Emotional and non-emotional facial behaviour in patients with unilateral brain damage

Abstract

Aspects of emotional facial expression (responsivity, appropriateness, intensity) were examined in brain-damaged adults with right or left hemisphere cerebrovascular lesions and in normal controls. Subjects were videotaped during experimental procedures designed to elicit emotional facial expression and non-emotional facial movement (paralysis, mobility, praxis). On tasks of emotional facial expression, patients with right hemisphere pathology were less responsive and less appropriate than patients with left hemisphere pathology or normal controls. These results corroborate other research findings that the right cerebral hemisphere is dominant for the expression of facial emotion. Both brain-damaged groups had substantial facial paralysis and impairment in muscular mobility on the hemiface contralateral to site of lesion, and the left brain-damaged group had bucco-facial apraxia. Performance measures of emotional expression and non-emotional movement were uncorrelated, suggesting a dissociation between these two systems of facial behaviour.

Full text

Journal
of
Neurology,
Neurosurgery,
and
Psychiatry
1988;51:826-832
Emotional
and
non-emotional
facial
behaviour
in
patients
with
unilateral
brain
damage
JOAN
C
BOROD,*t
ELISSA
KOFF,$
MARJORIE
PERLMAN
LORCH,§
MARJORIE
NICHOLAS,§
JOAN
WELKOWITZII
From
the
Departments
of
Psychology,
Queens
College,
and
Neurology,
Mount
Sinai
Medical
School,*
The
City
University
of
New
York,
Department
of
Psychiatry,t
New
York
University
Medical
School,
Department
of
Psychology,t
Wellesley
College,
Aphasia
Research
Center,§
Boston
VA
Medical
Center,
and
Department
of
Neurology,
Boston
University
Medical
School,
Department
of
Psychology,
11
New
York
University,
USA
SUMMARY
Aspects
of
emotional
facial
expression
(responsivity,
appropriateness,
intensity)
were
examined
in
brain-damaged
adults
with
right
or
left
hemisphere
cerebrovascular
lesions
and
in
normal
controls.
Subjects
were
videotaped
during
experimental
procedures
designed
to
elicit
emotional
facial
expression
and
non-emotional
facial
movement
(paralysis,
mobility,
praxis).
On
tasks
of
emotional
facial
expression,
patients
with
right
hemisphere
pathology
were
less
responsive
and
less
appropriate
than
patients
with
left
hemisphere
pathology
or
normal
controls.
These
results
corroborate
other
research
findings
that
the
right
cerebral
hemisphere
is
dominant
for
the
expres-
sion
of
facial
emotion.
Both
brain-damaged
groups
had
substantial
facial
paralysis
and
impairment
in
muscular
mobility
on
the
hemiface
contralateral
to
site
of
lesion,
and
the
left
brain-damaged
group
had
bucco-facial
apraxia.
Performance
measures
of
emotional
expression
and
non-emotional
movement
were
uncorrelated,
suggesting
a
dissociation
between
these
two
systems
of
facial
behav-
iour.
The
neuropsychological
mechanisms
involved
in
the
facial
expression
of
emotion
have
been
the
subject
of
a
number
of
recent
investigations.1
2
While
the
major-
ity
of
these
studies
have
used
normal
adult
subjects,
several
have
examined
facial
behaviours
in
brain-
damaged
populations.
Studies
in
normal
subjects
have
documented
that
the
left
side
of
the
face
(or
"hemiface"),
presumably
controlled
by
the
right cere-
bral
hemisphere,3
moves
more
extensively4-6
and
ap-
pears
more
intense7-
10
during
emotional
expression
than
does
the
right
hemiface
(but
see
also").
Studies
of
facial
emotional
expression
in
brain-damaged
pa-
tients,
'2
-15
also
have
suggested
a
special
role
for
the
right
hemisphere
in
emotional
expression.
In
these
studies,
right
brain-damaged
patients,
relative
to
left
brain-damaged
patients
and
normal
controls,
are
more
impaired
in
the
production
of
facial
emotional
expressions.
The
main
purpose
of
this
study
was
to
elucidate
Address
for
reprint
requests:
Dr
Joan
C
Borod,
Department
of
Psy-
chology,
Queens
College,
CUNY
Flushing,
NY
11367,
USA.
Received
21
August
1987
and
in
revised
form
16
December
1987.
Accepted
23
December
1987
aspects
of
facial
emotional
expression
that
have
pre-
viously
been
associated
with
unilateral
brain
damage.
Facial
expressions
from
an
earlier
study'3
were
re-
examined,
and
three
aspects
of
facial
behaviour
were
analysed:
appropriateness,
intensity,
and
re-
sponsivity.
These
parameters
were
selected
because
of
observations
in
the
clinical
literature
that
right
hemi-
sphere
pathology
is
associated
with
inappropriate
affect,'6
flattened
affect,'7
and
low
levels
of
arousal.'8
To
our
knowledge,
there
has
been
no
at-
tempt
to
examine
systematically
these
various
aspects
of
facial
expression
in
a
brain-damaged
population.
As
in
our
previous
work
with
normals,5
we
employed
measures
of
expression
appropriateness
to
address
the
issue
of
inappropriate
affect,
measures
of
intensity
to
address
the
issue
of
"flat
affect,"
and
measures
of
re-
sponsivity
to
address
the
issue
of
decreased
arousal.
In
addition
to
providing
a
more
qualitative
analysis
of
facial
expression
deficits,
the
approach
adopted
for
this
study
permits
an
examination
of
the
inter-
relationships
among
these
various
measures.
It
would
be
important
from
a
theoretical
perspective
to
ascer-
tain
whether
impairments
in
responsivity,
appropri-
ateness,
and
intensity
are
independent
and
separate
from
each
other
or
whether
they
reflect
a
common
826
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

Emotional
and
non-emotionalfacial
behaviour
in
patients
with
unilateral
brain
damage
underlying
deficit
in
facial
emotional
expression.
A
second
purpose
of
this
study
was
to
examine
how
facial
movement
relates
to
the
expression
of
facial
emotion.
While
outcomes
of
facial
emotional
expres-
sion
studies
typically
have
been
ascribed
to
central
(that
is,
cerebral)
processing
mechanisms,
the
possi-
bility
that
these
findings
could
reflect
other
character-
istics
of
the
face
has
to
be
considered.
Studies
of
normal
subjects
have
reported
that
asymmetries
in
emotional
facial
expression
are
not
correlated
with
asymmetries
in
non-emotional
muscular
mobility'9
or
in
hemiface
size.20
In
brain-damaged
patients,
it
remains
to
be
determined
whether
deficits
in
facial
motor
behaviour,
such
as
facial
paralysis,
might
con-
tribute
to
impairment
in
facial
emotional
expression.
Such
deficits
could
affect
the
brain-damaged
patient's
ability
to
execute
tasks
of
facial
emotion,
and
could
influence
the
examiner's
judgements
about
behav-
iours
associated
with
such
tasks.
For
example,
the
face
of
a
patient
with
a
dense
central
facial
paralysis
may
appear
so
distorted
or
unusual
that
any
emo-
tional
expression
would
be
perceived
as
inap-
propriate.
Also,
a
patient
with
restricted
mobility
of
the
facial
musculature
might
produce
emotional
ex-
pressions
of
limited
intensity,
creating
the
impression
of
flattened
affect.
Finally,
a
patient
with
bucco-facial
apraxia
may
be
unable
to
respond
to
a
command
to
use
his
or
her
facial
muscles
to
express
an
emotion.
In
the
current
study,
measures
of
facial
emotion
(under
posed
and
spontaneous
conditions)
were
cor-
related
with
measures
of
related
facial
movements.
Posed
expressions
were
studied
in
a
manner
designed
to
parallel
directed
clinical
examination
of
affect;
spontaneous
expressions
were
elicited
to
provide
a
naturalistic
analogue
for
"clinical
observations"
of
patient
behaviour.
Aspects
of
facial
movement
con-
sidered
in
this
report
were
non-emotional
hemiface
mobility,"2
facial
paralysis,22
and
bucco-facial
apraxia.23
Methods
Subjects
Background
data
Subjects
were
right-handed
(by
self-
or
family-report)
adult
males
with
unilateral
right
brain-
damage
(RBD)
(N
=
12)
or
left
brain-damage
(LBD)
(N
=
15);
16
normal
males
(NC)
served
as
controls.
The
three
subject
groups
did
not
differ
significantly
on
demographic
variables,
with
an
overall
mean
age
of
57
years
(SD
=
8
1),
13
years
of
education
(SD
=
2
8),
and
an
occupational
level
of
3-9
(SD
=
1-5)
as
measured
by
the
Hollingshead-Redlich
seven-point
scale.24
Neurological
data
Patients
were
tested
at
least
one
month
after
onset
of
illness
(median
=
9
months)
and
only
were
included
in
this
study
if
their
lesions
were
the
result
of
a
cerebrovascular
accident
and
if
there
was
a
negative
history
of
psychiatric
disorder,
psychotropic
drug
treatment,
or
sec-
ondary
neurological
disorder
(such
as
epilepsy,
dementia).
Evidence
of
the
unilateral
and
focal
nature
of
the
lesion
was
confirmed
by
CT
in
all
but
two
cases,
where
clinical
data
from
neurological
examination
were
used.
Fourteen
(63%)
of
the
patients
had
lesions
restricted
to
the
cortex
and/or
subcortical
white
matter,
while
the
remainder
had
cortical
lesions
which
extended
to
subcortical
grey
matter
structures
(for
example,
basal
ganglia,
thalamus).
The
patient
groups
were
similar
with
regard
to
the
distributions
of
their
lesions.
The
percentages
of
patients
with
lesions
in
a
particular
re-
gion,
for
RBDs
and
LBDs
respectively
were:
frontal-58%,
60%;
parietal-92%,
67%;
temporal-58%,
67%;
occipital-17%,
7%;
subcortical
white
matter-83%,
60%;
and
subcortical
grey
structures-50%,
27%.
Using
the
Fisher
exact
probability
test
to
examine
the
frequency
of
RBD
and
LBD
patients
with or
without
a
lesion
in
each
region,
none
of
the
comparisons
were
significant.
(Since
the
sensitivity
of
CT
scanning
is
limited,
other
lesions
that
were
not
detectable
by
this
method
may
have
been
present
in
these
patients.)
At
the
time
of
testing,
the
majority
of each
patient
group
had
sensory-motor
deficits
on
the
side
con-
tralateral
to
their
lesion.
These
deficits
were
hemiplegia
(58%
of
RBDs,
60%
of
LBDs);
central
facial
paralysis
(75%
of
RBDs,
80%
of
LBDs);
and
visual
field
loss
(58%
of
RBDs,
47%
of
LBDs).
Fisher
tests
were
conducted,
and,
again,
there
were
no
significant
group
differences.
Although
all
of
the
LBDs
were
aphasic
(3
Broca's,
2
Mixed
Nonfluents, 2
Globals,
2
Conductions,
4
Wernicke's,
2
Unclassifieds),
all
had
adequate
auditory
comprehension
(mean
Boston
Diagnostic
Aphasia
Examination23
z-score
=
-0-88,
SD
=
0
70)
and
cognitive
functioning
(mean
WAIS
Performance
IQ25
=
94,
SD
=
16)
to
perform
the
experi-
mental
tasks.
All
subjects
had
intact
basic
visual
perception,
as
assessed
by
the
ability
to
correctly
describe
two
slides,
that
is,
a
kindergarten
classroom
scene34
and
a
Thematic
Apper-
ception
Test
card.
Dimensions
of
non-emotional
facial
movement
In
order
to
determine
whether
the
brain-damaged
groups
were
equiv-
alent
on
facial
movement,
videotapes
of
subjects'
faces
were
rated
for
facial
paralysis
(resting
face)26
and
hemiface
mo-
bility.21
To
control
for
the
effect
of
central
facial
paralysis,
and
since
the
lower
face
predominantly
receives
con-
tralateral
innervation
while
the
upper
face
receives
bilateral
innervation,3
27-29
the
upper
and
the
lower
face
were
assessed
separately.
Bucco-facial
apraxia
also
was
evaluated
because
of
concern
that
it
might
influence
the
performance
of
LBDs
on
tasks
of
facial
emotional
expression.
Facial
paralysis
For
upper
face
paralysis,
pattern
of
wrin-
kles
in
the
forehead,
position
of
eyebrow,
and
size
of
eyelid
were
assessed;
for
lower
face
paralysis,
depth
of
the
na-
solabial
fold,
position
of
the
corner
of
the
mouth,
and
direc-
tion
to
which
the
middle
of
the
mouth
is
pointing
were
assessed,
according
to
standard
procedures.3032
For
each
of
these
parameters,
asymmetry
was
rated
as
I
(more
paraly-
sis
on
left),
2
(no
asymmetry),
or
3
(more
paralysis
on
right).
Faces
were
rated
by
two
independent
judges,
with
84%
com-
plete
agreement
on
a
subsample
of
67
observations;
the
mean
of
the
two
ratings
was
used
in
data
analysis.
Each
subject
received
one
score
for
the
upper
face
(i
of
the
three
parameters)
and
one
for
the
lower
face
(x
of
the three
par-
ameters).
One-way
analyses
of
variance
(ANOVAs)
on
the
asymmetry
ratings
revealed
no
difference
among
the
three
subject
groups
for
the
upper
face.
For
the
lower
face
827
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

Table
I
Mean
scores
(and
standard
deviations)
for
measures
of
non-emotional
facial
movements,
by
subject
group
Subject
group
Measure
Parameter
Score
range
RBDs
LBDs
NCs
Facial
paralysis
Asymmetry*
Upper
1-3
1
96
(046)
1-98
(055)
1-96
(0
35)
Lower
1-3
1
69
(062)
241
(0-54)
2-08
(0.41)
Hemiface
mobility
Adequacyt
Upper
Right
Hemiface
1-3
2
75
(062)
203
(1-01)
2
81
(0-54)
Left
Hemiface
1-3
2
17
(094)
2-70
(0-59)
2-88
(0
50)
Lower
Right
Hemiface
1-3
247
(0-56)
1-87
(080)
260
(0
37)
Left
Hemiface
1-3
2-08
(0
64)
2-24
(0
56)
2-52
(0
53)
Bucfco-acial
Executiont
1-3
2
80
(0.20)
257
(031)
291
(014)
apraxia
Accuracyt
0-1
094
(0-16)
0-72
(0
25)
091
(0
16)
*From
left-sided
(1)
to
right-sided
(3),
with
symmetry
(2)
as
the
midpoint.
tFrom
impaired
(low
score)
to
unimpaired
(high).
(F
=
6
75;
df
=
2
40;
p
<
0-01),
RBDs
showed
left-sided
paralysis,
LBDs
right-sided
paralysis,
and
NCs
no
paralysis/asymmetry.
Means
for
each
subject
group
are
dis-
played
in
table
1.
Facial
mobility
Subjects
were
instructed
to
make
four
uni-
lateral
facial
movements,
one
involving
eye
(upper
face)
and
three
involving
mouth
(lower
face).21
Each
movement
was
made
five
times
in
succession
on
each
side
of
the
face.
For
each
of
these
movements,
each
hemiface
was
rated
as
1
(in-
adequate),
2
(partially
adequate),
or
3
(adequate).
Two
inde-
pendent
raters
were
in
agreement
for
83%
of
a
subsample
of
80
observations,
and
the
mean
of
the
two
ratings
used
for
data
analysis.
Each
subject
received
one
score
for
the
upper
face
and
one
for
the
lower
face
(x
of
the
three
movements).
Two-way
ANOVAs
were
conducted
on
the
adequacy
ratings
to
examine
the
effects
of
Group
and
Hemiface,
separately
for
the
upper
and
lower
face.
(See
table
I
for
group
means.)
While
there
was
no
main
effect
of
Group
for
the
upper
face,
there
was
an
effect
of
Group
for
the
lower
face
(F
=
3 8
1;
df
=
2-40;
p
<
0
05),
wherein
LBDs
(x
=
2
06)
and
RBDs
(R
=
2
28)
were
more
impaired
than
NCs
(R
=
2
56).
There
was
a
Group
by
hemiface
interaction
for
both
the
upper
(F
=
7-56;
df
=
2
40;
p
<
0
005)
and
the
lower
(F
=
5
02;
df
=
2
40;
p
<
0-01)
faces.
For
both
the
upper
and
the
lower
face,
the
left
Hemiface
was
less
impaired
than
the
right
among
LBDs
while
the
right
was
less
impaired
than
the
left
among
RBDs;
there
were
no
differences
for
NCs.
Facial
apraxia
Subjects
were
requested
to
perform
four
movements:
suck
through
a
straw,
blow
out
a
match,
sniff
a
flower,
and
lick
an
envelope.33
Each
movement
was
rated
for
execution
as
1
(poorly
or
clumsily
executed),
2
(slightly
clumsy),
or
3
(well-executed),
and
for
accuracy
as
0
(inaccu-
rate)
or
I
(accurate).
To
receive
a
score
of
1
on
accuracy,
shape,
movement
(that
is,
direction),
and
location
had
to
be
correct.
Since
interrater
complete
agreement
was
high
for
both
execution
(84%)
and
accuracy
(100%)
on
a
subsample
of
80
observations,
the
mean
of
the
two
ratings
was
used
for
each
movement.
Subjects
received
separate
apraxia
scores
(x
of
the
four
movements)
for
execution
and
accuracy.
Sepa-
rate
one-way
ANOVAs
examined
the
effect
of
Group.
There
was
a
significant
main
effect
of
Group
for
both
execution
(F
=
8-73;
df
2
40;
p
<
0-001)
and
accuracy
(F
=
5-10;
df
=
2
40;
p
<
0-01).
In
both
cases,
LBDs
were
more
impaired
than
RBDs
or
NCs.
(See
table
1
for
group
means.)
Experimental
procedures
Subjects
were
videotaped,
as
part
of
a
larger
study,12
13
while
producing
positive
(that
is,
pleasant)
and
negative
(that
is,
unpleasant)
facial
emotional
expressions,
under
spontaneous
and
posed
conditions.
Valence
categories
(posi-
tive,
negative)
were
based
on
findings
from
an
earlier
study
of
normal
adults'
19
in
which
facial
expressions
were
rated
by
trained
judges
for
the
degree
of
pleasant
versus
unpleas-
ant
affect.
For
the
spontaneous
condition,
slides
designed
to
elicit
expressions
of
positive
and
negative
emotion34
were
presented
to
the
subject
on
a
Singer
Caramate
projector.
There
were
10
positive
and
six
negative
slides.
Positive
slides
were
selected
from
each
of
four
categories;
pleasant
scenes,
scenic
views,
sexual
material,
and
familiar
people.
Negative
slides
were
selected
from
each
of
three
categories:
disgusting
materials,
sad
scenes,
and
unusual
photographs.
Each
cate-
gory
was
represented
by
two
slides,
except
for
familiar
peo-
ple
where
there
were
four
slides.
For
example,
one
positive
slide
showed
a
little
baby
picking
flowers,
and
one
negative
slide
depicted
a
surgical
procedure.
Subjects
were
seated
facing
a
one-way
mirror,
behind
which
a
videocamera
recorded
their
facial
expressions.
Each
of
the
slides
was
presented
for
20
seconds;
after
the
first
6
seconds,
the
subject
was
requested
to
describe
his
feelings
about
and
reactions
to
the
slide.
For
the
posed
condition,
subjects
were
requested
to
delib-
erately
pose
three
positive
(happiness,
pleasant
surprise,
sex-
ual
arousal)
and
five
negative
(sadness,
anger,
fear,
disgust,
confusion)
emotional
expressions
to
oral
command,
such
as
"Look
happy."
The
cue
for
the
subject
to
actually
pose
the
expression
was
"Ready,
go."
Expressions
were
posed
two
times,
once
in
response
to
a
verbal
command
and
once
in
response
to
visual
imitation.
Both
auditory
verbal
(oral
vignettes)
and
visual
nonverbal
(photographs
of
proto-
typical
expressions)35
instructions
were
used
to
avoid
selec-
tively
priming
one
or
the
other
hemisphere
through
the
input
modality.
To
keep
the
spontaneous
and
posed
conditions
as
parallel
as
possible,
no
encouragement
was
given
to
max-
imise
or
intensify
responses.
Two
judges
viewed
the
videotapes
and
rated
each
facial
expression
for
responsivity,
appropriateness,
and
intensity.
For
the
spontaneous
condition,
the
first
complete
facial
ex-
pression
following
the
onset
of
each
slide
was
rated.
For
the
D
-
"Urod,
Koff,
Lorch,
Nicholas,
Welkowitz
828
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

Emotional
and
non-emotionalfacial
behaviour
in
patients
with
unilateral
brain
damage
posed
condition,
the
expression
following
"Ready,
go"
was
rated.
Half
of
the
videotapes
(containing
both
emotional
and
non-emotional
movements)
were
rated
under
normal
viewing
conditions
and
half
were
rated
under
mirror-image
conditions;
this
was
done
to
control
for
perceiver
bias."
Responsivity
ratings
were
made
on
a
two-point
scale
(0
=
no,
1
=
yes)
based
on
rater
consensus
as
to
whether
an
expression
had
occurred
or
not.
If
an
expression
occurred,
it
was
then
rated
for
appropriateness
and
intensity.
Appropri-
ateness
ratings
were
made
on
a
two-point
scale
(0
=
inap-
propriate,
1
=
appropriate).
For
a
response
to
be
scored
as
appropriate,
consensus
of
the
two
raters
was
necessary,
and
the
facial
expression
had
to
be
appropriate
to
the
stimulus
slide
or
command.
For
the
spontaneous
condition,
the
raters
were
familiarised
with
each
of
the
16
stimulus
slides.
For
the
posed
condition,
the
raters
were
familiarised
with
photo-
graphic
exemplars
of
prototypical
facial
expressions
of
hap-
piness,
surprise,
sadness,
anger,
fear,
disgust,35
36
sexual
arousal,
and
confusion.4"
Overall,
there
was
98%
complete
agreement
for
the
appropriateness
ratings.
Intensity
ratings
were
made
on
a
7-point
scale
from
"minimal"
to
"maximal"
intensity.
Since
interrater
agreement,
determined
on
a
sub-
sample
of
observations,
was
high
(Pearson's
r
=
+
0-74),
a
mean
score
of
the
two
ratings
was
used
for
this
variable.
The
categorisation
of
emotional
expressions
as
positive
or
nega-
tive
was
based
on
a
priori
classification
schemes34
36
and
on
post
hoc
analyses.
(Using
Chi
Square
Tests,
score
distribu-
tions
for
all
individual
expressions
within
a
valence
category
were
compared
to
each
other.
Responsivity,
appropri-
ateness,
and
intensity
were
examined
in
both
posed
and
spontaneous
conditions,
separately
by
subject
group.
Among
the
468
Chi
Square
Tests
conducted,
only
4%
were
significant
and
no
patterns
emerged).
For
each
of
the
three
dependent
variables
(responsivity,
appropriateness,
in-
tensity),
subjects
received
a
separate
mean
score
for
posed
positive
(the
i
of
the
six
expressions
[three
verbal,
three
visual]),
posed
negative
(the
x
of
the
10
expression
[five
ver-
bal,
five
visual]),
spontaneous
positive
(the
i
of
the
10
slides)
and
spontaneous
negative
(x
of
six
slides)
emotions.
Results
Group
comparisons
for
emotional
facial
expression
A
series
of
repeated
measures
analyses
of
variance
(ANOVAs)
were
conducted
on
the
emotional
facial
expression
variables,
that
is,
responsivity,
appropri-
ateness,
and
intensity.
For
each
variable,
the
effects
of
Group
(RBD,
LBD,
NC),
Condition
(Posed,
Sponta-
neous),
and
Valence
(Positive,
Negative)
were
exam-
ined
in
a
3
x
2
x
2
design.
Only
significant
effects
are
reported,
and
the
Newman-Keuls
multiple
com-
parison
procedure38
was
used
for
post
hoc
tests.
Responsivity
The
three-way
ANOVA
conducted
on
the
responsivity
scores
yielded
a
significant
main
effect
of
Group
(F
=
3-64;
df
=
2-40;
p
<
0-05).
Using
post
hoc
tests,
the
RBDs
were
significantly
(p
<
0-05)
less
responsive
than
either
the
LBDs
or
NCs.
(See
table
2
for
group
means.)
There
also
was
a
main
effect
for
Condition
(F
=
69-62;
df
=
1-40;
p
<
0-001),
such
that
subjects
were
more
responsive
in
the
posed
(x
=
0-89)
than
in
the
spontaneous
(i
=
0-48)
condition.
Finally,
there
was
a
Condition
by
Valence
interaction
(F
=
5-42;
df
=
2-40;
p
<
0-05).
Using
post
hoc
tests
to
examine
condition
differences
within
each
valance
category,
for
the
positive
category,
sub-
jects
were
significantly
more
responsive
in
the
posed
(x
=
0-90)
than
the
spontaneous
(x
=
0-43)
condi-
tion;
for
the
negative
category,
again,
subjects
were
significantly
more
responsive
in
the
posed
(x
=
0-87)
than
the
spontaneous
(x
=
0-52)
condition.
Using
post
hoc
tests
to
examine
valence
differences
within
each
condition,
in
the
spontaneous
condition,
sub-
jects
were
significantly
more
responsive
on
negative
(x
=
0-52)
than
positive
(x
=
0-43)
expressions;
there
were
no
differences
for
the
posed
condition.
Appropriateness
The
three-way
ANOVA
conducted
on
the
appropriateness
scores
yielded
a
significant
main
effect
of
Group
(F
=
4-05;
df
=
2-34;
p
<
0-05).
Using
post
hoc
tests,
the
RBDs
were
significantly
(p
<
0-05)
more
impaired
than
the
NCs.
(See
table
2
for
group
means.)
There
also
was
a
significant
interaction
between
Group
and
Valence
(F
=
3-46;
df
=
2-34;
p
<
0-05).
Using
post
hoc
tests
to
examine
group
differences
within
each
valence
category,
RBDs
(i
=
0-78)
were
significantly
(p
<
0-05)
less
appropriate
than
LBDs
(i
=
0-91)
or
NCs
(i
=
0-95)
for
positive
emotions;
the
groups
did
not
differ
for
negative
emo-
tions
(RBDs
=
0-88,
LBDs
=
0-84,
NCs
=
0-93).
Using
post
hoc
tests
to
examine
valence
differences
within
each
group,
RBDs
had
significantly
more
difficulty
with
positive
(x
=
0-78)
than
with
negative
(i
=
0-88)
emotions.
There
were
no
valence
differences
for
LBDs
or
NCs.
Finally,
there
was
a
main
effect
of
Condition
(F
=
19-68;
df
=
1-34;
p
<
0-001),
such
that
subjects
displayed
more
appropriate
expressions
during
the
spontaneous
(i
=
0-95)
than
during
the
posed
(i
=
0-81)
condition.
Table
2
Mean
scores
(and
standard
deviations)
for
measures
of
emotional
facial
expression,
by
subject
group
Subject
group
Measure
Score
range
RBDs
LBDs
NCs
Responsivity*
0-1
0-59
(0-19)
0-74
(0-14)
0-72
(0-14)
Appropriateness*
0-1
0-83
(0-14)
0-87
(0-09)
0-93
(0-06)
Intensityt
1-7
3-38
(060)
3-62
(0-65)
3 19
(0-56)
*From
impaired
(low)
to
unimpaired
(high).
tFrom
minimal
(low)
to
maximal
(high).
829
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

830
Intensity
The
three-way
ANOVA
conducted
on
the
intensity
scores
yielded
significant
main
effects
for
Condition
(F
=
9-66;
df
=
1-34;
p
<
0-005)
and
Valence
(F
=
6-11;
df
=
1-34;
p
<
0-05).
Expressions
were
produced
more
intensely
under
posed
(i
=
3.66)
than
spontaneous
(x
=
3-12)
conditions,
and
for
pos-
itive
(x
=
3
54)
than
negative
(x
=
3.24)
emotions.
There
were
no
significant
main
effects
or
interactions
as
a
function
of
subject
group.
(See
table
2
for
group
means.)
Relationships
between
responsivity,
appropriateness,
and
intensity
To
examine
the
relationships
among
the
three
mea-
sures
of
emotional
facial
expression,
we
inter-
correlated
the
scores
for
responsivity,
appro-
priateness,
and
intensity.
Based
on
the
ANOVA
findings
reported
above
for
emotional
expression,
since
there
was
only
one
Group
difference
as
a
func-
tion
of
condition
or
valence,
data
were
collapsed
across
these
distinctions
into
one
mean
score
for
each
expression
variable.
Spearman
rank-order
correlation
coefficients
were
computed
among
the
three
expres-
sion
variables
separately
for
each
subject
group
and
for
all
subjects.
These
correlation
coefficients
are
dis-
played
in
table
3.
As
can
be
seen
in
table
3,
the
only
significant
correlations
were
between
responsivity
and
intensity.
This
was
the
case
for
both
patient
groups
and
for
all
subjects
together.
This
finding
suggests
that
the
more
responsive
a
subject
was,
the
more
in-
tensely
he
expressed
facial
emotion.
The
relationship
between
emotional
expression
and
non-emotional
movement
To
examine
the
relationship
between
expression
and
movement,
we
correlated
(Spearman
rank-order)
the
scores
for
the
three
emotional
facial
expression
vari-
ables
(that
is,
responsivity,
appropriateness,
intensity)
with
scores
for
the
six
non-emotional
facial
move-
ment
variables:
asymmetry
of
facial
paralysis
(upper,
lower),
adequacy
of
hemiface
mobility
(upper,
lower),
execution
of
praxis
tasks,
and
accuracy
of
praxis
tasks.
This
analysis
was
conducted
only
with
the
brain-damaged
groups.
For
each
expression
variable,
data
were
collapsed
across
condition
and
valence
into
one
mean
score.
For
hemiface
mobility,
the
mean
rat-
ing
of
the
right
and
left
hemifaces
was
used.
RBDs
and
LBDs
were
treated
separately.
Among
the
36
cor-
relations
computed
(18
per
patient
group),
there
was
only
one
significant
(p
<
0O05)
correlation.
(Since
we
were
concerned
that
significant
relationships
might
have
been
obscured
by
collapsing
across
the
right
hemiface
and
the
left
hemiface
for
the
mobility
rat-
ings,
the
data
were
reanalysed
with
separate
scores
for
each
hemiface.
Among
the
24
correlations
computed,
12
per
patient
group,
only
two
were
significant
p
<
Borod,
Koff,
Lorch,
Nicholas,
Welkowitz
Table
3
Spearman
rank-order
correlation
rhos
between
parameters
offacial
emotional
expression,
by
subject
group.
Parameters
Subject
Responsivity
Responsivity
Intensity
group
vs
appropriateness
vs
Intensity
vs
appropriateness
RBDs
0
02
0-67*
0
37
LBDs
014
0-84t
-0-14
NCs
-0-44
0-31
-0-02
All
Ss
-0-06
0-63t
0
01
*p
<
0-05.
tp
<
0-001.
0
05,
and
no
systematic
patterns
emerged.)
The
lack
of
significant
correlations
(median
rho
=
0
11)
sug-
gests
that
the
ability
to
make
non-emotional
facial
movements
is
not
related
to
the
production
of
emo-
tional
facial
expressions.
Discussion
The
major
purpose
of
this
study
was
to
assess
system-
atically
the
effects
of
brain
damge
on
the
expression
of
facial
emotion.
It
was
hypothesised
that
patients
with
unilateral
lesions
of
the
right
hemisphere
would
be
more
impaired
on
tasks
of
emotional
facial
expression
than
patients
with
lesions
of
the
left
hemisphere
and
normal
controls.
Examination
of
facial
responsivity,
appropriateness,
and
intensity
suggested
that
patients
with
right
hemisphere
pathology,
compared
with
pa-
tients
with
left
hemisphere
pathology
or
normal
con-
trols,
were
less
responsive
and
less
appropriate
in
their
expression
of
facial
emotion.
For
the
appropri-
ateness
rating,
this
was
only
the
case
for
the
positive
emotions.
These
findings
were
obtained
for
both
posed
and
spontaneous
expressions.
There
were
no
significant
differences
among
the
three
groups
in
the
intensity
with
which
the
facial
expressions
were
pro-
duced.
The
finding
that
RBDs
were
more
impaired
than
LBDs
and
NCs
in
the
expression
of
facial
emotion
lends
support
to
the
theory
that
the
right
hemisphere
is
dominant
for
emotional
processing.
39-4
While
the
bulk
of
evidence
regarding
the
role
of
the
right
hemi-
sphere
in
the
facial
expression
of
emotion
has
been
based
on
studies
examining
facial
asymmetry
in
nor-
mal
subjects
(for
reviews,
seei
2
42
43),
this
study
ex-
amines
the
emotional
facial
behaviour
of
subjects
with
verified
lesions
in
either
the
right
or
left
cerebral
hemisphere.
When
valence
of
emotional
expression
was
consid-
ered,
there
were
no
systematic
differences
between
positive
and
negative
emotions
across
the
three
de-
pendent
variables.
When
the
interaction
between
va-
lence
and
subject
group
was
examined,
valence
did
not
discriminate
between
patient
groups
either
for
re-
sponsivity
or
for
intensity.
However,
RBDs
were
found
to
be
significantly
less
appropriate
than
LBDs
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

Emotional
and
non-emotionalfacial
behaviour
in
patients
with
unilateral
brain
damage
or
NCs
on
positive
emotions.
This
finding
seems
to
converge
with
findings
by
Gardner
and
his
colleagues45
on
the
deficit
of
right
hemisphere
pa-
tients
for
the
appreciation
of
humorous
materials.
In
general,
our
finding
of
a
selective
deficit
in
RBDs
for
positive
emotion
contrasts
with
the
literature
on
fa-
cial
expression
in
normal
controls
which
has
reported
evidence
for
right
hemisphere
mediation
of
negative
emotion
and
left
hemisphere
or
bilateral
mediation
of
positive
emotion
(see
review
in1).
It
also
contrasts
with
two
previous
clinical
studies
that
also
have
con-
sidered
this
issue.'444
Sackeim
and
colleagues44
found
evidence
for
right
hemisphere
mediation
of
negative
emotions
and
left
hemisphere
mediation
of
positive
emotion,
while
Buck
and
Duffy'4
found
no
significant
differences
between
RBDs
and
LBDs
as
a
function
of
valence.
Differences
in
patient
popu-
lations
(for
example,
aetiology,
lesion
location,
sub-
ject
characteristics)
and
in
methodologies
examining
emotional
behaviour
(such
as,
elicitation
conditons,
stimuli,
rating
schemes)
could
account
for
these
differences
in
results.
Additional
studies
with
larger
numbers
of
subjects
and
more
careful
delineation
of
lesion
location
are
necessary
to
assess
more
precisely
the
relationship
between
valence
and
lesion
site.
When
examining
the
relationships
among
the
three
parameters
of
emotional
expression,
there
was
a
significant
correlation
between
responsivity
and
in-
tensity
but
no
relation
between
appropriateness
and
either
of
these
variables.
The
neuropsychological
literature
suggests
the
possibility
that
the
relationship
between
responsivity
and
intensity
reflects
an
under-
lying
dimension
such
as
arousal.
In
the
current
study,
we
used
responsivity
as
an
indirect
measure
of
arousal.
If
one
has
an
arousal
deficit,
it
is
possible
that
the
amplitude
or
intensity
of
a
behavioural
response
also
may
be
dampened.
The
finding
that
responsivity
was
the
only
variable
for
which
RBDs
were
consis-
tenly
imparied
could
be
interpreted
as
support
for
a
relationship
between
a
deficit
in
arousal
and
some
sort
of
right
hemisphere
dysfunction.
There
is
a
recent
body
of
literature
suggesting
that
arousal
deficits
are
associated
with
the
right
hemisphere
syndrome.
For
example,
right
hemisphere
pathology
has
been
associ-
ated
with
abnormal
patterns
of
autonomic
nervous
system
responding,
as
measured
by
heart
rate46
and
skin
conductance,
18
4'
and
also
with
increases
in
reac-
tion
time.48
49
When
we
examined
the
influence
of
non-emotional
facial
movement
on
the
three
aspects
of
facial
emo-
tional
expression
among
the
brain-damaged
subject
groups,
correlations
were
nonsignificant.
The
brain-
damaged
groups
showed
similar
levels
of
per-
formance
on
measures
of
both
paralysis
and
mobility,
indicating
that
group
differences
in
facial
mobility
were
not
responsible
for
the
lack
of
correlation.
Al-
though
LBDs
did
demonstrate,
as
would
be
expected,
significant
impairments
in
bucco-facial
apraxia,
exe-
cution
and
accuracy
measures
of
apraxia
were
not
systematically
related
to
our
measures
of
facial
emo-
tional
expression.
These
data,
combined
with data
from
related
studies
using
normal
subjects,19
20
argue
for
the
dissociation
between
systems
controlling
facial
emotional
expression
and
non-emotional
facial
move-
ment.
Conclusion
In
conclusion,
we
have
observed
deficits
in
the
re-
sponsivity
and
appropriateness
with
which
emotional
facial
expressions
are
produced
as
a
function
of
right
hemisphere
pathology.
These
findings
support
the
no-
tion
that
the
right
hemisphere
is
dominant
for
certain
aspects
of
facial
emotional
expression.
The
group
differences
for
responsivity
and
the
significant
cor-
relations
between
responsivity
and
intensity
suggest
that
it
may
be
fruitful
for
future
research
to
examine
the
relationship
between
impairments
in
emotional
facial
expression
and
deficits
in
arousal
among
RBDs.
Both
patient
groups
had
central
facial
paralysis
and
deficits
in
non-emotional
muscular
mobility
on
the
hemiface
contralateral
to
the
lesion.
The
LBD
group
was
apractic.
It
was
not
the
case,
however,
that
the
ability
to
carry
out
specific
facial
expressions
or
to
respond
in
emotional
situations
was
affected
by
fac-
tors
such
as
degree
of
facial
paralysis,
muscular
facil-
ity,
or
bucco-facial
apraxia.
This
work
was
supported,
in
part,
by
USPHS
Grant
No
MH37952
to
New
York
University
School
of
Medicine,
USPHS
Grant
No
NS06209
to
Boston
University
School
of
Medicine,
and
Biomedical
Re-
search
Support
Grants
RRO7186-02
to
Wellesley
Col-
lege
and
RR070604
to
Queens
College.
We
are
grateful
to
Dr
Ross
Buck,
Dr
Mary
Hyde,
Jerry
Mar-
tin,
Karen
Olsen,
and
the
Boston
VA
Medical
Center.
This
paper
was
presented,
in
part,
at
the
International
Neuropsychology
Society
Meeting,
Denver,
Colo-
rado,
February,
1986.
References
1
Borod
JC,
Koff
E.
Asymmetries
in
affective
facial
expression.
In:
Fox
N,
Davidson
R,
eds.
The
Psychobiology
of
Affective
Development.
Philadelphia:
L
Erlbaum
and
Associates,
1984.
2
Rinn
WB.
The
neuropsychology
of
facial
expression:
A
review
of
the
neurological
and
psychological
mechanisms
for
producing
facial
expression.
Psychol
Bull
1984;95:52-77.
3
Kuypers
HGJM.
Corticobulbar
connexions
to
the
pons
and
lower
brain-stem
in
man-An
anatomical
study.
Brain
1958;81:364-88.
4
Borod
JC,
Caron
HS.
Facedness
and
emotion
related
to
lateral
dominance,
sex,
and
expression
type.
Neuropsychologia
1980;18:237-42.
5
Borod
JC,
Koff
E,
White
B.
Facial
asymmetry
in
posed
and
spon-
831
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

832
taneous
expressions
of
emotion.
Brain
Cogn
1983;2:165-75.
6
Moscovitch
M,
Olds
J.
Asymmetries
in
spontaneous
facial
expression
and
their
possible
relation
to
hemispheric
special-
ization.
Neuropsychologia
1982;20:71-81.
7
Campbell
R.
Asymmetries
in
interpreting
and
expressing
a
posed
facial
expression.
Cortex
1978;14:327-42.
8
Dopson
WG,
Beckwith
BE,
Tucker
DM,
Bullard-Bates
PC.
Asymmetry
of
facial
expression
in
spontaneous
emotion.
Cortex
1984;20:243-52.
9
Heller
W,
Levy
J.
Perception
and
expression
of
emotion
in
right-
handers
and
left-handers.
Neuropsychologia
1981;19:363-72.
10
Sackeim
HA,
Gur
RC.
Lateral
asymmetry
in
intensity
of
emo-
tional
expression.
Neuropsychologia
1978;16:473-81.
11
Hager
JC,
Ekman
P.
The
asymmetry
of
facial
actions
is
inconsis-
tent
with
models
of
hemispheric
specialization.
Psycho-
physiology
1985;22:307-3
18.
12
Borod
JC,
Koff
E,
Perlman
Lorch
M,
Nicholas
M.
Channels
of
emotional
expression
in
patients
with
unilateral
brain
damage.
Arch
Neurol
1985;42:345-8.
13
Borod
JC,
Koff
E,
Perlman
Lorch
M,
Nicholas
M.
The
expres-
sion
and
perception
of
facial
emotion
in
brain-damaged
patients.
Neuropsychologia
1986;24:169-80.
14
Buck
R,
Duffy
RJ.
Nonverbal
communication
of
affect
in
brain-
damaged
patients.
Cortex
1980;16:351-61.
15
Ross
ED,
Mesulam
M.
Dominant
language
functions
of
the
right
hemisphere?
Prosody
and
emotional
gesturing.
Arch
Neurol
1979;36:144-8.
16
Gardner
H.
The
shattered
mind:
The
person
after
brain
damage.
New
York:
Alfred
A
Knopf,
1975.
17
Gainotti
G.
Emotional
behavior
and
hemispheric
side
of
the
lesion.
Cortex
1972;8:41-55.
18
Heilman
KM,
Schwartz
H,
Watson
RT.
Hypoarousal
in
patients
with
the
neglect
syndrome
and
emotional
indifference.
Neu-
rology
1978;28:229-32.
19
Borod
JC,
Koff
E.
Hemiface
mobility
and
facial
expression
asym-
metry.
Cortex
1983;19:355-61.
20
Sackeim
HA,
Gur
RC.
Asymmetry
in
the
face.
Science
1980;209:834-6.
21
Koff
E,
Borod
JC,
White
B.
Asymmetries
for
hemiface
size
and
mobility.
Neuropsychologia
1981;19:825-30.
22
Gordon
AS,
Friedberg
J.
Current
status
of
testing
for
seventh
nerve
lesions.
Otolaryngol
Clin
North
Am
1978;11:301-24.
23
Goodglass
H,
Kaplan
E.
The
Assessment
of
Aphasia
and
Related
Disorders.
Philadelphia:
Lea and
Febiger,
1972.
24
Hollingshead
AB,
Redlich
FC.
Social
Class
and
Mental
Illness.
New
York:
Wiley,
1958.
25
Wechsler
D.
The
Measurement
and
Appraisal
of
Adult
Intel-
ligence.
Baltimore:
Williams
and
Wilkins,
1958.
26
Borod
JC,
Koff
E,
Perlman
Lorch
M,
Nicholas
M.
Deficits
in
facial
expression
and
movement
as
a
function
of
brain
damage.
In:
Nespoulous
JL,
Perron
P,
Le
Cours
AR,
eds.
The
Biological
Foundations
of
Gestures:
Motor
and
Semiotic
aspects.
Hillsdale
NJ:
L
Erlbaum
Associates,
1986.
27
Chusid
JG,
McDonald
JF.
Correlative
Neuroanatomy
and
Func-
tional
Neurology
(17th
ed.).
Los
Altos
CA:
Lange
Medical
Publications,
1976.
28
De
Jong
RN.
The
Neurologic
Examination
(4th
ed.).
Hagerstown
MD:
Harper
&
Row,
1979.
Borod,
Koff,
Lorch,
Nicholas,
Welkowitz
29
Diamond
C,
Frew
I.
The
Facial
Nerve.
New
York:
Oxford
Uni-
versity
Press,
1979.
30
Devriese
PP.
Experiments
on
the
Facial
Nerve.
Amsterdam:
North-Holland
Publishing
Company,
1972.
31
Miehlke
A.
Surgery
of
the
Facial
Nerve.
Philadelphia
PA:
WB
Saunders,
1973.
32
Van
Gelder
R.
Facing
the
Facial
Nerve.
Unpublished
manuscript.
Amsterdam:
Interfaculteit
Lichamelijke
Opvoeding,
Vrjie
Universiteit,
1981.
33
Albert
ML,
Goodglass
H,
Helm
NA,
Rubens
AB,
Alexander
MP.
Clinical
Aspects
of
Dysphasia.
New
York:
Springer-Verlag,
1981:49.
34
Buck
R.
The
slide
viewing
technique
for
measuring
non-verbal
sending
accuracy:
a
guide
for
replication.
Catal
Sel
Doc
Psy-
chol
1978;8:62.
35
Ekman
P,
Friesen
W.
Unmasking
the
Face.
Englewood
Cliffs
NJ:
Prentice-Hall,
1975.
36
Izard
CD.
The
Face
of
Emotion.
New
York:
Appleton-Century-
Crofts,
1971.
37
Campbell
R.
Asymmetries
in
moving
faces.
Br
J
Psychol
1982;73:95-103.
38
Winer
BJ.
Statistical
Principles
in
Experimental
Design.
New
York:
McGraw-Hill
Book
Company,
1971.
39
Borod
JC,
Koff
E,
Caron
H.
Right
hemispheric
specialization
for
the
expression
and
appreciation
of
emotion:
A
focus
on
the
face.
In:
Perecman
E,
ed.
Cognitive
Processes
in
the
Right
Hemisphere.
New
York:
Academic
Press,
1983.
40
Ley
RG,
Bryden
MP.
Consciousness,
emotion,
and
the
right
hemisphere.
In:
Stevens
R,
Underwood
G,
eds.
Aspects
of
Consciousness.
New
York:
Academic
Press,
1982.
41
Tucker
DM.
Lateral
brain
function,
emotion,
and
concep-
tualization.
Psychol
Bull
1981;89:19-46.
42
Hager
JC.
Asymmetries
in
facial
expression.
In:
Ekman
P,
ed.
Emotion
in
the
Face:
Guidelinesfor
Research
and
an
Integration
of
Findings.
Cambridge:
Cambridge
University
Press,
1982.
43
Sackeim
HA,
Gur
RC.
Facial
asymmetry
and
the
communication
of
emotion.
In:
Cacioppo
JT,
Petty
RE,
eds.
Social
Psycho-
physiology.
New
York:
Guilford
Press,
1982.
44
Sackeim
H,
Greenberg
M,
Weiman
A,
Gur
R,
Hungerbuhler
J,
Geshwind
N.
Functional
brain
asymmetry
in
the
expression
of
positive
and
negative
emotions:
Lateralization
of
insult
in
cases
of
uncontrollable
emotional
outbursts.
Arch
Neurol
1982;19:210-8.
45
Bihrle
AM,
Brownell
HH,
Powelson
JA,
Gardner
H.
Compre-
hension
of
humorous
and
nonhumorous
materials
by
left
and
right
brain-damaged
patients.
Brain
Cogn
1986;5:399-41
1.
46
Boller
F,
Yokoyama
K, Ackles
P,
Hood
P,
Jennings
R.
Lack
of
heart
rate
changes
during
an
attention
task
in
patients
with
right
hemisphere
lesions.
Neurology
1986;36
(Suppl
1):
131.
47
Morrow
L,
Vrtunski
B,
Kim
Y,
Boller
F.
Arousal
responses
to
emotional
stimuli
and
laterality
of
lesion.
Neuropsychologia
198
1;19:65-71.
48
DeRenzi
E,
Faglioni
P.
The
comparative
efficiency
of
intelligence
and
vigilance
in
detecting
hemisphere
damage.
Cortex
1965;1:410-33.
49
Howes
D,
Boller
F.
Simple
reaction
times:
Evidence
for
focal
impairment
from
lesions
of
the
right
hemisphere.
Brain
1975;98:317-32.
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from

doi: 10.1136/jnnp.51.6.826
1988 51: 826-832J Neurol Neurosurg Psychiatry
J C Borod, E Koff, M P Lorch, et al.
unilateral brain damage.
facial behaviour in patients with
Emotional and non-emotional
http://jnnp.bmj.com/content/51/6/826
Updated information and services can be found at:
These include:
References http://jnnp.bmj.com/content/51/6/826#related-urls
Article cited in:
service
Email alerting
the online article.
article. Sign up in the box at the top right corner of
Receive free email alerts when new articles cite this
Notes
http://group.bmj.com/group/rights-licensing/permissions
To request permissions go to:
http://journals.bmj.com/cgi/reprintform
To order reprints go to:
http://group.bmj.com/subscribe/
To subscribe to BMJ go to:
group.bmj.com on July 11, 2011 - Published by jnnp.bmj.comDownloaded from