frontotemporal dementia

[ Pobierz całość w formacie PDF ]
Neuropsychologia 43 (2005) 1673–1687
Impaired recognition of negative facial emotions in
patients with frontotemporal dementia
Diego Fernandez-Duque
a
,
b
,
∗
, Sandra E. Black
b
a
DepartmentofPsychology,VillanovaUniversity,800LancasterAvenue,Villanova,PA19085,USA
b
DepartmentofMedicine,SunnybrookandWomen’sCollegeHealthScienceCentre,UniversityofToronto,Canada
Received 5 January 2005; accepted 13 January 2005
Available online 16 March 2005
Abstract
Patients with behavioral variant of frontotemporal dementia (FTD) have difficulties recognizing facial emotions, a deficit that may contribute
to their impaired social skills. In three experiments, we investigated the FTD deficit in recognition of facial emotions, by comparing six patients
with impaired social conduct, nine Alzheimer’s patients, and 10 age-matched healthy adults. Experiment 1 revealed that FTD patients were
impaired in the recognition of negative facial emotions. Experiment 2 replicated these findings when participants had to determine whether
two faces were expressing the same or different emotions. Experiment 3 was a control study in which participants had to discriminate whether
two faces were of the same sex. In this non-emotional processing task, both patient groups performed worse than normal participants, but
FTD patients performed as well as Alzheimer’s patients. We conclude that FTD patients are impaired in the recognition of negative facial
emotions.
© 2005 Elsevier Ltd. All rights reserved.
Keywords:
Emotion; Face recognition; Neuropsychology; Orbitofrontal cortex
Frontotemporal lobar degeneration encompasses a hetero-
geneous group of dementias with varied clinical and patho-
logical presentations. One of its clinical presentations, the be-
havioral variant of frontotemporal dementia (FTD), is char-
acterized by changes in personality, impaired social skills,
poor decision making, lack of empathy and lack of insight,
implying injury to the orbitofrontal cortex (
McKhann et al.,
2001
;
Mychack, Rosen, & Miller, 2001
;
Neary et al., 1998
).
1
Although it has an insidious onset and a gradual progression,
FTD in this clinical presentation bears close resemblance to
cases of orbitofrontal damage caused by traumatic brain in-
jury (
Rosen et al., 2002
). Those patients are often impaired
not only in social behavior, but also in more basic aspects
of social communication, such as the ability to recognize
facial emotions (
Hornak, Rolls, & Wade, 1996
). Given the
similarities in their impaired social behavior and in anatomi-
cal correlates between the two groups, we hypothesized that
FTD patients, like patients with orbitofrontal lesions, would
be impaired in the recognition of facial emotions.
Besides the clinical implications of FTD, the question
of whether patients with this type of dementia are im-
paired in recognizing facial emotions is important for un-
derstanding the neural architecture underlying emotion and
face processing. Both theoretical and empirical arguments
have been gathered in support of specialized brain areas
that separately recognize facial identity and facial emotion
(
Bruce & Young, 1986
). Thus, some prosopagnosic patients
are sometimes unimpaired at recognizing facial emotions
(
Humphreys, Donnelly, & Riddoch, 1993
;
Tranel, Damasio,
& Damasio, 1988
), and patients with normal recognition of
∗
Corresponding author. Tel.: +1 610 519 6207; fax: +1 610 519 4269.
E-mailaddress:
diego.fernandezduque@villanova.edu
(D. Fernandez-Duque).
1
Several taxonomies exist in the literature on frontotemporal dementia,
and this has sometimes led to confusion. Cases such as the ones described
in this article, in which personality changes are the chief initial symptom,
are sometimes called ‘frontal variant’ of FTD, a label that highlights the
contribution of orbitofrontal cortex to those symptoms (
Keane et al., 2002
).
However, other times they are referred to as ‘temporal variant’, highlight-
ing the contribution of right anterior temporal lobe structures to behavioral
disinhibition (
Rosen et al., 2002
). Some researchers have proposed a classifi-
cation based on clinical features. The cases described in this article belonged
to the behavioral variant of FTD in such a classification, as opposed to the
variants in which progressive language deficits are the main feature (e.g.,
semantic dementia, primary progressive aphasia) (
McKhann et al., 2001
).
0028-3932/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.neuropsychologia.2005.01.005
 1674
D.Fernandez-Duque,S.E.Black/Neuropsychologia43(2005)1673–1687
facial identity sometimes have difficulties recognizing emo-
tional expressions (
Anderson, Spencer, Fullbright, & Phelps,
2000
;
Young et al., 1993
). In functional neuroimaging stud-
ies, emotional and non-emotional facial features activate
different brain areas. The structural aspects of face pro-
cessing activate ventral occipitotemporal areas (
Kanwisher,
McDermott, & Chun, 1997
), while emotional features acti-
vate a network of limbic structures that includes the amyg-
dala, insula, and orbitofrontal cortex (
Blair, Morris, Frith,
Perret, & Dolan, 1999
;
Calder, Lawrence, & Young, 2001
;
Phillips et al., 1997; Whalen et al., 1998
). Those limbic struc-
tures are affected in FTD, while occipitotemporal areas are
relatively spared (
Boccardi et al., 2002; Rosen et al., 2002
).
Thus, it is reasonable to hypothesize that FTD patients will
be impaired in the recognition of facial emotion, but not in
the recognition of non-emotional facial features. On the other
hand, certain brain areas that may be implicated in the recog-
nition of facial emotion, such as somatosensory cortex, are
relatively spared in FTD, raising the possibility that FTD
patients may be capable of normal facial emotion recogni-
tion (
Adolphs, Damasio, Tranel, Cooper, & Damasio, 2000
;
Bocti, Rockel, Roy, Gao, & Black, 2004
).
The issue of specific processing of facial attributes can be
taken a step further by asking whether certain emotions will
be more affected than others. It is a matter of current debate
whether separate brain areas represent individual emotions
such as anger, fear, and disgust, or instead the brain encodes
dimensions such as valence and arousal from which a space
of emotional experiences arise. This debate notwithstanding,
there is some evidence that the limbic structures affected in
FTD are critical for the recognition of many negative emo-
tions (
Adolphs, Tranel, Damasio, & Damasio, 1994
;
Blair
et al., 1999; Calder et al., 2001
;
Harmer, Thilo, Rothwell,
& Goodwin, 2001
;
Hornak, Rolls, & Wade, 1996
). The so-
cial misconduct and personality changes exhibited by FTD
patients also hint at the possibility of a specific impairment
in the perception of emotions. Anecdotal evidence suggests
that FTD patients behave as if they are unable or unwilling
to make appropriate use of the social feedback conveyed in
expressions of anger, sadness, fear or disgust.
The hypothesis that FTD patients will be specifically im-
paired in the recognition of negative emotions is complicated
by the fact that even normal participants have more difficulties
recognizing negative emotions than positive ones (
Ekman &
Friesen, 1975
;
Russell, 1994
). It is unclear whether negative
facial emotions per se are more difficult to recognize, or in-
stead the difference is due to a test stimulus artifact. In either
case, the difference between negative and positive displays
raises the possibility that task difficulty might underlie pa-
tients’ poor performance. In other words, FTD patients, due
to their general cognitive deficits, may be disproportionately
impaired in the most difficult trials, which happen to be the
ones depicting negative emotions (
Rapcsak et al., 2002
). We
addressed this problem in two ways. First, our study included
a group of Alzheimer’s (AD) patients, which was matched to
the FTD group for cognitive ability. If poor recognition of
negative emotions stemmed from pictures of negative emo-
tions posing a more difficult task, then both groups should
be equally impaired. The inclusion of a cognitively impaired
comparison group also minimized the chances of obtaining
ceiling effects, which often muddle the interpretation of inter-
actions. Second, we compared patients’ recognition of ‘dif-
ficult’ and ‘easy’ negative emotions. Past literature reveals
that healthy adults often err in the recognition of facial ex-
pressions of fear, but are almost flawless in the recognition of
facial expressions of anger (
Ekman & Friesen, 1975
;
Rapcsak
et al., 2002
). Thus, a level-of-difficulty account would pre-
dict that FTD patients should be severely impaired in the
recognition of fear (a difficult emotion to recognize) while
being relatively spared in the recognition of anger (an easy
emotion to recognize). An account based on a specific deficit
for negative emotions would predict, instead, that both ‘easy’
and ‘difficult’ negative emotions should pose a challenge for
FTD patients.
The current study builds upon previous studies of facial
emotion recognition in FTD (
Fernandez-Duque & Black,
2002
;
Keane, Calder, Hodges, & Young, 2002
;
Lavenu,
Pasquier, Lebert, Petit, & Van der Linden, 1999
;
Perry et
al., 2001; Rosen et al., 2002
). The evidence from these stud-
ies converges to suggest that the inability to recognize fa-
cial emotions in FTD is caused by an inability to recognize
emotions rather than an inability to recognize facial fea-
tures. In fact, recognition of non-emotional features, such
as face identity, appeared to be relatively unimpaired. How-
ever, these studies did not allow a direct comparison be-
tween emotional and non-emotional tasks because different
stimuli and paradigms were used. Another problem of inter-
pretation stems from ceiling or near-ceiling performance in
many of the non-emotional tasks. This raises the possibility
that the emotional tasks were generally more difficult, which
may explain patients’ poor performance. The argument for a
specific impairment in facial emotion recognition would be
bolstered by increasing the difficulty of the non-emotional
task, thus reducing ceiling effects, and showing group by
task cross-over interactions. Our study aimed to provide such
evidence.
In summary, our study investigated facial emotion recog-
nition in patients with FTD, whether their emotion recog-
nition deficit was most severe for negative emotions, and
whether it could be accounted for by general cognitive
deficits. Experiment 1 asked participants to choose the cor-
rect label for a face displaying a basic emotion. We hypoth-
esized that the FTD group would be impaired relative to the
cognitively matched AD group, that the impairment would
be most severe for negative emotions, and that both ‘easy’
and ‘difficult’ negative emotions would pose a challenge for
patients with FTD. Experiment 2 extended the findings to
a same/different-emotion discrimination with reduced cog-
nitive demands. Experiment 3 provided a measure of non-
emotional facial processing by using a same/different sex
discrimination task. Also, in Experiment 3 we explored the
automatic processing of facial emotions: we hypothesized
D.Fernandez-Duque,S.E.Black/Neuropsychologia43(2005)1673–1687
1675
that performance in the sex discrimination task would be in-
fluenced by the emotion information in the healthy elderly
and AD groups, but not in the FTD group. The results of
the three experiments were largely consistent with our hy-
potheses, and together they support the view that FTD pa-
tients are selectively impaired in the recognition of negative
emotions.
tients with mild dementia were selected, based on a cut-off
score of 20 in the Mini-Mental State Examination.
2
To certify that the AD and FTD groups were matched
for cognitive abilities, patients completed a neuropsycholog-
ical assessment. Five normal participants were also tested
and their performance was compared to the patient groups.
Table 1
shows the results of the neuropsychological tests
for the three groups (for a more detailed description, see
Appendix A
).
As expected, both patient groups were impaired relative to
the normal participants in most domains. More importantly,
however, there was no cognitive domain in which FTD pa-
tients were significantly worse than AD patients. The FTD
group never performed more than one standard deviation be-
low the AD group, and performance by the FTD group was
indistinguishable from the AD group in visuospatial ability
(Line Orientation Task) and in the recognition of unfamiliar
faces (e.g., Benton Face Recognition Task).
Behavioral symptoms were assessed with the Frontal Be-
havioral Inventory (
Kertesz, Nadkarni, Davidson, & Thomas,
2000
), the Neuropysychiatric Inventory area (
Cummings et
al., 1994
), and the Cornell Scale for Depression in Demen-
tia (
Alexopoulos et al., 1988
Alexopoulos, Abrams, Young,
& Shamoian, 1988) (for a more detailed description, see
Appendix A
). All six FTD patients had some signs of neu-
ropsychiatric dysfunction, including disinhibition, aberrant
motor behavior, apathy, and changes in appetite. In contrast,
only two of the nine AD patients had neuropsychiatric prob-
lems. Consistent with the overlap of symptoms between FTD
and depression in terms of apathy, changes in appetite, and
irritability, four FTD patients had high scores in the Cornell
Depression Scale. FTD patients were being treated for de-
pressive symptoms or behavioral abnormalities with SSRIs
(
N
= 4) or atypical neuroleptics (
N
= 2). No patient was psy-
chotic nor met clinical depression criteria at time of testing.
The abnormal scores on the depression symptom scale
raise the question as to whether impaired emotion recognition
may be secondary to depression. However, the patterns of
results found in depressed patients are opposite to the ones
hypothesized for FTD patients in this study. In particular,
depressed patients sometimes show a negative bias, with high
accuracy for labeling sadness and relatively poor accuracy
labeling happiness (
Mandal and Bhattacharya, 1985
Mandal
& Bhattacharya, 1985).
To rule out contributions from other pathologies, MRI
was performed with a 1.5 T GE Signa scanner using stan-
dard protocol (
Callen, Black, Gao, Caldwell, & Szalai, 2001
).
Apart from atrophy consistent with their dementia, the scans
showed no other pathology. Cerebral blood flow was measure
in both patient groups using single-photon emission com-
puted tomography (SPECT). Five of the six FTD patients
1. Experiment 1
Experiment 1 provided an initial assessment of whether
patients with frontotemporal dementia are impaired in the
ability to recognize facial emotions. Faces depicting emotions
were displayed one at a time and participants were instructed
to select the corresponding emotional label.
We also investigated some more specific questions. First,
we asked whether FTD patients’ poor performance could
be accounted for by general cognitive deficits. For this, we
compared FTD and AD groups matched for cognitive impair-
ment. Second, we asked whether emotion recognition in FTD
patients would be most impaired for expressions carrying a
negative valence. To test this, we assessed participants’ re-
sponses to each emotion separately. A third question, related
to the previous ones, was whether poor recognition of nega-
tive emotions could be accounted for by a level-of-difficulty
explanation. A level-of-difficulty explanation would predict
that both patient groups should show a larger impairment to
the most difficult emotions (i.e., the emotions that healthy
elderly have most difficulty with). An explanation based in a
selective deficit of negative emotion recognition would pre-
dict that the impairment should be of similar magnitude for
easy and difficult negative emotions, and be present only in
patients with FTD.
Another question we asked in experiment 1 was whether
FTD patients were capable of categorizing emotions as pos-
itive and negative. For this, we looked at whether errors
crossed emotional valence (e.g., a happy face labeled as sad,
or an angry face labeled as happy). Finally, we explored
whether the error patterns were similar across patient groups,
or instead there were systematic deviations in what different
groups perceived.
1.1. Method
1.1.1. Participants
Six patients with clinical diagnosis of frontotemporal
dementia (FTD), nine patients with clinical diagnosis of
Alzheimer’s disease (AD), and ten age-matched normal par-
ticipants (NCs) participated in the study. All FTD patients
met Lund–Manchester Criteria (
Neary et al., 1998
), and all
the AD patients met criteria for probable Alzheimer’s dis-
ease, as established by the workgroup of the National In-
stitute of Neurological and Communicative Disorders and
Stroke–Alzheimer’s Disease and Related Disorders Associ-
ation (NINCDS–ADRDA) (
McKhann et al., 1984
). Only pa-
2
Patients were recruited primarily through the Cognitive Neurology Unit
at Sunnybrook and Women’s Health Sciences Centre in Toronto, where the
project received approval from the Ethics Board. Consent for participation
in the study was obtained from the patients and their caregivers.
1676
D.Fernandez-Duque,S.E.Black/Neuropsychologia43(2005)1673–1687
Table 1
Demographic, neuropsychiatric, and neuropsychological information
Maximum score
NC
AD
FTD
Age
80
65.1 (8.4)
70.1 (7)
63.7 (6.4)
Sex: male–female ratio
4/6
5/4
5/1
Years of education
15.7 (3.6)
15.9 (3.5)
16.5 (3.8)
Frontal behavioral inventory
72
n/a
18 (11)
37.8 (12)
Neuropsychiatric inventory
144
n/a
14 (18)
31.6 (18)
Cornell scale for depression
38
n/a
9.1 (6.8)
13.6 (8)
MMSE
30
29 (0.7)
24.8 (2.0)
26.5 (2.3)
DRS (total)
144
140 (1.1)
125.7 (9.9)
125.7 (6.3)
Boston naming
30
27.8 (1.3)
21.3 (7.4)
21.6 (6.8)
WAB comprehension
10
9.97 (.07)
9.92 (0.07)
9.82 (0.34)
Verbal fluency (FAS)
47.9 (15)
29.3 (15)
25.2 (12)
Semantic fluency
19 (6)
10.3 (4)
13.2 (5)
Pyramids and Palms
a
52
n/a
n/a
48.8 (3.2)
CVLT acquisition
b
80
46 (7.7)
24.2 (9.7)
30.8 (12.9)
CVLT long delay free recall
80
9.2 (3.4)
0.9 (1.5)
4.4 (3.4)
Line orientation task
30
25.6 (6)
22.2 (5)
20 (9)
Visual memory immediate
41
32 (3)
16.7 (5)
17.7 (4)
Visual memory delayed
41
23.8 (4)
2.7 (3)
3.4 (4)
Forward digit span
12
9.1 (1.6)
9.3 (1.9)
8.3 (2.6)
Backward digit span
12
7.75 (1.7)
6.3 (2.5)
5.5 (2.4)
Trails A
n/a
36.7 (9)
47.5 (17)
36.5 (8)
Trails B
c
n/a
79.2 (22)
178 (85)
117 (45)
B to A ratio
n/a
2.2 (0.4)
3.8 (1.9)
3.3 (1.3)
WCST correct
d
64 44 (9) 39.6 (10) 45.6 (11)
Benton face recognition 54 48.2 (3.7) 42.3 (3) 41.7 (2.1)
MMSE: Mini-Mental State Examination; DRS: Dementia Rating Scale; WAB: Western Aphasia Battery; CVLT: California Verbal Learning Test; WCST:
Wisconsin Card Sorting Task.
a
Cut-off score for impairment is 46.8 (90%).
b
FTD case 4 completed the Hopkins Verbal Learning Test instead of the CVLT, and performed within normal limits.
c
No data were collected for one AD patient, who failed to understand trails B instructions.
d
No data were collected for FTD case 1, as the patient refused to complete the task.
showed frontal temporal hypoperfusion, and eight of the nine
AD patients showed posterior hypoperfusion patterns consis-
tent with AD (
Neary et al., 1987
).
displayed onto the white background of the computer screen.
The emotion labels were displayed in black 26 pt Courier
New font, along each side of the photograph. ‘Sad’, ‘happy’,
and ‘surprised’ appeared from top to bottom on the left side,
‘disgusted’, ‘frightened’, and ‘angry’ were displayed from
top to bottom on the right, and ‘neutral’ was centered below
the photograph. The labels remained on the screen during the
total duration of the experiment. Each label had a response
area delimited by a black rectangle, 7 cm
1.1.2. Equipment
All the experiments were carried out on a Dell Inspiron
laptop computer with Windows 98 operating system, and a
15 in. monitor, set to a screen resolution of 1024
768 pix-
els. Stimulus display and response collection were achieved
using E-prime, a commercial experiment application. Touch
responses were collected by an attachable touchscreen (Ed-
mark Touchwindow E 1014), and relayed to the computer via
a USB connector.
×
3 cm in size. The
border of the rectangles were 2 cm away from the outer border
of the photograph, and there was a 2.5 cm vertical distance
between each rectangle’s borders and those of its neighbors.
×
1.1.4. Procedure
Each participant completed two sessions, on separate
days.
3
In the initial session, participants were taught how
to use the touchscreen and practiced until they reported feel-
ing comfortable with its use. Participants were instructed that
1.1.3. Stimuli
Photographs of neutral faces and the six basic emotions
(sad, happy, surprised, angry, disgusted, frightened) from the
Ekman and Friesen series were selected. For each emotion,
we chose the seven faces that led to highest recognition lev-
els in previously reported norms. For ‘fear’ and ‘disgust’,
an eighth photograph was added after a preliminary study
revealed unusual difficulties in recognizing the emotions de-
picted by one of the photographs in these categories (see Sec-
tion
1.1.5
). Each photograph was 13.5 cm
3
The face recognition tasks reported in this article (Experiments 1–3)
were a subset of a larger battery which also included tasks on theory of
mind, emotional understanding in short vignettes, empathic accuracy in
videotaped interviews, and a set of personality questionnaires. To mini-
mize carry-over effects, the facial recognition tasks were intermixed with
other parts of the battery. The findings from those other tasks are reported
elsewhere (
Fernandez-Duque, Hodges, & Black, 2005
).
×
9 cm in size, had
a gray background surrounded by a thin black frame, and was
 D.Fernandez-Duque,S.E.Black/Neuropsychologia43(2005)1673–1687
1677
responses inside the rectangular area would be recorded and
would trigger a feedback tone. Participants had the option to
report their answer by touch or verbally, in which case the
experimenter entered the response via touchscreen. Partici-
pants were encouraged to make a response in every trial but
accuracy was emphasized over speed. Faces were displayed
one at a time and remained on the screen until response or
for a maximum of 30 s. In the rare occasions in which time
expired before the participant made a response, the trial was
repeated at the end of the session.
There were seven practice trials – one for each emotion –
which were not included in the data analysis. The same seven
photographs were used as practice for all participants. For
each practice trial, the experimenter read the seven labels,
at a rate of 1 s
−
1
, from top to bottom, starting on the left-
hand side (happy, sad, surprised), continuing on the right-
hand side (disgusted, frightened, angry), and finishing on the
bottom (neutral). No accuracy feedback was given during
practice nor during actual testing. The only feedback that
participants received, besides the auditory tone announcing
that a response had been recorded, occurred in practice trials
in which participants selected the ‘neutral’ response. In those
trials, the experimenter said “Remember, we choose neutral
when the face is not showing any emotion. If the face is
showing no emotion, you will choose neutral. If the face is
showing an emotion but you are not sure which one, you will
make a guess from one of the other labels”. We included this
feedback because in previous pilot studies participants would
sometimes choose ‘neutral’ to mean ‘I don’t know’.
During testing, the photographs were presented in random
order. There were a total of 51 test trials per session (eight
trials for fear, eight for disgust, and seven for each of the
other emotions). At various points during the session, the ex-
perimenter would remind participants of the instructions by
saying “how is s/he feeling? Is s/he
(7), neutral 98.6 (4), surprised 98.6 (4), sad 89 (15), disgust
84 (20), fear 83.8 (15), angry 94 (8). Overall performance
was very good at 92.1% accuracy, suggesting that the pho-
tographs we selected depicted highly recognizable emotions.
However, there were two faces (one depicting fear, the other
depicting disgust) that were mislabeled by more than 40%
of participants. To compensate for these unusually difficult
trials, we added one other photograph of fear and one other
photograph of disgust to the stimuli set.
1.2. Results
For each participant, data from the two sessions were ag-
gregated, and an average was calculated for each emotion.
We compared performance across groups in each of the emo-
tions (see
Table 2
). We report mostly non-parametric tests,
which protect against violations of the normal distribution.
Analyses of variance yielded comparable results to the non-
parametric tests, and are reported if they provide additional
information.
There were group differences for emotions of fear, anger,
disgust, and surprise (Kruskal–Wallis non-parametric test, 2
d.f.,
H
6.4,
p
< 0.05). There was also a non-significant trend
for perception of sadness (Kruskal–Wallis non-parametric
test, 2 d.f.,
H
= 4.7,
p
< 0.09). Follow-up analyses revealed
that, relative to age-matched normal participants, FTD pa-
tients were impaired in the recognition of all negative
emotions [anger:
U
= 2.5,
Z
= 3.1,
p
< 0.002; disgust:
U
=9,
Z
= 2.3,
p
< 0.02; fear:
U
= 8.5,
Z
= 2.3,
p
< 0.02; sadness:
U
= 13.5,
Z
= 1.8,
p
< 0.07]. Relative to AD patients, FTD
patients were impaired in the recognition of anger (
U
=6,
Z
= 2.5,
p
< 0.01), disgust (
U
= 8.5,
Z
= 2.2,
p
< 0.03), fear
(
U
= 2.5,
Z
= 2.9,
p
< 0.004), and surprise (
U
= 6.5,
Z
= 2.5,
p
< 0.01). No differences were found between AD patients
and normal participants for any of the emotions (
p
> 0.10).
To explore this question more thoroughly, data were submit-
ted to a mixed analysis of variance that had Group (AD, NC)
as a between-subjects factor and Emotion as a within-subject
factor. This more powerful analysis also failed to reveal a dif-
ference between the two groups,
F
(1, 17) = 0.001, ns, or an
interaction between emotion and group,
F
(6, 102) = 0.8, ns.
Could the impaired recognition of negative emotions be
accounted for by a level-of-difficulty explanation? To explore
this question, we selected the most difficult negative emo-
tion (i.e., the one to which healthy subjects made the most
errors) and the easiest one (i.e., the one to which healthy sub-
jects made the fewest errors). Consistent with previous liter-
ature, these were fear and anger, respectively. Next, we asked
≥
” and then reading the
seven labels in the aforementioned fixed order. Participants
were reminded of the instructions whenever they made sev-
eral errors in a row. Participants who made few errors were
reminded of the instructions approximately three times in
each session.
...
1.1.5. Preliminarystudy
To confirm that the facial emotions in the photographs
we selected were highly recognizable, we conducted a pi-
lot study on 20 undergraduate students from University of
Toronto (mean age: 20 years; S.D. = 2.7). We used the same
procedure described above. Percent accuracy in young adults
was as follows (standard deviation in parenthesis): happy 97
Table 2
Percent correct (and standard deviations) for facial emotion recognition in Experiment 1
Happiness
Neutral
Surprise
Sad
Disgust
Fear
Anger
Average
NC
95 (5
.
7)
88 (20.2)
89 (10.8)
85 (22.1)
90 (8.8)
66 (24.2)
96 (6.9)
87 (9.6)
AD
95 (6
.
0)
92 (7.9)
95 (7.1)
78 (18.0)
88 (7.2)
70 (14.4)
91 (11.8)
87 (5.1)
FTD
100 (0
.
0)
75 (21.2)
75 (21.6)
62 (25.4)
65 (19.1)
34 (16)
55 (27.0)
66 (9.5)
  [ Pobierz całość w formacie PDF ]

zanotowane.pl

doc.pisz.pl

pdf.pisz.pl

radiodx.htw.pl