Lin
et
al.
(2013).
Although
found
an
increase
in
ACTH
from
baseline
to
post-narrative,
the
change
did
not
reach
signifi-
cance
(baseline
=
38.79
pg/mL,
narrative
=
42.16
pg/mL;
two-tailed
t-test,
p
=
0.15).
This
non-replication
could
be
due
to
the
larger
age
distribution
(Lin
et
al.,
age
range
=
18–35)
or
differences
in
the
stim-
ulus
(a
self-relevant,
visceral,
and
negative
stimulus
in
Lin
et
al.).
The
current
research
contributes
to
the
emerging
literature
on
the
neurobiology
of
influence
and
persuasion.
Previous
research
has
shown
that
central
nervous
system
activity
(BOLD
activity
in
medial
prefrontal
cortex)
during
presentation
of
an
anti-smoking
public
service
ads
(PSAs)
is
a
better
predictor
of
population
level
success
of
the
PSA
than
subjective
smoker
ratings
or
even
ratings
from
professionals
(Falk
et
al.,
2012).
We
show
here
that
peripheral
physiology
can
serve
the
same
function.
From
a
practical
stand-
point,
autonomic
measures
are
much
easier
to
collect
and
can
be
done
inside
and
outside
of
the
lab.
However,
autonomic
physiology
does
not
provide
a
fine-grained
view
of
particular
psychological
processes
that
may
be
involved
(e.g.,
affective
versus
cognitive).
Yet,
since
peripheral
neural
systems
coordinate
interactions
with
the
environment,
these
measures
may
be
as
successful
in
capturing
influence
that
leads
to
an
action.
In
short,
physiological
resonance
with
the
environment
may
be
able
to
differentiate
when
some
may
act
where
others
sit
idly
by.
Author’s
contribution
J.A.
Barraza
and
P.J.
Zak
developed
the
study
concept
and
design.
Protocol
testing,
data
collection,
and
data
preparation
was
per-
formed
by
J.A.
Barraza,
L.E.
Beavin,
V.
Alexander,
and
E.
Terris.
Data
analysis
was
performed
by
V.
Alexander
in
consultation
with
J.A.
Barraza
and
P.J.
Zak.
The
manuscript
was
written
by
J.A.
Barraza
and
P.J.
Zak,
with
contributions
to
methods
and
results
sections
by
V.
Alexander.
All
authors
approved
the
final
submitted
version
of
the
manuscript.
Declaration
of
conflicting
interests
The
authors
declared
that
they
had
no
conflicts
of
interest
with
respect
to
their
authorship
or
the
publication
of
this
article.
Funding
This
work
was
supported
by
a
Defense
Advanced
Research
Projects
Agency
Grant
(N11AP20033)
awarded
to
Paul
J.
Zak
(PI)
and
Jorge
A.
Barraza
(co-PI).
The
views,
opinions,
and/or
findings
contained
in
this
article
are
those
of
the
author
and
should
not
be
interpreted
as
representing
the
official
views
or
policies,
either
expressed
or
implied,
of
the
Defense
Advanced
Research
Projects
Agency
or
the
Department
of
Defense.
Acknowledgements
We
are
grateful
to
all
that
have
assisted
in
the
data
collec-
tion
including
Kenneth
Pyle,
Jesse
Bettinger,
Saadullah
Bashir,
Giti
Zahedzadeh,
Sandra
Garcia,
Kevin
Guttenplan,
Jafar
Soltan-
zadeh,
Katya
Vaydylevich,
and
Cristina
Chavez.
Special
thanks
to
Dr.
Patrick
Williams
(Univ.
of
Chicago)
for
programming
assistance
and
Dr.
Michael
Spezio
(Scripps
College)
for
his
advice.
J.A.
Barraza
et
al.
/
Biological
Psychology
105
(2015)
138–143
143
Appendix
A.
Supplementary
data
Supplementary
data
associated
with
this
article
can
be
found,
in
the
online
version,
at
http://dx.doi.org/10.1016/j.biopsycho.2015.
01.008.
References
Appel,
M.,
&
Richter,
T.
(2010).
Transportation
and
need
for
affect
in
narrative
per-
suasion:
A
mediated
moderation
model.
Media
Psychology,
13(2),
101–135.
Appelhans,
B.
M.,
&
Luecken,
L.
J.
(2006).
Heart
rate
variability
as
an
index
of
reg-
ulated
emotional
responding.
Review
of
General
Psychology,
10(3),
229–240.
http://dx.doi.org/10.1037/1089-2680.10.3.229
Barraza,
J.
A.,
&
Zak,
P.
J.
(2009).
Empathy
toward
strangers
triggers
oxytocin
release
and
subsequent
generosity.
Annals
of
the
New
York
Academy
of
Sciences,
1167,
182–189.
Batson,
C.
D.,
Sager,
K.,
Garst,
E.,
Kang,
M.,
Rubchinsky,
K.,
&
Dawson,
K.
(1997).
Is
empathy-induced
helping
due
to
self-other
merging?
Journal
of
Personality
and
Social
Psychology,
73(3),
495.
Beauchaine,
T.
(2001).
Vagal
control,
development,
and
Gray’s
motivational
the-
ory:
Toward
an
integrated
model
of
autonomic
nervous
system
functioning
in
psychopathology.
Development
and
Psychopathology,
13(2),
183–214.
Boksem,
M.
A.,
Mehta,
P.
H.,
Van
den
Bergh,
B.,
van
Son,
V.,
Trautmann,
S.
T.,
Roelofs,
K.,
et
al.
(2013).
Testosterone
inhibits
trust
but
promotes
reciprocity.
Psychological
Science,
24,
2306–2314.
Bollerslev,
T.
(1986).
Generalized
autoregressive
conditional
heteroskedasticity.
Journal
of
Econometrics,
31,
307–327.
Born,
J.,
Fehm,
H.
L.,
&
Voigt,
K.
H.
(1986).
ACTH
and
attention
in
humans:
A
review.
Neuropsychobiology,
15,
165–186.
Bos,
P.
A.,
Panksepp,
J.,
Bluthé,
R.
M.,
&
Honk,
J.
V.
(2012).
Acute
effects
of
steroid
hormones
and
neuropeptides
on
human
social–emotional
behavior:
A
review
of
single
administration
studies.
Frontiers
in
Neuroendocrinology,
33,
17–35.
Boucsein,
W.
(2012).
Electrodermal
activity.
Boston,
MA:
Springer
Science
&
Business
Media.
http://dx.doi.org/10.1007/978-1-4614-1126-0
Bradley,
M.
M.,
&
Lang,
P.
J.
(2000).
Measuring
emotion:
Behavior,
feeling,
and
phys-
iology.
In
R.
D.
Lane,
&
L.
Nadel
(Eds.),
Cognitive
neuroscience
of
emotion
(pp.
242–276).
New
York,
NY:
Oxford
University
Press.
Brainard,
D.
H.
(1997).
The
Psychophysics
Toolbox.
Spatial
Vision,
10,
433–436.
Busselle,
R.,
&
Bilandzic,
H.
(2009).
Measuring
narrative
engagement.
Media
Psychol-
ogy,
12(4),
321–347.
Condon,
P.,
&
Barrett,
L.
F.
(2013).
Conceptualizing
and
experiencing
compassion.
Emotion,
13,
817–821.
Davis,
M.
H.
(1983).
Measuring
individual
differences
in
empathy:
Evidence
for
a
multidimensional
approach.
Journal
of
Personality
&
Social
Psychology,
44,
113–126.
Dawson,
M.
E.,
Schell,
A.
M.,
&
Filion,
D.
L.
(1989).
The
electrodermal
system.
In
J.
T.
Cacioppo,
&
al.
et
(Eds.),
Handbook
of
psychophysiology,
1989
(pp.
200–223).
Cambridge,
UK:
Cambridge
University
Press.
de
Wied,
D.
(1989).
Neurotrophic
effects
of
ACTH/MSH
neuropeptides.
Acta
neuro-
biologiae
experimentalis,
50,
353–366.
Dickerson,
S.
S.,
&
Kemeny,
M.
E.
(2004).
Acute
stressors
and
cortisol
responses:
A
theoretical
integration
and
synthesis
of
laboratory
research.
Psychological
Bul-
letin,
130(3),
355.
Eisenberg,
N.,
Fabes,
R.
A.,
Miller,
P.
A.,
Fultz,
J.,
Shell,
R.,
Mathy,
R.
M.,
et
al.
(1989).
Relation
of
sympathy
and
personal
distress
to
prosocial
behavior:
A
multi-
method
study.
Journal
of
Personality
and
Social
Psychology,
57,