Recognizing threat: a simple geometric shape activates neural circuitry for threat detection.
ABSTRACT The urgent need to recognize danger quickly has been shown to rely on preferential processing in dedicated neural circuitry. In previous behavioral studies examining the pattern of the face when displaying anger, we found evidence that simple noncontextual geometric shapes containing downward-pointing V-shaped angles activate the perception of threat. We here report that the neural circuitry known to be mobilized by many realistic, contextual threatening displays is also triggered by the simplest form of this V-shaped movement pattern, a downward-pointing triangle. Specifically, we show that simple geometric forms containing only downward-pointing V-shapes elicit greater activation of the amygdala, subgenual anterior cingulate cortex, superior temporal gyrus, and fusiform gyrus, as well as extrastriate visual regions, than do presentations of the identical V-shape pointing upward. Thus, this simple V-shape is capable of activating neural networks instantiating detection of threat and negative affect, suggesting that recognition of potential danger may be based, in part, on very simple, context-free visual cues.
- [show abstract] [hide abstract]
ABSTRACT: Echo-Planar functional magnetic resonance imaging (EP-fMRI) was used to study the activity of the amygdala while three normal female subjects viewed alternating blocks of affectively neutral and affectively negative still pictures. Bilateral activation in the amygdala that was significantly correlated with the changing valence of the visual stimuli was found in all three subjects. These findings are consistent with the large corpus of data from non-human studies suggesting that the amygdala is a key structure for extracting the affective significance from external stimuli. This is the first known report of phasic amygdala activation detected with EP-fMRI in normal human subjects responding to affective stimuli.Neuroreport 08/1996; 7(11):1765-9. · 1.40 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Pathological disturbances of mood may follow a 'bipolar' course, in which normal moods alternate with both depression and mania, or a 'unipolar' course, in which only depression occurs. Both bipolar and unipolar disorders can be heritable illnesses associated with neurochemical, neuroendocrine and autonomic abnormalities. The neurobiological basis for these abnormalities has not been established. Using positron emission tomographic (PET) images of cerebral blood flow and rate of glucose metabolism to measure brain activity, we have now localized an area of abnormally decreased activity in the prefrontal cortex ventral to the genu of the corpus callosum in both familial bipolar depressives and familial unipolar depressives. This decrement in activity was at least partly explained by a corresponding reduction in cortical volume, as magnetic resonance imaging (MRI) demonstrated reductions in the mean grey matter volume in the same area of 39 and 48% in the bipolar and unipolar samples, respectively. This region has previously been implicated in the mediation of emotional and autonomic responses to socially significant or provocative stimuli, and in the modulation of the neurotransmitter systems targeted by antidepressant drugs.Nature 05/1997; 386(6627):824-7. · 38.60 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Commensurate with the importance of rapidly and efficiently evaluating motivationally significant stimuli, humans are probably endowed with distinct faculties and maintain specialized neural structures to enhance their detection. Here we consider that a critical function of the human amygdala is to enhance the perception of stimuli that have emotional significance. Under conditions of limited attention for normal perceptual awareness-that is, the attentional blink-we show that healthy observers demonstrate robust benefits for the perception of verbal stimuli of aversive content compared with stimuli of neutral content. In contrast, a patient with bilateral amygdala damage has no enhanced perception for such aversive stimulus events. Examination of patients with either left or right amygdala resections shows that the enhanced perception of aversive words depends specifically on the left amygdala. All patients comprehend normally the affective meaning of the stimulus events, despite the lack of evidence for enhanced perceptual encoding of these events in patients with left amygdala lesions. Our results reveal a neural substrate for affective influences on perception, indicating that similar neural mechanisms may underlie the affective modulation of both recollective and perceptual experience.Nature 06/2001; 411(6835):305-9. · 38.60 Impact Factor
Recognizing Threat: A Simple Geometric Shape
Activates Neural Circuitry for Threat Detection
Christine L. Larson1, Joel Aronoff2, Issidoros C. Sarinopoulos2,
and David C. Zhu2
& The urgent need to recognize danger quickly has been shown
to rely on preferential processing in dedicated neural circuitry.
In previous behavioral studies examining the pattern of the
face when displaying anger, we found evidence that simple
noncontextual geometric shapes containing downward-pointing
V-shaped angles activate the perception of threat. We here re-
port that the neural circuitry known to be mobilized by many
realistic, contextual threatening displays is also triggered by the
simplest form of this V-shaped movement pattern, a downward-
pointing triangle. Specifically, we show that simple geometric
forms containing only downward-pointing V-shapes elicit greater
activation of the amygdala, subgenual anterior cingulate cortex,
superior temporal gyrus, and fusiform gyrus, as well as extra-
striate visual regions, than do presentations of the identical
V-shape pointing upward. Thus, this simple V-shape is capable of
activating neural networks instantiating detection of threat and
negative affect, suggesting that recognition of potential danger
may be based, in part, on very simple, context-free visual cues. &
The successful regulation of human interaction rests on
the accurate comprehension of the intentions of others.
This basic requirement led Darwin (1872/1998) to suggest
that human facial displays of emotion provided a reliable
communication system that used a common biological
foundation to express and comprehend meanings in
similar ways across all cultures. In order to determine
key stimuli that convey these semantic meanings, and the
mechanisms associated with their recognition, the re-
search reported here used functional magnetic resonance
imaging (fMRI) to determine whether a configurational
pattern initially associated with threatening facial expres-
sions, but stripped of all contextual meaning, leaving only
a simple geometric shape, is sufficient to trigger specific
emotion-related neural circuitry that previously has been
shown to respond to threat-related stimuli.
Darwin’s proposal regarding universal signals con-
tained in emotional expressions found support in major
programs of research directed by Ekman (1973, 2003)
and others that identified the global expressions charac-
teristic of each emotion as well as the exact movements
(Ekman, Friesen, & Hager, 2002) of the facial muscles
through which these displays are formed. Ekman (2003)
also suggested that these displays are identified by sets
of feature detectors that permit the rapid recognition of
a facial expression. Although Ekman reviewed evidence
clarifying the expressions associated with each emotion,
how an emotion is recognized is understood much less
well. Given the evolutionary advantage of rapid detec-
tion of threat (Lundqvist & O ¨hman, 2005; Hansen &
Hansen, 1988), we focused on the mechanisms under-
lying the recognition of facial expressions of anger. As
these investigators have maintained, because of the
survival advantage conferred by early recognition of
potential danger, this process likely relies on neural
circuitry that triggers rapidly, relatively automatically,
and with minimal sensory input (O ¨hman & Mineka,
2001; LeDoux, 2000). Further, as rapid detection is facil-
itated when visual signals of threat share easily iden-
tifiable features, thus reducing the need for thorough
processing of all the features that compose a threatening
stimulus, it would be highly efficient if such appraisal
systems were organized to respond to an overall visual
configuration formed by the facial features, rather than
require the inspection of each movement of each facial
landmark. As Ekman’s proposed feature detectors are
thought to be associated with dedicated neural circuitry
activated by the visual configurations formed when ex-
pressing an emotion, this study focuses on identifying
the most essential underlying visual cues need to ex-
press threat-related emotion.
The behavioral basis for this hypothesis rests on
studies that used multiple methods to isolate key stimuli
that convey affective meaning. Seeking to identify the
specific sign stimuli that conveyed the meaning of anger
and happiness across a wide range of primarily non-
literate tribal cultures, cross-cultural research (Aronoff,
Barclay, & Stevenson, 1988) found that the affective
1University of Wisconsin-Milwaukee,2Michigan State University
D 2008 Massachusetts Institute of TechnologyJournal of Cognitive Neuroscience 21:8, pp. 1523–1535
identity of the display was carried in the overall con-
figuration made by the major landmarks of the face,
rather than by the specific facial features themselves.
Anger was shown to be conveyed by angular and diag-
onal forms made by the facial features (e.g., eyebrows),
especially acute angles pointing downward, and happi-
ness was conveyed by curved patterns. This configura-
tional hypothesis is also supported by Bassili’s (1978)
pioneering point-light experiment, which studied the
overall geometric pattern formed by the movement of
the face as a whole when displaying an emotional ex-
pression. Bassili placed luminescent dots on subjects’
faces and, in a dark room, asked them to assume happy
and angry expressions. When portraying a happy ap-
pearance, the burst of dots expanded outward to form
a rounded shape, whereas in the angry representation,
the points of light imploded downward and inward to
form a V-shaped figure. The ability of the V-shaped fig-
ure (usually representing eyebrows) to convey an an-
gry subjective state is also shown in the many studies
using schematic faces (Lundqvist, Esteves, & O ¨hman,
1999) and confirmed by studies (Lundqvist, Esteves, &
O ¨hman, 2004) which show that V-shaped figures elicit
this emotional meaning even when presented without
any other facial feature. The power of more rounded
facial shapes to convey emotionally positive semantic
meanings (Zebrowitz, 1998; Hildebrandt & Fitzgerald,
1983) is similarly well established.
These initial findings have been confirmed by addi-
tional naturalistic and experimental research (Aronoff,
Woike, & Hyman, 1992; Aronoff et al., 1988), which ex-
amined the affective identity of a wide range of nonrep-
resentational visual stimuli, such as angular or curved
lines, as well as similar geometric shapes using figures far
removed from actual representations of the human face.
Abstract shapes and everyday objects (e.g., watch, sofa)
containing sharp angles of various orientations have also
been found to be less preferred than the similar shapes
or objects containing curved forms (Bar & Neta, 2006).
Bar and Neta (2006) posited that the lower preference
for sharp innocuous objects ‘‘stemmed from a feeling of
threat, and that this feeling was triggered by the sharp-
ness of the angles per se’’ (p. 647). Other work in our
laboratory attempted to identify the key features un-
derlying facial cues of emotion by gradually stripping
away contextual information and reducing the stimuli to
their most fundamental geometric components. In sev-
eral studies using semantic differential scales (Osgood,
Suci, & Tannenbaum, 1957) to record the subjective
meaning elicited by these nonrepresentational visual
stimuli (i.e., the degree of ‘‘badness,’’ ‘‘potency,’’ and
‘‘activity’’ perceived in the stimuli), and following tech-
niques introduced by Tinbergen (Eibl-Eibesfeldt, 1989),
who pioneered the use of models to study signaling de-
vices, we presented increasingly angular, diagonal, or
curvilinear models for examination and found that an-
gular V-shaped figures alone (similar to the angles found
in the eyebrows, cheeks, chin, and jaws in angry ex-
pressions) and rounded figures alone (similar to the
curves found in the cheeks, eyes, and mouth in happy
expressions) conveyed the same affective meanings as
that evoked by actual angry and happy facial represen-
tations (Aronoff, 2006; Aronoff et al., 1988, 1992). Using
the same technique, we recently extended this finding,
reporting that simple geometric shapes are perceived as
having affective value, and furthermore, that the orienta-
tion of the angle is an important determinant of this
value. Specifically, simple shapes containing a downward-
pointing acute angle (e.g., a ‘‘V,’’ a triangle) were rated
by participants as more threatening than the exact same
shapes with the V-angle pointing upward (Larson, Aronoff,
& Stearns, 2007).
Additional support for the configurational hypothe-
sis is provided by the many studies which examine the
efficiency of shapes to signal potential threat. Resting
again on Darwin’s suggestion that speedy detection of
threat confers an evolutionary advantage (Niedenthal &
Kitayama, 1994), Hansen and Hansen (1988) used the
visual search paradigm to show that the search for an
angry face in a crowd of happy faces is more efficient
(i.e., rapid) than the reverse. Subsequent work of this
type, using both real and schematic faces (Horstmann
& Bauland, 2006; O ¨hman, Lundqvist, & Esteves, 2001),
confirmed that angry faces consistently lead to briefer
search times. In an experiment to isolate capture of atten-
tion effects for the pure geometrical form of a downward-
pointing acute angle from other facial features and from
other confounding perceptual factors, we used a simple
triangle whose vertex was pointed either up or down as
our target shape (Larson et al., 2007). Triangles were used
rather than a ‘‘V’’ itself to avoid the possibility that cap-
ture of attention was facilitated because V is a letter,
rather than due to any inherent emotional or attentional
properties. Thus, having a simple shape that varied only
in the orientation of the acute angle, we demonstrated
that a triangle with a downward-pointing vertex is rec-
ognized more rapidly than the identical shape with an
upward-pointing vertex. Further, this study also provided
evidence that the downward-pointing triangle has the
power to elicit sustained attention, in keeping with the
ability of realistic threat-related stimuli to disrupt or
slow performance of ongoing cognitive tasks (Eastwood,
Smilek, & Merikle, 2003; Vuilleumier, Armony, Driver, &
Dolan, 2001; White, 1995).
Thus, both subjectively and attentionally, the simple
downward V-shape appears to function much like a typi-
cal contextually based threat stimulus. These results lead
us to hypothesize the presence of a neural network of
targeted, and thus, highly efficient brain regions that
can identify molar shapes that signal the presence of bi-
ologically relevant affective stimuli. Such configurational
detection mechanisms seem to be a much more parsi-
monious way to account for the decoding of emotional
displays than would be the conjecture of a wide set of
1524Journal of Cognitive NeuroscienceVolume 21, Number 8
specialized neural circuitry isomorphic with each indi-
vidual muscular movement. Thus, the evolutionary ad-
vantage for threat detection may be due, in part, to
facilitated recognition of a simple geometric form, com-
mon to a number of threat-related objects, thereby
minimizing the need to process a stimulus fully and in
context in order to identify a potential threat. For these
reasons, in the present study, we investigate whether
viewing the simple shape of a downward-pointing acute
angle recruits the same neural circuitry as that demon-
strated previously to respond to representational and
contextual cues of threat, such as threat-related facial
expressions, aversive scenes, and phobogenic stimuli.
A large body of research has implicated the amygdala
in directing attention toward biologically relevant, af-
fectively salient information, as well as in the processing
of aversive stimuli (LeDoux, 2000); functions that ren-
der the amygdala particularly important in the detection
of potential threat. Human neuroimaging studies have
demonstrated increased amygdala activation in response
to a number of aversive stimuli, including negatively va-
lenced pictures (Irwin et al., 1996), angry faces (Kesler-
West et al., 2001), and fear faces (Vuilleumier, Armony,
Driver, & Dolan, 2003; Morris et al., 1998). Other work
has further supported the notion that the amygdala in-
stantiates rapid, automatic detection of threat, even
under conditions in which attention is limited (Anderson
& Phelps, 2001) or directed away from threat-related
stimuli (Vuilleumier et al., 2001). Human amygdala acti-
vation has also been detected in response to coarse,
spatially degraded facial cues of threat (Vuilleumier
et al., 2003). Importantly, evidence from fear condition-
ing studies in animals demonstrates that an intact amyg-
dala is sufficient for rats to learn to fear very simple stimuli
(e.g., simple tones) even when input from higher corti-
cal regions is disrupted through lesions of the relevant
modality-specific sensory cortex (Doron & LeDoux, 1999).
Of particular relevance for the present study, Bar and Neta
(2007) found that neutral objects (abstract figures, every-
day objects) containing sharp as compared to curved con-
tours activated the amygdala, among other regions. Similar
to the principles guiding the present work, the authors
interpreted these findings as further support for the no-
tion that sharp, angular objects may signal danger.
In light of recent evidence, we also expected to see
activation in visual sensory pathways in response to the
downward V-shape. In a recent review, Vuilleumier and
Driver (2007) emphasized that affective, not just percep-
tual, properties of visual stimuli can be potent modu-
lators of the visual cortex. Consistent with this premise,
emotional, particularly unpleasant emotional, faces acti-
vate the face-sensitive region of the fusiform gyrus more
strongly than neutral faces (Vuilleumier et al., 2001).
Similarly, unpleasant scenes elicit more robust activation
of the extrastriate cortex than neutral scenes (Sabatinelli,
Lang, Keil, & Bradley, 2007). This enhanced sensory
processing is likely a function of the heightened salience
of these stimuli, which necessitates increased recruit-
ment of attentional resources, including the sensory
cortex. The amygdala has been shown to be crucial to
facilitate this enhanced processing in the ventral visual
pathway (Vuilleumier, Richardson, Armony, Driver, &
Dolan, 2004), and thus, seems to be the key structure
initiating preferential processing of salient, biologically
relevant visual stimuli.
We sought to determine whether the amygdala and
related circuitry is recruited in response to a very simple
geometric shape which is devoid of contextual affective
cues, but has been shown to depict threat. Specifically,
we predicted that shapes with downward-pointing V’s
would elicit greater amygdala activation than the iden-
tical shape whose vertex pointed upward. Consistent
with the work of Vuilleumier and Driver (2007), we also
predicted greater activation in ventral visual processing
regions, as well as increased connectivity between the
amygdala and these areas. Whole-brain fMRI scans were
conducted while participants viewed a large set of images
of three simple geometric shapes: downward-pointing
triangles, upward-pointing triangles, and circles that var-
ied in number, size, and location (see Figure 1) in a block
design. To maintain attention, participants made simple
judgments about the number of shapes depicted in each
image. As with our previous behavioral work, triangles
were used rather than an open V to avoid the possible
Figure 1. fMRI paradigm and
examples of shape stimuli.
Across four runs, 12 blocks
of each of the three stimulus
triangles, and circles, were
presented in a random order.
Each block was 25 sec long
and consisted of 10 different stimuli of the same shape, each presented for 2.5 sec. Each stimulus consisted of one to seven identical shapes.
Shapes of four different sizes were presented, but size was held constant within any given image. The number, size, and location of the shapes
were counterbalanced within a condition and were equated across the three different shape conditions. To ensure that participants maintained
attention to the stimulus presentation, they were asked to press a button indicating whether greater or fewer than four shapes were presented
in each image (no images contained four shapes). Each block was followed by a 15-sec rest period during which a fixation cross was displayed.
Larson et al. 1525
confound of V being a letter, and thus, being more sa-
lient. Upward-pointing triangles were used as a compar-
ison condition in order to test the effects of the same
shape when inverted. Circles provided another control
condition and were selected to test the general effects of
angular compared to nonangular stimuli.
Twenty right-handed, healthy college students participated
in this study and signed consent forms approved by the
Michigan State University Institutional Review Board. Data
from three subjects were discarded, one due to incorrect
positioning of the equipment, another due to irregular
anatomy, and a third due to lack of activation in primary
visual areas suggesting lack of attention during the study.
For the same reason, data from two functional runs were
discarded in one of the remaining subjects, and data
from one functional run were discarded in another subject.
Seventeen subjects (10 men, mean age = 20.6 years,
range = 18–26 years) were included in the data analysis.
Stimuli included 120 unique pictures for each of the three
conditions: circles, upward triangles, and downward tri-
angles (Figure 1). Each of the 120 pictures per condition
was unique in either number of objects, object size, or
object positions. Each picture contained one, two, three,
five, six, or seven outlines of each shape on a white
background. Shapes were presented in four different
sizes, but for each picture the size was held constant.
Finally, the positions were randomized within condition
across all 120 images. Stimuli were matched across con-
ditions such that, for each shape condition, there was
an image that exactly matched the other conditions on
number of objects, object size, and object position. Stim-
uli were displayed in color on a 640 ? 480 LCD monitor
mounted on top of the RF head coil. The LCD subtended
128 ? 168 of visual angle.
A block-design paradigm was controlled by an IFIS-SA
system (Invivo, Gainesville, FL). Two ergonomic keypads
were placed under the hands of each subject. The sub-
ject was requested to press the right index finger button
when there were more than four objects on a picture,
and to press the left index finger button when there
were less than four objects (no image contained exactly
four objects). Accuracy for this judgment was 99.7%
and there were no reaction time differences between
the three shape conditions.1Before entering the scan-
ner, all subjects were trained by viewing a 2-min practice
paradigm so that they became familiar with the task,
and they were asked to pay close attention to the shape
of the objects in each picture. Participants were told that
the study aimed to understand how the visual system
responds to different geometric shapes. The experiment
was divided into four functional runs each lasting 6 min
15 sec. In each run, subjects were presented nine blocks
of visual stimulation after an initial 15-sec ‘‘resting’’ pe-
riod. In each block, 10 unique pictures from one condi-
tion were presented. Within a block, each picture was
presented for 2.5 sec with no interstimulus interval. A
15-sec baseline condition (a white screen with a black
fixation cross at the center) followed each block. Each
condition was shown in three blocks per run. Both the
order of conditions within each run and the order of
pictures within a block were randomly determined. The
four functional runs were presented to eight subjects in
a forward order and others in a reverse order.
Data were collected on a 3-T GE Signa EXCITE scanner
(GE Healthcare, Milwaukee, WI) with an eight-channel
head coil. During each session, images were first ac-
quired for the purpose of localization, followed by first
and higher-order shimming procedures to improve mag-
netic field homogeneity (Kim, Adalsteinsson, Glover, &
Spielman, 2002). To study brain function, echo-planar
images, starting from the most inferior regions of the
brain, were then acquired with the following parame-
ters: 34 contiguous 3-mm axial slices in an interleaved
order, TE = 27.7 msec, TR = 2500 msec, flip angle =
808, FOV = 22 cm, matrix size = 64 ? 64, ramp sam-
pling, and with the first four data points discarded.
Each volume of slices was acquired 146 times during
each of the four functional runs while subjects viewed
the pictures, resulting in a total of 584 volumes of im-
ages over the course of the entire experiment. After
functional data acquisition, high-resolution volumetric
T1-weighted spoiled gradient-recalled images with cere-
brospinal fluid suppressed were obtained to cover the
whole brain with one hundred twenty 1.5-mm sagittal
slices, 88 flip angle, and 24 cm FOV. These images were
used to identify anatomical locations.
fMRI Data Preprocessing and Analysis
All fMRI data preprocessing and analysis was conducted
with AFNI software (Cox, 1996). For each subject, ac-
quisition timing difference was first corrected for dif-
ferent slice locations. With the first functional image as
the reference, rigid-body motion correction was done in
three translational and three rotational directions. The
amount of motion in these directions was estimated
and then the estimations were used in data analysis.
For each subject, spatial blurring with a full width half
1526Journal of Cognitive NeuroscienceVolume 21, Number 8
maximum of 4 mm was applied to reduce random noise
(Parrish, Gitelman, LaBar, & Mesulam, 2000), and also
to reduce the issue of intersubject anatomical variation
and Talairach transformation variation during group analy-
sis. For the group analysis, all images were converted to
Talairach and Tournoux (1988) coordinate space with an
interpolation to 1 mm3voxels.
For analysis of each individual subject, the reference
function throughout all functional runs for each picture
category was generated on the basis of the convolution
of the stimulus input and a gamma function, modeled as
the impulse response when each picture was presented.
The acquired functional data were compared with the
reference functions using the 3dDeconvolve software
for multiple linear regression analysis and general linear
tests (Ward, 2002). Multiple linear regressions were ap-
plied on a voxelwise basis for t-statistic tests and to find
the magnitude change when each picture condition was
presented, compared to the reference functions. BOLD
percent signal change relative to the baseline state was
then calculated. General linear tests were also applied
on a voxelwise basis to find the statistical significance of
pairwise comparisons for all the picture conditions. For
the above analysis, in addition to applying the reference
functions for the three picture conditions, MRI signal
modeling also included the subject motion estimations
in the three translational and the three rotational direc-
tions, and the constant, linear, and quadratic trends for
each of the four functional runs.
Monte Carlo simulation of the effect of matrix and
voxel sizes of the imaging volume, spatial correlation of
voxels, voxel intensity thresholding, masking, and clus-
ter identification was applied to estimate overall statis-
tical significance with respect to the whole brain (Ward,
2000). Because the anterior cingulate cortex (ACC) was
a specific region of interest (ROI), a similar procedure
was carried to estimate the overall statistical significance
with respect to this ROI.
Whole-brain Group Analysis
After the percent signal change was estimated with re-
spect to each picture condition for each subject, an
ANOVA was performed for group analysis with a mixed-
effect two-factor model, with picture condition (three
levels) modeled as a fixed effect and subject modeled
as the second factor as a random effect. The ANOVA
results were used to extract the activated voxels for all
pairwise condition contrasts (voxel-based p value <.005
and whole-brain corrected p value <.023). The active
voxel selection criteria required that the voxels were
nearest-neighbor and within a cluster size of 248 mm3.
Based on application of these criteria to the whole brain
(the medium size of the seventeen brains), the voxel-
based p value <5 ? 10?3was corrected to be an equiva-
lent of whole-brain corrected p value <.023.
Given the importance of the ACC in emotion and at-
tention, we were also interested in activation in this re-
gion, particularly the subgenual ACC. As no cluster in
this region met the aforementioned threshold criteria,
we used the same Monte Carlo procedure to determine
cluster size threshold using the ACC as the volume of
interest rather than the whole brain. The combined
right and left ACC with the medium size of all 17 subjects
was used for this Monte Carlo simulation. It included
413 voxels at the resolution of echo-planar images,
equivalently 14,662 mm3. The ANOVA results were then
used to extract the activated voxels for all pairwise con-
dition contrasts in the ACC (voxel-based p value <5 ?
10?3and ROI corrected p value <.026). The active voxel
selection criteria required that the voxels were nearest-
neighbor and within a cluster size of 106 mm3. Based
on application of these criteria to the whole ACC (the
medium size based on the 17 subjects), the voxel-based
p value <5 ? 10?3was corrected to be an equivalent of
ROI corrected p value <.026.
Functional Connectivity Analysis
To assess increases in coupling between the amygdala
and other brain regions associated with threat, psycho-
physiological interaction analyses were conducted (Friston
et al., 1997) using the amygdala cluster showing greater
activation to downward compared to upward triangles
as the seed region. A voxel-based multiple regression
model including two first-order (main effect) terms and
the interaction of the two was computed. The key term
for determining increased coupling is the second-order
interaction term involving BOLD fMRI signal extracted
from a seed region, in this case, a 10-mm spherical ROI
centered on the peak amygdala voxel for the contrast
between downward- and upward-pointing triangles. The
two first-order terms included: (1) a psychological vari-
able, in this case, a modeled epoch for the downward- or
the upward-pointing triangles convolved with a canonical
hemodynamic response; and (2) the BOLD fMRI signal
extracted from the amygdala seed region. Note that both
the physiological and psychological variables were entered
into the multiple regression analyses; in this way, results
cannot be explained by the main effects of either, but only
by the interaction. Resultant interaction beta weights
from the downward- and upward-pointing triangle voxel-
based multiple regressions were compared to one an-
other using a paired t test. The t test results were used to
extract the activated voxels (voxel-based p value <5 ?
10?3and whole-brain corrected p value <.023). The ac-
tive voxel selection criteria required that the voxels were
nearest-neighbor and within a cluster size of 248 mm3.
Based on application of these criteria to the whole brain
(medium size of the 17 brains), the voxel-based p value
<.005?3was corrected to be an equivalent of whole-brain
corrected p value <.023.
Larson et al.1527
Whole-brain Analysis: Brain Regions Activated
by Viewing of Simple Shapes
Downward- Compared to Upward-pointing Triangles
Voxelwise group analyses revealed a greater BOLD re-
sponse in the amygdala for downward- compared to
upward-pointing triangles (t = 3.63, p = .0002, [?18,
?4, ?10], Figure 2A; Table 1; all t values, coordinates,
and p values in text are reported for peak t value for
that region). As the two conditions in this contrast in-
volve presentation of the identical shape, this contrast
illustrates that not only is the amygdala activated by a
simple geometric shape but also that the orientation of
the V angle is crucial for this activation. In addition to
the amygdala, examination of other brain regions known
to react to emotionally significant and threatening stim-
uli showed that there were significant BOLD responses
for the downward minus upward triangle contrast in the
subgenual ACC (Figure 2B; t = 4.61, [?9, 27, ?7]), the
left insula (t = 3.82, [?41, ?28, 16]), and the bilateral
superior temporal gyrus [STG] (L: t = 3.80, [?41, ?30,
16]; R: t = 3.82, [56, ?57, 14]). Greater activation for
downward- compared to upward-pointing triangles was
also observed in a region of the right fusiform gyrus that
is somewhat anterior to, but, given the cluster size
and extent, likely overlapping with the typically ob-
served ‘‘fusiform face area’’ (Kanwisher, McDermott, &
Chun, 1997; t = 5.15, [39, ?41, ?16]; Figure 2C).
In addition to these regions previously implicated in
processing of affective stimuli, additional regions of the
ventral visual pathway, as well as parietal visual attention
areas, were also engaged in response to downward-
pointing triangles (see Table 1). This includes the fusi-
form region mentioned above, as well as portions of
the lingual and parahippocampal gyri in Brodmann’s
area 19, and the cuneus and the precuneus. A full list of
the regions activated in response to downward- but not
upward-pointing triangles is presented in Table 1.
No significant clusters were found indicating greater
activation for upward- compared to downward-pointing
triangles, suggesting that, in general, the brain preferen-
tially processes downward acute angles.
Downward-pointing Triangles Compared to Circles
Interestingly, none of the predominantly affect perception-
related regions listed above or in Table 1 were more
strongly activated during viewing of downward-pointing
triangles than circles in the whole-brain voxelwise analy-
sis, nor was the face-sensitive region of the fusiform gyrus
(see Table 1). The only regions demonstrating greater
activation to downward-pointing triangles were areas of
the ventral visual stream including the parahippocampal
and lingual gyri, which were also identified in the con-
trast with upward-pointing triangles. However, as detailed
in Table 1, the spatial extent of these areas of activation
was, in general, much smaller for the comparison with
Figure 2. Illustration of
greater activation for
downward triangle minus
upward triangle contrast
in the (A) amygdala, (B)
subgenual ACC, and (C)
fusiform gyrus. A t statistic
comparing percent signal
change for each voxel was
calculated and is plotted on
the images in the top row.
Mean percent signal change
for the cluster is plotted in
the bar graphs. Clusters of
248 mm3with a voxel-based
p threshold of <.005
(a whole-brain corrected
p threshold of .023) were
significant for (A) and (C).
For the ACC, full ACC
for clustering and statistical significance were applied requiring a cluster exceeding 106 mm3with a voxel-based p threshold of <.005
(a volume-corrected p of .026). The color scale to the left of the brain images represents the t statistic. Orange and red tones indicate greater
percent signal change for downward compared to upward triangles. For panels A and C, images are presented in radiological convention,
such that the right side of the image represents the left hemisphere. Greater activation for downward triangles was observed for all three
regions. Maximum t value for each cluster and the Talairach coordinates of this value are as follows: left amygdala: t = 3.63, [?18, ?4, ?10];
right subgenual ACC: t = 4.61, [9, 29, ?6]; right fusiform gyrus: t = 5.15, [39, ?41, ?16]. In each figure, the crosshairs are centered on the
coordinates of the peak t value for the downward–upward triangle comparison. The coordinates of the fusiform gyrus activation overlap
with the boundaries of the previously defined fusiform face area. Although other regions also showed greater activation for this contrast, the
focus of the present article is on regions primarily implicated in affective processing.
1528Journal of Cognitive Neuroscience Volume 21, Number 8
Table 1. Regions Activated by Simple Geometric Shapes
Downward Triangle >
Triangle > Circle
Circle > Upward
x, y, z
x, y, z
x, y, z
Predominantly Affect and Affect-perception Related Regions
AmygdalaL 3.62 151
?18, ?4, ?10
?41, ?28, 16
9, 29, ?6
8, ?28, 36
?3, ?14, 36
30, ?8, 2
?22, ?11, 7
9, ?19, 3
Anterior cingulateR 3.59 114
Lentiform nucleusR 3.86362
Mediodorsal thalamusR 3.86572
Predominantly Sensory (Visual) Processing Regions
Fusiform gyrusR 3.81 50139, ?41, ?16
?32, ?77, ?12
6, ?73, 2
?14, ?68, ?5
56, ?57, 14
Lingual/pPHC gyrusR 4.28 2218 4.07817 21, ?65, ?2
?8, ?73, 3
3.87230712, ?90, ?14
?15, ?87, 3L 3.681749 3.811184 3.66680
?41, ?30, 16
?41, ?61, 14
?62, ?25, 7
?54, ?31, 4Middle temporal gyrus L 3.63 4563.50 1167
Middle occipital gyrusR 3.80 77332, ?83, 20
CuneusR4.1514566, ?75, 6
?7, ?80, 7
8, ?62, 43
3.721102 3, ?75, 11
?12, ?97, 3L3.865323.60 261
?12, ?56, 46
13, ?87, 3
?20, ?91, 1
Precentral gyrusR3.78629 42, ?16, 35
?43, ?17, 34L3.86 838
?35, ?59, 52
Inferior frontal gyrusL3.86 633
?43, 11, 29
The mean t value for all clusters exceeding threshold for cluster size (248 mm3) and p-threshold (voxel-based p < .005 and whole-brain corrected
p < .023) are presented. For the ACC cluster, the following thresholds were a cluster size of 106 mm3with a voxel-based p < .005 and ROI corrected
p < .026. pPHC = posterior parahippocampal gyrus.
Larson et al.1529
circles. As with the downward versus upward triangle
comparison, no brain regions showed greater activation
for circles compared to downward-pointing triangles.
Circles Compared to Upward-pointing Triangles
Although not central to the understanding of the neu-
ral circuitry underlying perception of the downward
V-shape, some interesting findings arose in the circle
versus upward triangle contrast that are consistent with
earlier work (Bar & Neta, 2006; Aronoff et al., 1992) that
suggests that circles and curvilinear forms may also be
salient visual cues. Whereas upward triangles did not
elicit greater activation than circles in any region of the
brain, a large number of visual association, visual at-
tention, and other areas were activated by circles (see
Table 1). This suggests that circles are a more potent,
salient visual stimulus than upward-pointing triangles, a
notable finding given that this is the same triangle which
robustly recruits a broad range of affect and visual pro-
cessing areas, but is simply inverted.
Functional Connectivity Analysis: Affective–visual
Perception Networks Recruited by Viewing of
In light of the existence of widespread cortical projec-
tions from the amygdala (Amaral & Price, 2002), and
previous work suggesting that the amygdala modulates
activity in the ventral visual pathway, including the fusi-
form gyrus (Vuilleumier et al., 2004; Anderson & Phelps,
2001; Morris et al., 1998), we examined the degree
to which amygdala activation in response to the down-
ward V-shape was positively coupled with activation in
the fusiform gyrus. To this end, we tested for condition-
dependent changes in connectivity (i.e., psychophysio-
logical interaction analysis; Friston et al., 1997). Consis-
tent with previous data indicating amygdalar modulation
of fusiform activation, results revealed that this region
exhibited increased coupling with the amygdala during
the presentation of downward- compared to upward-
pointing triangles (Figure 3; t = 4.42, [38, ?38, ?16]).
Although the data are consistent with a modulatory role
for the amygdala, this kind of analysis cannot provide
conclusive evidence on the issue of directionality. Posi-
tive coupling with the amygdala BOLD response was
also evident in several other regions that showed acti-
vation to the downward-pointing triangles, including the
subgenual cingulate cortex (see Figure 3), lingual gyrus,
and STG. Additional regions demonstrating positive cou-
pling with the amygdala cluster based on the downward
compared to upward triangle contrast are presented in
Table 2. In general, the connectivity analyses replicate
the findings of the initial whole-brain ANOVA and lay
the foundation for future work exploring circuitry re-
cruited by the downward V-shape.
Amygdala Activation to the
These findings support our contention that signals of
danger can be represented by a fundamental visual stim-
ulus, specifically a downward-pointing acute angle, thus
facilitating efficient detection of threat. The present data
are further consistent with arguments made by numer-
ous researchers that evolution has selected for a neural
mechanism that permits the rapid detection of threat
based on simple stimulus features, and that the amyg-
Figure 3. Illustration of
greater coupling of amygdala
activation with subgenual
ACC (A) and fusiform gyrus
(B) activity in response
to downward triangles
compared to upward triangles.
A t statistic comparing coupling
coefficients for each voxel
was calculated. Clusters of
248 mm3with a voxel-based
p threshold of .005 (a
p threshold of .023) were
significant. Orange and
red tones indicate greater
coupling for downward compared to upward triangles. Images are presented in radiological convention, such that the right side of the image
represents the left hemisphere. Greater coupling with amygdala activation was observed for downward triangles compared to upward triangles
in both regions. Maximum t value for each cluster and the Talairach coordinates of this value are as follows: subgenual ACC: t = 4.08, [12,
31, ?1]; right fusiform gyrus: t = 4.42, [38, ?38, ?16]. In each figure, the crosshairs are centered on the coordinates of the peak t value
for the downward–upward triangle coupling comparison. Although other regions also showed differential coupling with the amygdala, the
focus here is on cortical regions identified in the main analyses to respond to downward triangles.
1530Journal of Cognitive Neuroscience Volume 21, Number 8
dala is a key component of this system (O ¨hman &
Mineka, 2001; LeDoux, 2000). Although previous fear
conditioning studies in rats have found that an intact
amygdala is sufficient to condition animals to very sim-
ple stimuli, such as pure tones, even following lesions
to the relevant sensory cortex (Doron & LeDoux, 1999),
studies in humans have typically assessed the role of
the amygdala in responses to contextually laden affec-
tive stimuli. However, recent data from multiple sources
now suggest that, across multiple sensory modalities,
the amygdala is reactive to very simple cues of threat
or anger (Whalen et al., 2004; Vuilleumier et al., 2003)
including those that have been stripped of associated
contextual cues, such as anger prosody (Sander et al.,
2005). The present data suggest that the visual stimuli
necessary for this recruitment may be even more ele-
mental than previously thought. Previous work demon-
strating that schematic faces that include eyebrows in
a downward V-shape (Wright, Martis, Shin, Fischer, &
Rauch, 2002) and sharp, angular objects (Bar & Neta,
2007) recruit the amygdala have provided hints that the
geometric configuration of a stimulus may be linked
with recognition of threat. The present findings extend
this work in two important ways. First, we show that not
only are geometric properties capable of engaging the
amygdala but that this process can be reduced even
further to a very simple shape. Second, no region in the
brain showed greater activation for upward-pointing
V-shapes than either the downward-pointing V or the
circles. Thus, the present data clearly indicate that not
any angular shape is sufficient to convey threat or en-
gage the amygdala and associated brain regions known
to underlie negative affect and detection of threat, but
that the orientation of the angle is critical.
Although amygdala activation was not observed when
contrasting downward-pointing triangles with circles, this
lack of differential activation is consistent with behavioral
data indicating that circles and curvilinear forms convey
the emotion of happiness (Larson et al., 2007; Zebrowitz,
1998; Aronoff et al., 1988) and functional neuroimaging
data implicating the amygdala in responses to positively
valenced affective stimuli and situations (Britton et al.,
2006; Fitzgerald, Angstadt, Jelsone, Nathan, & Phan, 2006;
Ernst et al., 2005).
As reported in the Introduction, there is growing evi-
dence from subjective ratings (Larson et al., 2007; Bar
& Neta, 2006; Aronoff et al., 1988, 1992), attentional
bias (Larson et al., 2007), and neuroimaging (Bar & Neta,
2007) work that cumulatively add to the body of evi-
dence supporting the validity of the downward V-shape
as a depiction of threat. Indeed, the regions activated by
the downward-pointing V-shape show a striking simi-
larity to those recruited by angular neutral and abstract
objects in the study by Bar and Neta (2007). In light
of this work and the extensive literature implicating
the amygdala in threat-related processes, we have inter-
preted amygdala activation to downward-pointing trian-
gles as additional evidence supporting the hypothesis
that this shape is an effective threat cue.
However, several additional factors need to be con-
sidered when interpreting these results. As the amygdala
has also been found to be activated by a number of
other tasks seemingly unrelated to threat or emotion,
such as neutral faces (Fitzgerald et al., 2006), gaze moni-
toring (Hooker et al., 2003), unpredictable innocuous
tones (Herry et al., 2007), and more rapid responses
in a working memory task (Schaefer et al., 2006), the
mere presence of amygdala activation to the downward-
pointing triangles does not, in itself, provide incontro-
vertible evidence that the downward-pointing V-shape
signals threat. Factors such as different patterns of eye
movements or attentional orientation may be evoked by
Table 2. Regions Correlated with Activated Cluster in the Left Amygdala
Region Mean t StatisticVolume (mm3) Coordinates x, y, z
Anterior cingulateR 4.08 724 12, 31, ?1
38, ?38, ?16
16, 17, 5
Fusiform gyrusR 4.42424
Caudate/putamenR 8.61 2725
CaudateL 3.75 356
?18, 14, 2
50, 37, 15Prefrontal cortexR 4.51 441
Temporal operculumR 6.742718 60, ?5, 5
45, ?27, 7
?64, ?52, 8
6, ?86, 3
?14, ?89, ?7
Superior temporal gyrusR4.51 894
Middle temporal gyrusL 4.81894
Lingual gyrusR4.87 265
Lingual gyrusL 4.53 315
All regions from the downward-pointing compared to upward-pointing triangle contrast showing positive coupling with the amygdala cluster
demonstrating greater BOLD responses for downward- compared to upward-pointing triangles. Clusters were designated as significant if they
exceeded the cluster threshold of 248 mm3and the p-threshold criteria (voxel-based p < .005 and whole-brain corrected p < .023).
Larson et al.1531
these visual forms, and the contribution of such factors
to the recognition process should be addressed with
future work. Additionally, greater activation to the
downward- compared to upward-pointing triangle may
be due to the fact that the upward-pointing triangle is
a more ‘‘standard’’ (or commonly seen) version of the
shape, and thus, is less novel. Although this is certainly
possible, we have previously found that search times
are faster for a downward-pointing V (literally the letter
‘‘V’’) than the same shape inverted. As ‘‘V’’ is a letter,
and thus, the more standard variant of the shape, these
findings are not consistent with the novelty interpreta-
tion. Finally, it is also possible that the numerosity task
used to maintain the participant’s attention may be a
factor; however, the fact that performance was at ceiling
and did not differ as a function of shape suggests that
the observed BOLD differences were likely not due to
Activation to the Downward-pointing V-shape
in Other Regions
As predicted, we observed increased activation of the
ventral visual pathway and other visual cortical regions,
including the fusiform gyrus, during viewing of the down-
ward V-shape. The lateral portion of the mid-fusiform
gyrus responds strongly to faces (Kanwisher et al., 1997)
and its activation is heightened by affective expressions
(Kesler-West et al., 2001). Importantly, previous studies
using fearful faces as affective stimuli have demonstrated
that the amygdala and the fusiform gyrus form a func-
tional network for processing these stimuli, as evidenced
by increased coupling of these two regions during view-
ing of fearful faces (Morris et al., 1998) and the fact that
an intact amygdala potentiates responses in the fusiform
cortex (Vuilleumier & Pourtois, 2007). The present data
indicate that, much like the response to emotional faces,
not only does viewing of the downward-pointing V-shape
activate the amygdala and a region of the fusiform gyrus
that overlaps with the face processing fusiform region but
it also enhances connectivity between these two regions.
These data further highlight that responses to this sim-
ple shape are similar to those recruited by contextually
rich affective stimuli, such as affective faces.
With respect to visual cortical regions more generally,
as would be predicted by Vuilleumier’s model of affect
modulation of visual processing (Vuilleumier & Driver,
2007), we found widespread increases in activation in
regions of the ventral visual pathway and visual atten-
tion areas for the downward-pointing triangles com-
pared to the other shapes. Interestingly, we also found
that circles were more potent activators of these re-
gions than the upward-pointing triangles, albeit to a
lesser extent than the downward-pointing triangles. On
the premise that activation in visual cortical regions is
potentiated by attention-grabbing and affective stimuli
(Vuilleumier & Driver, 2007), these data are once again
consistent with previous behavioral work demonstrat-
ing that downward-pointing V and circular forms convey
emotion (Larson et al., 2007; Bar & Neta, 2006; Aronoff
et al., 1988, 1992). The augmented activation associated
with the threat-related shape compared to the happiness-
related shape in these affect-modulated visual regions
is further in keeping with previous work identifying a
negativity bias, such that negatively valenced information
receives preferential processing (Cacioppo, 1994).
Another region with face- and affect-sensitive response
properties, the superior temporal gyrus and sulcus (Allison,
Puce, & McCarthy, 2000) (STG/STS), showed greater ac-
tivation for both of the affective shapes, the downward
triangle and the circle. Activation in the STG/STS is elic-
ited by viewing of faces, facial features (Puce, Allison,
Bentin, Gore, & McCarthy, 1998), and emotional faces
(Phillips et al., 1998), including those depicting anger
(Hooker et al., 2003). The STG has also been found to
be responsive to biologically meaningful eye gaze di-
rection (Hooker et al., 2003). In addition to the face,
the STG has also repeatedly been found to be engaged
during viewing of biological motion, both when watch-
ing real human movement (Allison et al., 2000) and
when viewing point-light representations of movement
(Bonda, Petrides, Ostry, & Evans, 1996). Importantly for
the present work, this STG activation to biological mo-
tion has been found to depend upon processing of the
configural aspects of the stimulus (Thompson, Clarke,
Stewart, & Puce, 2005). Relatedly, Downing, Jiang, Shuman,
and Kanwisher (2001) identified a region of the STG that
responds preferentially to images of the human body,
including simple graphic forms, such as stick figures and
silhouettes. Notably, previous work examining the con-
figural bases of emotion found that angular patterns of
body movement and position are perceived as threaten-
ing (Aronoff et al., 1992). Further reinforcing the notion
that the STG responds to simple and fundamental bio-
logically relevant cues, voice prosody during angry non-
sense speech also activates this region (Grandjean et al.,
2005). Thus, the STG appears to be broadly activated
by biological signals from the face and the body (Allison
et al., 2000), and is responsive to cues representing the
configural forms underlying these signals. The fact that
the STG is robustly activated by affective and biologically
relevant stimuli has led some researchers to suggest that
the STG is an important node in the neural circuitry un-
derlying social cognition (Adolphs, 2002). With respect
to our findings, activation of this region in response
to configural shapes associated with emotion further
underscores their potential role as fundamental cues of
biologically relevant information.
In addition to the amygdala, several other regions re-
peatedly implicated in emotion-related processes were
activated by downward-pointing V’s, but not to either
of the other shapes. Activation of the subgenual ACC
in response to the downward-pointing V-shape is con-
sistent with previous work indicating that this shape
1532Journal of Cognitive NeuroscienceVolume 21, Number 8
carries a connotation of negative affect. The ACC has
been implicated in a wide range of affective and cogni-
tive functions, with the ventral portion, including the
subgenual ACC, predominantly demonstrating sensitivity
to affect-related processes (Bush, Luu, & Posner, 2000).
Negative affect (NA), in particular, has been linked with
activation of this region, both in terms of induced nega-
tive mood (Liotti et al., 2000) and more trait-like cases of
NA, including self-reported high-trait NA (Zald, Mattson,
& Pardo, 2002) and major depressive disorder (Drevets
et al., 1997). In addition to increased activation, we
also found positive coupling of activation in the amyg-
dala and the ventral ACC during viewing of downward-
compared to upward-pointing triangles, again implicating
an amygdalar–cortical network in perception of this
shape. Although initial work on the role of the ventral
ACC in emotion focused more on the pathogenesis
of mood pathology (Drevets et al., 1997), these data
also underscore the role of this region in processing of
stimulus-driven negative affect, including simple context-
free representations of threat.
The insular cortex has been implicated in a broad
range of functions including visceral responses, respond-
ing to other’s facial expressions (Carr, Iacoboni, Dubeau,
Mazziotta, & Lenzi, 2003), pain (Sawamoto et al., 2000),
and a number of unpleasant emotions (Phillips et al.,
1998). Interestingly, another pain-responsive region in
the cingulate cortex (Sawamoto et al., 2000; Rainville,
Duncan, Price, Carrier, & Bushnell, 1997) also showed
increased BOLD response to the downward V-shape. Al-
though we have not conceptualized the downward V
form as being explicitly related to pain, the experience
and expectancy of pain is clearly amplified by unpleas-
ant emotional states and threat stimuli (Rainville et al.,
1997), and thus, may be part of the larger affective pro-
cessing network that instantiates recognition of and
preparation for responses to potential threat.
The present work certainly does not deny the important
effects of contextual cues in eliciting affect and percep-
tion of threat; nor does it suggest that other attributes
of a particular stimulus, such as size (Tipples, 2007),
are unimportant. Furthermore, this work is not meant
to suggest that all visual depictions of threat are based
on the V-shape. Rather, these data highlight the fact
that it is possible for visual depictions of threat to be
reduced to a simple geometric configuration. In addi-
tion, although the spatial resolution of the imaging data
does not permit conclusions at the cellular level, given
the potential biological relevance of this shape, it is in-
triguing to speculate that there may be specific cells (or
sets of cells) which are tuned to respond primarily to the
downward acute angle, just as there are cells known
to respond to diagonal lines and other geometric forms
(Livingstone & Hubel, 1988).
In sum, consistent with the notion that simple, repeat-
able patterns aid in efficient detection of threat (Herry
et al., 2007), the current data provide additional sup-
port for the idea that visual threat can be conveyed by a
simple downward V-shape, even when completely de-
void of other contextual or affective cues. Recognition
of this shape is instantiated in a network of regions pre-
viously implicated in the processing of contextually rich
affective and biologically relevant stimuli, such as threat-
related faces, including the amygdala, the subgenual
ACC, the STG/STS, and the fusiform gyrus. This pattern
of activation supports the findings from our previous
behavioral studies indicating that the downward V-shape
operates as a signal of threat, both at a subjective and
attentional level. Recognition of this configural repre-
sentation is likely facilitated by the amygdala and the
STG/STS, with the amygdala then eliciting heightened
processing in visual sensory regions. Importantly, these
results suggest that this neural circuitry may achieve ef-
ficient recognition of threat through detection of the
underlying geometry of the stimulus. Thus, the essence
of visual threat can be signaled and detected based on far
less stimulus information than previously demonstrated
to be necessary.
We thank Jordan Robinson, Jeremy Grounds, and Jeff Stearns,
and the Michigan State University Radiology Department for
their efforts and support on this project. We would also like to
thank Jim Zacks, Rose Zacks, and Tom Carr for many helpful
comments. This work was supported by funds from the MSU
Intramural Research Grant Program and NIMH (MH071275-01)
to Christine L. Larson and the Michigan State University Core
Reprint requests should be sent to Christine L. Larson, Depart-
ment of Psychology, University of Wisconsin-Milwaukee, 2441
E. Hartford Avenue, Milwaukee, WI 53211, or via e-mail: larsoncl@
Numerosity (1, 2, 3, 4, 5, 7) repeated measures ANOVA, the
main effect for shape was not significant (F < 1; means:
downward triangle = 651.88 msec, upward triangle = 653.41
msec, circle = 649.72 msec). There was a significant main
effect for numerosity [F(5, 95) = 19.07, p < .001]. Subse-
quent analyses revealed that participants took longer to de-
termine that fewer than four shapes were present (mean of
1–3 shapes = 661.18) than greater than four [mean of 5–
7 shapes = 631.27; t(19) = 4.50, p < .001]. This finding was
primarily driven by the fact that the two conditions closest
to the choice point (4 shapes) were the most difficult distinc-
tions to make. In particular, responses to the condition in
which three shapes were presented were slower than all other
numerosity conditions (ps < .05), including the five-shape
2. A Shape ? Numerosity ANOVA, in which numerosity was
composed of two factors (1–3 or 5–7 shapes), was calculated
In a Shape (downward triangle, upward triangle, circle) ?
Larson et al.1533
on a voxelwise basis. Importantly, no Shape ? Numerosity
interaction was present for any of the affect-related regions
reported in the main analysis. Three regions did show signifi-
cant interactions surviving cluster threshold. Both the left lin-
gual gyrus (?10, ?89, 11; 2683 voxels) and the left posterior
cingulate/precuneus (?19, ?15, 14; 703 voxels) showed greater
activation for more compared to fewer shapes for upward-
pointing triangles. Also, the right cuneus (17, ?72, 9, 944 voxels)
showed an advantage for more compared to fewer shapes for
the downward-pointing triangle.
Adolphs, R. (2002). Neural systems for recognizing emotion.
Current Opinion in Neurobiology, 12, 169–177.
Allison, T., Puce, A., & McCarthy, G. (2000). Social
perception from visual cues: Role of the STS region.
Trends in Cognitive Sciences, 4, 267–2778.
Amaral, D. G., & Price, J. L. (2002). Amygdalo-cortical
projections in the monkey (Macaca fascicularis).
Journal of Comparative Neurology, 230, 465–496.
Anderson, A. K., & Phelps, E. A. (2001). Lesions of the
human amygdala impair enhanced perception of
emotionally salient events. Nature, 411, 305–309.
Aronoff, J. (2006). How we recognize angry and happy
emotion in people, places, and things. Cross-Cultural
Research, 40, 83–105.
Aronoff, J., Barclay, A. M., & Stevenson, L. A. (1988). The
recognition of threatening facial stimuli. Journal of
Personality and Social Psychology, 54, 647–655.
Aronoff, J., Woike, B. A., & Hyman, L. M. (1992). Which
are the stimuli in facial displays of anger and happiness?
Configural bases of emotion recognition. Journal of
Personality and Social Psychology, 62, 1050–1066.
Bar, M., & Neta, M. (2006). Humans prefer curved visual
objects. Psychological Science, 17, 645–648.
Bar, M., & Neta, M. (2007). Visual elements of subjective
preference modulate amygdala activation.
Neuropsychologia, 45, 2191–2200.
Bassili, J. N. (1978). Facial motion in the perception of faces
and of emotional expression. Journal of Experimental
Psychology: Human Perception and Performance, 4,
Bonda, E., Petrides, M., Ostry, D., & Evans, A. (1996).
Specific involvement of human parietal systems and
the amygdala in the perception of biological motion.
Journal of Neuroscience, 16, 3737–3744.
Britton, J. C., Phan, K. L., Taylor, S. F., Welsh, R. C., Berridge,
K. C., & Liberzon, I. (2006). Neural correlates of social
and nonsocial emotions: An fMRI study. Neuroimage,
Bush, G., Luu, P., & Posner, M. I. (2000). Cognitive and
emotional influences in anterior cingulate cortex. Trends
in Cognitive Sciences, 4, 215–222.
Cacioppo, J. T. (1994). Relationship between attitudes and
evaluative space: A critical review, with emphasis on
the separability of positive and negative substrates.
Psychological Bulletin, 115, 401–423.
Carr, L., Iacoboni, M., Dubeau, M. C., Mazziotta, J. C., &
Lenzi, G. L. (2003). Neural mechanisms of empathy in
humans: A relay from neural systems for imitation to
limbic areas. Proceedings of the National Academy of
Sciences, U.S.A., 10, 5497–5502.
Cox, R. W. (1996). AFNI: Software for analysis and visualization
of functional magnetic resonance neuroimages. Computers
and Biomedical Research, 29, 162–173.
Darwin, C. (1872/1998). The expression of the emotions
in man and animals (3rd ed.). New York: Oxford
Doron, N. N., & LeDoux, J. E. (1999). Organization of
projections to the lateral amygdala from auditory and
visual areas of the thalamus in the rat. Journal of
Comparative Neurology, 412, 383–409.
Downing, P. E., Jiang, Y., Shuman, M., & Kanwisher, N.
(2001). A cortical area selective for visual processing
of the human body. Science, 293, 2470–2473.
Drevets, W. C., Price, J. L., Simpson, J. R., Jr., Todd, R. D.,
Reich, T., Vannier, M., et al. (1997). Subgenual prefrontal
cortex abnormalities in mood disorders. Nature, 386,
Eastwood, J. D., Smilek, D., & Merikle, P. M. (2003).
Negative facial expression captures attention and disrupts
performance. Perception & Psychophysics, 65, 352–358.
Eibl-Eibesfeldt, I. (1989). Human ethology. New York:
Aldine de Gruyter.
Ekman, P. (1973). Universal facial expressions in emotion.
Studia Psychologica, 15, 140–147.
Ekman, P. (2003). Emotions revealed: Recognizing faces
and feelings to improve communication and emotional
life. New York: Times Books.
Ekman, P., Friesen, W., & Hager, J. C. (2002). Facial action
coding system. Salt Lake City: Research Nexus.
Ernst, M., Nelson, E. E., Jazbec, S., McClure, E. B., Monk,
C. S., Leibenluft, E., et al. (2005). Amygdala and nucleus
accumbens in responses to receipt and omission of gains
in adults and adolescents. Neuroimage, 25, 1279–1291.
Fitzgerald, D. A., Angstadt, M., Jelsone, L. M., Nathan, P. J.,
& Phan, K. L. (2006). Beyond threat: Amygdala reactivity
across multiple expressions of facial affect. Neuroimage,
Friston, K. J., Buechel, C., Fink, G. R., Morris, J., Rolls, E., &
Dolan, R. J. (1997). Psychophysiological and modulatory
interactions in neuroimaging. Neuroimage, 6, 218–229.
Grandjean, D., Sander, D., Pourtois, G., Schwartz, S., Seghier,
M. L., Scherer, K. R., et al. (2005). The voices of wrath:
Brain responses to angry prosody in meaningless speech.
Nature Neuroscience, 8, 145–146.
Hansen, C. H., & Hansen, R. D. (1988). Finding the face
in the crowd: An anger superiority effect. Journal of
Personality and Social Psychology, 54, 917–924.
Herry, C., Bach, D. R., Esposito, F., Di Salle, F., Perrig,
W. J., Scheffler, K., et al. (2007). Processing of temporal
unpredictability in human and animal amygdala. Journal
of Neuroscience, 27, 5958–5966.
Hildebrandt, K. A., & Fitzgerald, H. E. (1983). The infant’s
physical attractiveness: Its effect on bonding and
attachment. Infant Mental Health Journal, 4, 3–12.
Hooker, C. I., Paller, K. A., Gitelman, D. R., Parrish, T. B.,
Mesulam, M. M., & Reber, P. J. (2003). Brain networks
for analyzing eye gaze. Cognitive Brain Research, 17,
Horstmann, G., & Bauland, A. (2006). Search asymmetries
with real faces: Testing the anger–superiority effect.
Emotion, 6, 193–207.
Irwin, W., Davidson, R. J., Lowe, M. J., Mock, B. J., Sorenson,
J. A., & Turski, P. A. (1996). Human amygdala activation
detected with echo-planar functional magnetic resonance
imaging. NeuroReport, 7, 1765–1769.
Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The
fusiform face area: A module in human extrastriate cortex
specialized for face perception. Journal of Neuroscience,
Kesler-West, M. L., Andersen, A. H., Smith, C. D., Avison,
M. J., Davis, C. E., Kryscio, R. J., et al. (2001). Neural
1534Journal of Cognitive Neuroscience Volume 21, Number 8
substrates of facial emotion processing using fMRI.
Cognitive Brain Research, 11, 213–226.
Kim, D., Adalsteinsson, E., Glover, G., & Spielman, D.
(2002). Regularized higher-order in vivo shimming.
Magnetic Resonance in Medicine, 48, 715–722.
Larson, C. L., Aronoff, J., & Stearns, J. (2007). The shape
of threat: Simple geometric forms evoke rapid and
sustained capture of attention. Emotion, 7, 526–534.
LeDoux, J. E. (2000). Emotion circuits in the brain. Annual
Review of Neuroscience, 23, 155–184.
Liotti, M., Mayberg, H. S., Brannan, S. K., McGinnis, S., Jerabek,
P., & Fox, P. T. (2000). Differential limbic–cortical correlates
of sadness and anxiety in healthy subjects: Implications
for affective disorders. Biological Psychiatry, 48, 30–42.
Livingstone, M., & Hubel, D. H. (1988). Segregation of form,
color, movement and depth: Anatomy, physiology, and
perception. Science, 240, 740–749.
Lundqvist, D., Esteves, F., & O ¨hman, A. (1999). The face
of wrath: Critical features for conveying facial threat.
Cognition and Emotion, 13, 691–711.
Lundqvist, D., Esteves, F., & O ¨hman, A. (2004). The face of
wrath: The role of features and configurations in conveying
social threat. Cognition and Emotion, 18, 161–182.
Lundqvist, D., & O ¨hman, A. (2005). Emotion regulates attention:
The relation between facial configurations, facial emotion,
and visual attention. Visual Cognition, 12, 51–84.
Morris, J. S., Friston, K. J., Buchel, C., Frith, C. D., Young,
A. W., Calder, A. J., et al. (1998). A neuromodulatory role
for the human amygdala in processing emotional facial
expressions. Brain, 121, 47–57.
Niedenthal, P. M., & Kitayama, S. (1994). The heart’s eye:
Emotional influences in perception and emotion. San
Diego, CA: Academic Press.
O ¨hman, A., Lundqvist, D., & Esteves, F. (2001). The face in
the crowd revisited: A threat advantage with schematic
stimuli. Journal of Personality and Social Psychology,
O ¨hman, A. E., & Mineka, S. (2001). Fears, phobias, and
preparedness: Toward an evolved module of fear and
fear learning. Psychological Science, 108, 438–522.
Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957).
The measurement of meaning. Urbana, IL: University
of Illinois Press.
Parrish, T. B., Gitelman, D. R., LaBar, K. S., & Mesulam, M.
(2000). Impact of signal-to-noise on functional MRI.
Magnetic Resonance in Medicine, 44, 925–932.
Phillips, M. L., Young, A. W., Scott, S. K., Calder, A. J.,
Andrew, C., Giampietro, V., et al. (1998). Neural responses
to facial and vocal expressions of fear and disgust.
Proceedings of the Royal Society of London, Series B,
Biological Sciences, 265, 1809–1817.
Puce, A., Allison, T., Bentin, S., Gore, J. C., & McCarthy, G.
(1998). Temporal cortex activation in humans viewing eye
and mouth movements. Journal of Neuroscience, 18,
Rainville, P., Duncan, G. H., Price, D. D., Carrier, B., &
Bushnell, M. C. (1997). Pain affect encoded in human
anterior cingulate but not somatosensory cortex. Science,
Sabatinelli, D., Lang, P. J., Keil, A., & Bradley, M. M. (2007).
Emotional perception: Correlation of functional MRI and
event-related potentials. Cerebral Cortex, 17, 1085–1091.
Sander, D., Grandjean, D., Pourtois, G., Schwartz, S., Seghier,
M. L., Scherer, K. R., et al. (2005). Emotion and attention
interactions in social cognition: Brain regions involved
in processing anger prosody. Neuroimage, 28, 848–858.
Sawamoto, N., Honda, M., Okada, T., Hanakawa, T., Kanda, M.,
Fukuyama, H., et al. (2000). Expectation of pain enhances
responses to nonpainful somatosensory stimulation in the
anterior cingulate cortex and parietal operculum/posterior
insula: An event-related functional magnetic resonance
imaging study. Journal of Neuroscience, 20, 7438–7445.
Schaefer, A., Braver, T. S., Reynolds, R., Jr., Burgess, G. C.,
Yarkoni, T., & Gray, J. R. (2006). Individual differences in
amygdala activity predict response speed during working
memory. Journal of Neuroscience, 26, 10120–10128.
Talairach, J., & Tournoux, P. (1988). Co-planar stereotaxic
atlas of the human brain: 3-Dimensional proportional
system: An approach to cerebral imaging (1st ed.).
Thompson, J. C., Clarke, M., Stewart, T., & Puce, A. (2005).
Configural processing of biological motion in human
superior temporal sulcus. Journal of Neuroscience, 25,
Tipples, J. (2007). Wide eyes and an open mouth enhance
facial threat. Cognition and Emotion, 21, 535–557.
Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2001).
Effects of attention and emotion on face processing in
the human brain: An event-related fMRI study. Neuron,
Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2003).
Distinct spatial frequency sensitivities for processing
faces and emotional expressions. Nature Neuroscience,
Vuilleumier, P., & Driver, J. (2007). Modulation of visual
processing by attention and emotion: Windows on causal
interactions between human brain regions. Philosophical
Transactions of the Royal Society of London, Series B,
Biological Sciences, 362, 837–855.
Vuilleumier, P., & Pourtois, G. (2007). Distributed and
interactive brain mechanisms during emotion face
perception: Evidence from functional neuroimaging.
Neuropsychologia, 45, 174–194.
Vuilleumier, P., Richardson, M. P., Armony, J. L., Driver, J.,
& Dolan, R. J. (2004). Distant influences of amygdala
lesion on visual cortical activation during emotional
face processing. Nature Neuroscience, 7, 1271–1278.
Ward, B. (2000). Simultaneous inference for fMRI data.
Milwaukee, WI: Biophysics Research Institute, Medical
College of Wisconsin.
Ward, B. (2002). Deconvolution analysis of fMRI time
series data. Milwaukee, WI: Biophysics Research Institute,
Medical College of Wisconsin.
Whalen, P. J., Kagan, J., Cook, R. G., Davis, F. C., Kim, H.,
Polis, S., et al. (2004). Human amygdala responsivity to
masked fearful eye whites. Science, 306, 2061.
White, M. (1995). Preattentive analysis of facial expressions
of emotion. Cognition and Emotion, 9, 439–460.
Wright, C. I., Martis, B., Shin, L. M., Fischer, H., & Rauch,
S. L. (2002). Enhanced amygdala responses to emotional
versus neutral schematic facial expressions. NeuroReport,
Zald, D. H., Mattson, D. L., & Pardo, J. V. (2002). Brain
activity in ventromedial prefrontal cortex correlates with
individual differences in negative affect. Proceedings of
the National Academy of Sciences, U.S.A., 99, 2450–2454.
Zebrowitz, L. A. (1998). Reading faces: Window to the soul?
Boulder, CO: Westview Press.
Larson et al.1535