Neuropsychologia 42 (2004) 1657–1665

What are you looking at?

Impaired ‘social attention’ following frontal-lobe damage
Shaun P. Vecera a,∗ , Matthew Rizzo b
a Department of Psychology, E11 Seashore Hall, University of Iowa, Iowa City, IA 52242-1407, USA
b Department of Neurology and College of Engineering, University of Iowa, Iowa City, IA, USA

Received 2 January 2004; received in revised form 15 April 2004; accepted 21 April 2004

Abstract
Humans are able to predict the behavior of others. Several studies have investigated this capability by determining if social cues, such
as eye gaze direction, can influence the allocation of visual attention. When a viewer sees a face looking to the left, the viewer’s attention
is allocated in the gazed-at direction. These ‘social attention’ studies have asked if this allocation of attention is automatic or under
voluntary control. In this paper, we show that a patient with frontal-lobe damage is impaired at allocating attention to peripheral locations
voluntarily, although attention can be allocated there automatically. The patient, EVR, can use peripheral cues to selectively process one
location over another but cannot use symbolic cues (words) to allocate attention. EVR is also impaired in using eye gaze cues to allocate
attention, suggesting that ‘social attention’ may involve frontal-lobe processes that control voluntary, not automatic, shifts of visuospatial
attention.
© 2004 Elsevier Ltd. All rights reserved.
Keywords: Attention; Frontal lobes; Eye gaze direction; Social attention; Theory-of-mind

1. Introduction Kirkpatrick-Steger, Thomas, Allen, & Blumberg, 1995;

Reaux, Theall, & Povinelli, 1999). Second, gaze perception
Humans appear to be endowed with the ability to make appears to be supported by gaze-selective neural responses
predictions regarding others’ behavior or intentions. For ex- in a “face-cell” area in the superior temporal sulcus of the
ample, a professor may predict that an approaching student macaque (Perrett, Hietanen, Oram, & Benson, 1992), an
is going to ask for an extension on an assignment if the area which may correspond to the “fusiform face area” of
assignment is due in 2 days. Some theories label this abil- human extrastriate cortex (Kanwisher, McDermott, & Chun,
ity as a “mind reading” ability or a “theory-of-mind” (e.g., 1997). Third, sensitivity to gaze direction emerges in early
Baron-Cohen, 1995; Leslie, 1991; Premack & Woodruff, life (see Vecera & Johnson, 1995 for relevant results), sug-
1978), and some theories suggest that such an ability allows gesting a role in developing human social skills. Given
humans to understand the social interactions that are im- these reasons, theorists such as Baron-Cohen (1995) hy-
portant in the elaborate social hierarchies of primates (e.g., pothesize distinct cognitive modules for detecting eyes and
Cheney & Seyfarth, 1990). perceiving another’s eye gaze direction. Initial research
Perception of another person’s eye gaze direction supporting these theories and processes came from special
plays a key role in many theory-of-mind theories (e.g., populations (e.g., children with autism). More recent studies
Baron-Cohen, 1995) for several reasons. First, gaze direc- of cognitively normal individuals support theory-of-mind
tion may signal the upcoming target or goal of another theories by showing that visuospatial attention is influenced
person’s behavior, which helps predict behavior; for exam- by another’s gaze direction (Driver et al., 1999; Friesen &
ple, chimpanzees appear to be able to use another’s line of Kingstone, 1998; Kingstone, Friesen, & Gazzaniga, 2000;
gaze to orient attention (Povinelli & Eddy, 1996; Povinelli, Langton & Bruce, 1999, 2000).
Nelson, & Boysen, 1990), although such abilities may Previous research on visuospatial attention has demon-
not require an attribution of mental states (see Gagliardi, strated that attention can be allocated to locations by dif-
ferent types of cues that appear before a target stimulus
∗ Corresponding author. Tel.: +1 319 335 0839; fax: +1 319 335 0191. appears (Posner, 1980; Posner, Snyder, & Davidson, 1980).
E-mail address: [email protected] (S.P. Vecera). In Posner’s now-classic task, participants are asked to

0028-3932/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.neuropsychologia.2004.04.009
1658 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665

detect a visual target which appears at a peripheral location. tional allocation. Purely reflexive orienting to gaze would
Prior to target onset, a predictive cue appears. “Valid” cues not allow different social contexts to influence attentional
correctly predict a target’s subsequent location, whereas allocation; in contrast, voluntary orienting could allow so-
“invalid” cues are inaccurate and misleading. Participants cial context to modulate attentional orienting based on gaze
are generally faster to detect validly cued targets than in- direction. Current reflexive accounts of gaze-directed atten-
validly cued targets. Also, two cue types have been studied tion do not explain how attention can distinguish between
in this task, and these cues differ in their effects on at- appropriate and inappropriate contexts.
tentional orienting. Peripheral cues flicker briefly at the Neuropsychological data can provide additional evidence
predicted target location, whereas centrally-presented (sym- regarding the orienting produced by eye gaze cues. We focus
bolic) cues indicate a target’s probable location by means on patients with frontal-lobe lesions, who exhibit a variety
of symbolic information such as a word or arrow. Peripheral of cognitive impairments that can broadly be classified as
cues appear to capture spatial attention automatically or re- problems with cognitive control (see Kimberg, D’Esposito,
flexively (Jonides, 1981; Yantis & Hillstrom, 1994), cannot & Farah, 2000; Miller & Cohen, 2001, for recent reviews).
be ignored, and are not interfered with by symbolic cues Some of these impairments in cognitive control appear as
(Jonides, 1981; Müller & Rabbit, 1989). Peripheral cues attentional impairments. Patients with frontal-lobe lesions
summon attention even when they do not reliably predict are distracted by irrelevant stimuli (Chao & Knight, 1995),
target location; infrequently-occurring validly cued targets are impaired at voluntarily sustaining attention (Wilkins,
are detected faster than frequently-occurring invalidly cued Shallice, & McCarthy, 1987), and are impaired at using ad-
targets. In contrast, symbolic cues require participants to vance information in a variety of tasks, including spatial
shift attention voluntarily to the cued location. These sym- cuing tasks (Alivisatos, 1992; Alivisatos & Milner, 1989;
bolic cues can be ignored and are interfered with by periph- Koski, Paus, & Petrides, 1998). Because frontal-lobe patients
eral cues (Jonides, 1981; Müller & Rabbit, 1989), although do not appear to be impaired in highly-practiced, automatic
these cues need not predict an upcoming target’s location to tasks, these patients can be studied to explore dissociations
direct attention to a cued location (Hommel, Pratt, Colzato, between automatic and voluntary attentional processes.
& Godijn, 2001; Tipples, 2002). Studies investigating voluntary attention in frontal pa-
An important question is whether gaze cues orient spatial tients have used variants of Posner’s spatial cuing task with
attention reflexively, as peripheral cues, or voluntarily, as symbolic cues. For example, Alivisatos and Milner (1989)
symbolic cues. If eye gaze is critical to inferring another’s presented patients with word cues that either signaled
mental state, as suggested by theory-of-mind accounts, then the upcoming target’s location (valid trials) or provided
one might expect gaze cues to summon attention automati- no information about the target’s location (neutral trials).
cally to gazed-at locations. This prediction finds some sup- Frontal-lobe patients showed a smaller attentional benefit
port: When another person’s eyes are gazing to our left, (the difference in performance between valid and neutral
validly cued targets that appear there are processed faster trials) than either control participants or temporal lobe
and more accurately than invalidly cued targets that appear patients. Koski et al. (1998) reported similar results from
to our right (Driver et al., 1999; Friesen & Kingstone, 1998; centrally-presented arrow cues that either validly predicted
Langton & Bruce, 1999). Further, eye gaze can affect atten- the upcoming target’s location or did not predict the target’s
tional orienting when a target appears at the gazed-at (validly location. Again, frontal patients showed smaller attentional
cued) location infrequently, suggesting that gaze cues ori- benefits than both control participants and temporal lobe
ent attention in a reflexive manner. Attention is summoned patients. In the foregoing studies, the frontal patients had
to a cued (i.e., gazed at) location even when the target is varied lesion locations that included dorsolateral and ven-
more likely to appear at the uncued (i.e., not gazed at) lo- tromedial frontal areas.
cation. When the gaze cue is counter-predictive, gaze cues, To determine if eye gaze cues orient attention in an au-
like peripheral cues, cannot be ignored, and attention is tomatic or voluntary manner, we investigated attentional
driven to the gazed-at location (Driver et al., 1999; Friesen & orienting in patient EVR, who had regions of both frontal
Kingstone, 1998; Langton & Bruce, 1999). lobes excised during removal of a tumor (Eslinger &
Several issues remain to be addressed before we conclude Damasio, 1985). EVR performed a simple spatial cuing task
that gaze direction cues automatically influence the orient- in which he detected the onset of a target that appeared in
ing of spatial attention. For example, if participants in a the visual periphery (see Fig. 1). The target was preceded by
gaze precuing study are good “mind readers,” they might a spatial cue that either predicted the target’s location (valid
correctly guess the expected outcome of the experiment and cue) or did not predict the target’s location (invalid cue).
adjust their behavior accordingly (i.e., voluntarily orienting As shown in Fig. 1, we tested EVR with three types of spa-
attention in the direction of the gaze cue). Also, in many so- tial cues to assess his attentional orienting: peripheral cues,
cial situations, reflexive orienting by gaze is inappropriate or symbolic cues (e.g., words, such as “left”), and gaze cues.
unwarranted, such as when a speaker casually glances up- Previous findings from frontal-lobe patients (Alivisatos &
ward during conversation. Such casual glances differ from Milner, 1989; Koski et al., 1998) lead us to hypothesize
more meaningful glances that would be important for atten- that EVR would be unable to use symbolic word cues to
 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665 1659

+ + +
* * *
50 ms
+ + +

0 or
+ 100 ms RIGHT

time
100 ms
+ + +

(A) (B) (C)

1000 ms

Fig. 1. Sequence of events in the experiment. (A) A valid peripheral cue. (B) An invalid word cue. (C) A valid eye gaze cue (note: stimuli not drawn
to scale).

orient attention. Further, we expected to find that EVR 1998). His working memory span, speech, verbal fluency,
would have no difficulty orienting to peripheral cues be- verbal comprehension, and face perception all appear to be
cause peripheral cues orient attention automatically (e.g., normal, as does his executive function (as measured by the
Jonides, 1981; Müller & Rabbit, 1989; Yantis & Jonides, Wisconsin Card Sort task, the Stroop task, and by his perfor-
1984) and because frontal-lobe patients typically do not mance on solving the Tower of Hanoi puzzle; see Bechara,
have impairments in ‘automatic’ performance. The critical Damasio, Tranel, & Anderson, 1998).
issue concerns gaze cues. If gaze cues summon attention Despite EVR’s preserved performance on classic
automatically, then EVR should be unimpaired orienting to ‘frontal-lobe tasks’ (e.g., Wisconsin Card Sort), he demon-
gaze cues. Specifically, he should be faster to detect targets strates a long-lasting difficulty in generating learned re-
validly cued by gaze than targets invalidly cued by gaze. If sponses to complex visual stimuli (Tranel et al., 1995). For
gaze cues orient attention in a voluntary manner, however, example, EVR’s skin conductance responses do not dis-
then EVR should be impaired orienting to gaze cues (in criminate between familiar and unfamiliar faces, although
addition to being impaired orienting to symbolic cues). his perception and recognition of faces is flawless (Tranel
et al., 1995).
At the time of the current testing, EVR was 62 years old.
2. Case report His low-level visual functions were preserved; his corrected
acuity was 20/25, and he had no visual field defects (e.g.,
At age 35, EVR was diagnosed with a cerebral tumor, scotomas). EVR’s contrast sensitivity was within the normal
a large orbitofrontal meningioma. The tumor was surgi- limits, and he was able to detect coherent motion against a
cally removed, and EVR recovered. EVR’s frontal-lobe static background within normal limits. High-level visual
damage in the chronic phase of recovery corresponded functions also appeared to be intact in EVR. For example,
to regions F07, F11, and F12 in Damasio and Damasio’s face recognition, two- and three-dimensional block con-
(1989) lesion analysis schema. The removal of the tumor struction, and copying the Rey-Osterreith complex figure
and frontal-lobe tissue left EVR with lasting impairments in were all performed within the normal range. EVR showed
decision making, personality, and some forms of cognitive a subtle attentional impairment on a standardized dual at-
control. For example, EVR seems to have an impairment tention task: EVR’s performance was outside the normal
with goal-directed behaviors—when waking up, he does not range when he simultaneously performed a visual discrim-
automatically generate the goals of showering and eating as ination at fixation and a peripheral target localization task,
he used to (Eslinger & Damasio, 1985). When making deci- although his performance on the central discrimination
sions, EVR often labors over minutia and fails to appreciate task alone was within the range of normal control partic-
the global purpose (i.e., the goal) of reaching a decision. ipants. EVR’s impaired dual-task performance suggests a
Despite his noted cognitive impairments, EVR appears difficulty with high-level, “executive” attentional processes
normal on many cognitive functions. EVR’s intelligence has (e.g., Baddeley, 1996), processes that are important for exe-
remained stable and in the superior range since his case was cuting goal-directed behaviors and in generating previously
initially presented (WAIS verbal IQ of 120 in 1985; 131 in learned responses.
1660 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665

3. Method The cue-target intervals in this procedure and the 50 ms

target presentation are too brief to permit eye movements.
3.1. Participants Nevertheless, we monitored eye movements to prevent
any anticipatory eye movements to the cued location. Eye
Both EVR and ten older control participants performed movements were monitored in both EVR and control par-
a spatial cuing task (Fig. 1) in which a target appeared ticipants using an ETL-500 head mounted eye tracking
at a peripheral visual location. Prior to the target, a cue system from ISCAN Inc. (Burlington, MA). Participants
appeared. Three cues were tested in different blocks of wore a baseball cap containing a miniature scene camera
trials: peripheral cues, symbolic word cues, and eye gaze and a miniature eye camera. The eye camera monitored
cues. The control participants had a mean age of 69.3 years the pupil using corneal reflection of the participant’s left
(S.D. = 4.6 years). eye. Eye position was indicated by a crosshair on a re-
mote video screen; eye position was monitored contin-
uously throughout the experiment and was recorded on
3.2. Stimuli
videotape for post hoc analysis. Eye position was moni-
tored for each trial, and any trials that contained a visi-
Participants sat approximately 60 cm from a Macintosh
ble eye movement were excluded from further analyses.
iMac computer (15 in. monitor). Each trial began with a
This monitoring excluded less than 0.5% of the trials,
central fixation point and two 0.95◦ × 0.95◦ boxes which
and when eye movements were made, they were highly
appeared 6.1◦ of visual angle to the left and right of fix- visible saccades that ended on or near the peripheral
ation. The peripheral cues were a 1.6◦ box that appeared placeholders.
around the placeholder boxes. Peripheral cues were sin- There were three blocks of trials, one for each cue type.
gle cues that appeared around the left or right placeholder The three blocks were shown to EVR in a fixed order because
box. The symbolic cues were the words “left” and “right” of limited testing time that prevented full counterbalancing
that appeared at fixation in 36 point Helvetica font; the of block order; the fixed order was: peripheral cues, word
words ranged from 2.8 to 3.7◦ wide. The eye gaze cues cues, and gaze cues. The control participants viewed the
consisted of a schematic face that appeared at fixation three blocks in the same order as EVR to control for any
and had eyes looking left or right. The face measured order effects.
5.7◦ tall and 4.4◦ wide. Each individual eye measured To increase the likelihood of observing a cuing effect in
0.50◦ tall and 0.77◦ wide; the averted gaze directions were EVR with the word and gaze cues, valid cues appeared more
created by moving the pupils 2 mm to the left or right frequently than invalid cues for the word cues and gaze cues.
of the eye’s center. The target was a small asterisk that For these cue types, 75% of the cues were valid and 25%
measured approximately 0.40◦ tall and 0.40◦ wide. All were invalid, a ratio which should encourage participants
stimuli were drawn in black and presented on a white to orient attention in the cued direction. Further, to provide
background. the most stringent test of automatic orienting in all partic-
In a control task, EVR viewed the cues only and was ipants, the peripheral cues were uninformative (50% valid
asked to report (1) if the peripheral flash occurred to the left and 50% invalid). Based on these cue probabilities, we in-
or right, (2) if a letter string was the word “left” or “right,” creased our chances of observing a cuing effect with the
and (3) if the eyes were looking to the left or right. The word and gaze cues and decreased our chances of observ-
stimuli were identical to those used in the spatial cuing task. ing a cuing effect with peripheral cues—observations that
The control participants were not tested on these control would run counter to our hypothesized results. Thus, we
tasks. created an experimental situation that works to falsify our
hypotheses.
3.3. Procedure All participants viewed three blocks of trials, with cue
type held constant within each block. Each block consisted
In the spatial cuing task, each trial began with the fixation of 192 trials in which a target appeared and 24 “catch” trials
point and placeholder boxes visible for 1000 ms. A cue then in which a cue appeared but no target followed. Participants
appeared for 100 ms; after the cue disappeared, a target ap- were instructed to withhold their responses on these catch
peared. On half of the trials, the target appeared immediately trials. Within each block, participants were given rests after
after the cue disappeared (100 ms stimulus onset asynchrony every 54 trials (48 target trials plus 6 catch trials). All trials
(SOA)) or after a 100 ms delay (200 ms SOA). Two SOAs were presented randomly.
were used to discourage anticipation of the target’s appear- Finally, in the control tasks administered only to EVR, a
ance following the cue. The target appeared for 50 ms, and cue was presented for 100 ms. EVR verbally reported the di-
it appeared in the left placeholder box on half of the trials rection or location of the cue (left or right) after the cue dis-
and in the right placeholder box on half of the trials. Par- appeared, and the experimenter recorded his response. EVR
ticipants pressed the spacebar on a standard keyboard when performed the control tasks before performing the spatial
they detected the onset of the target. cue task to ensure that he could correctly perceive the cue.
 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665 1661

Peripheral Cue validly cued targets than invalidly cued targets, t(9) = 3.8,
Cuing Effects by Cue Type
45 Word Cue P < 0.005.
40 An inspection of individual control participant’s results
35 Gaze Cue
(Invalid RT minus Valid RT)

30 was consistent with the averaged group responses. For pe-

25
ripheral cues, all of the control participants responded faster
Cuing Effect

20
15 to validly-cued targets than to invalidly-cued targets. For
10
5 word cues, two control participants showed slightly (but
0 not significantly) faster to detect invalidly-cued targets than
-5
-10 validly-cued targets; for one participant, invalidly-cued tar-
-15
-20 gets were detected 2.2 ms faster than validly cued targets,
-25 and for the other participant, invalidly-cued targets were de-
-30
Control Participants EVR tected 3.9 ms faster than validly-cued targets. Finally, for
gaze cues, all participants detected validly-cued targets faster
Fig. 2. Cuing effects (invalid trial RTs minus valid trial RTs) from 10 than invalidly-cued targets.
neurologically normal older control participants and patient EVR. Error
bars are 95% confidence intervals on the comparison between the cuing
These results generally replicate previous studies from
effect and 0 ms; error bars that do not overlap 0 ms indicate a statistically spatial precuing tasks. All cue types showed faster detec-
significant cuing effect in which validly cued targets are detected faster tion times to validly cued targets than invalidly cued targets.
than invalidly cued targets. These cuing effects are important because they indicate that
our procedure is sensitive to such effects in older control
participants.
4. Results
4.2. EVR
4.1. Control participants
EVR performed flawlessly in the control tasks, indicating
Reaction times (RTs) over 1000 ms were excluded from he could perceive and interpret the direction of the three
the analyses as outliers, and this trimming excluded less than different cue types.
2% of the data. There was no evidence of any systematic As with control participants, RTs over 1000 ms were ex-
anticipatory responses (RTs < 150 ms). The control partic- cluded from the analyses, and this trimming excluded less
ipants made few catch trial errors (<2%). For each partic- than 1% of EVR’s data. EVR made few errors on catch trials
ipant, we computed the cuing effect across each cue type (<1%). We performed two analyses on EVR’s results to as-
(peripheral, word, or gaze), collapsed across SOA. The cu- sess his ability to orient attention from the various cue types.
ing effect was defined as the invalid cue RT minus the valid We first performed an analysis on his RTs to determine if
cue RT; positive scores indicated a cuing effect in which he exhibited cuing effects for the three cue types; that is,
the participant was faster to detect validly-cued targets than we compared RTs on validly cued trials to RTs on invalidly
invalidly-cued targets. Preliminary analyses indicated that cued trials. As one reviewer noted, however, such an anal-
SOA did not interact with any factors, so we collapsed ysis may be problematic because of serial effects in the RT
across SOA. The average cuing effects for the control par- data which violate the independence assumption of standard
ticipants appear in Fig. 2. For the sake of completeness, all statistical tests. To overcome such difficulties, we also com-
RTs for the control participants appear in Table 1. pared EVR’s mean cuing effects to the control participants’
Fig. 2 shows that the control participants exhibited sta- cuing effects to determine if EVR differed from the control
tistically significant cuing effects for each of the cue types: participants.
following peripheral cues, control participants were 33.5 ms
faster to detect validly cued targets than invalidly cued tar- 4.2.1. Cuing effect analysis: valid versus invalid trials
gets, t(9) = 6.1, P < 0.0005. Following word cues, control EVR’s cuing effects were compared with t-tests between
participants were 13.5 ms faster to detect validly cued targets valid and invalid trials for each cue type. These cuing effects
than invalidly cued targets, t(9) = 3.0, P < 0.02. Following are plotted in Fig. 2, and, for the sake of completeness, all
gaze cues, control participants were 22.0 ms faster to detect of EVR’s RTs appear in Table 2.

Table 1
Control participants’ mean RTs (in ms) for all conditions
Cue type Peripheral Word Gaze

100 ms SOA 200 ms SOA 100 ms SOA 200 ms SOA 100 ms SOA 200 ms SOA

Valid 460.1 (22.4) 435.7 (23.8) 470.6 (17.1) 418.0 (20.1) 444.1 (24.4) 402.8 (21.2)
Invalid 503.2 (26.3) 460.2 (26.0) 474.3 (23.8) 441.5 (20.1) 461.5 (25.7) 429.0 (20.7)
Note: Standard errors appear in parentheses.
1662 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665

Table 2
EVR’s mean RTs (in ms) for all conditions
Cue type Peripheral Word Gaze

100 ms SOA 200 ms SOA 100 ms SOA 200 ms SOA 100 ms SOA 200 ms SOA

Valid 436.8 (13.9) 383.2 (11.6) 469.3 (7.6) 421.1 (7.6) 467.7 (7.4) 435.6 (9.0)
Invalid 447.4 (8.4) 428.4 (11.5) 454.9 (10.5) 422.5 (17.6) 462.6 (12.6) 438.3 (15.7)
Note: Standard errors appear in parentheses.

EVR exhibited large cuing effects for peripheral cues, peripheral location. However, EVR could not reliably use
responding 28 ms faster to validly-cued targets than to centrally-presented word cues to allocate visual attention
invalidly-cued targets, t(190) = 2.4, P < 0.02. Unlike his to peripheral locations, despite preserved perception of
responses to the peripheral cues, EVR did not exhibit any these cues, replicating the results of previous studies that
cuing effects for either word cues (he was 6.5 ms slower reported frontal-lobe patients’ inability to orient attention
to detect validly-cued targets than invalidly-cued targets), from symbolic cues (Alivisatos & Milner, 1989; Koski
t(190) < 1, or gaze cues (he was 2.4 ms slower to detect et al., 1998). Most important, EVR could not reliably al-
validly-cued targets than invalidly-cued targets), t(190) < 1. locate attention using eye gaze cues, indicating that gaze
These failures to find cuing effects to word and gaze pre- cues share attentional mechanisms with known voluntary
cues occurred despite these cues being highly predictive of cues such as words. Our findings indicate that there are at
the target’s location. Had EVR responded on the basis of least two dissociable components to attentional orienting,
the cue’s frequency only, he should have been faster to de- automatic orienting from peripheral cues and voluntary
tect validly-cued targets than invalidly-cued targets. EVR’s orienting from symbolic cues. Gaze cues appear to affect
lack of a cuing effect with word and gaze cues replicates attention in a voluntary or controlled manner, not in an
previous work that demonstrates diminished attentional automatic or reflexive manner. These results are supported
orienting from symbolic cues in patients with frontal-lobe by two separate analyses—a comparison of EVR’s reaction
damage (Alivisatos & Milner, 1989; Koski et al., 1998). times and a comparison of EVR’s cuing effects to those of
age-matched control participants.
4.2.2. Comparison to control participants One important issue to consider is if EVR might exhibit
For converging evidence of EVR’s lack of a cuing effect cuing effects for the word and gaze cues if given additional
for the word and gaze cues, we compared his cuing effects to time to orient attention. That is, EVR’s impairment may be a
those from the control participants (see Fig. 2). We compared slowing of attentional deployment following word and gaze
the control participants’ cuing effect against EVR’s cuing cues (and, presumably, other symbolic cues). We have un-
effect (i.e., we used EVR’s cuing effect as the hypothesized published results that argue against a delayed orienting ac-
mean). For the peripheral cues, EVR’s cuing effect was well count. We tested EVR in a spatial cuing task that used longer
within normal limits (28 ms) and did not differ from the con- delays between the cue and target (650 ms SOA) with eye
trol participants’ cuing effect (33.5 ms), t(9) < 1. For the movement monitoring and found that he was no faster to
word and gaze cues, however, EVR’s cuing effect was sig- detect a validly-cued target than a neutrally-cued target for
nificantly smaller than the control participants’ effects. For gaze cues (451.9 ms for validly-cued targets versus 441.1 ms
the word cues, the control participants had a 13.5 ms cuing for neutrally-cued targets, P > 0.40). Moreover, although
effect and EVR had a −6.5 ms cuing effect, and these cuing EVR might have ‘sluggish’ attentional orienting, such an im-
effects were significantly different, t(9) = 4.4, P < 0.005. pairment does not affect our main point in this paper—that
The gaze cues produced similar results: control participants gaze cues orient attention in a voluntary, controlled manner,
had a cuing effect of 22.0 ms and EVR had a cuing effect similar to word cues (and dissimilar to peripheral cues). If
of −2.4 ms, and these cuing effects were significantly dif- EVR’s attentional difficulties do stem from slowed atten-
ferent, t(9) = 4.2, P < 0.005. As with the above analysis of tional orienting, then orienting is slowed for both word and
EVR’s RTs, EVR appears unable to orient attention volun- gaze cues but not for peripheral cues. Thus, a generalized
tary symbolic cues such as words and eye gaze, suggesting slowing account would not explain our results fully, and the
that gaze cues may require voluntary, not automatic, atten- conclusion remains that word cues and gaze cues behave
tional orienting. similar to each other and differently from peripheral cues.
In general, this same conclusion holds for other differences
among cue types, such as the possibility that gaze cues (and,
5. Discussion presumably, word cues) are somehow less salient than pe-
ripheral cues.
EVR demonstrated significant cuing effects to periph- There are other, alternative accounts that could explain
eral cues at short cue-target intervals, indicating that his our dissociation between automatic peripheral cues and gaze
spatial attention can be summoned ‘automatically’ to a cues. As with the ‘sluggish’ attention alternative discussed
 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665 1663

above, these other alternative accounts do not fully explain the arrowhead falls closer to the validly cued target location
the present results. For example, readers could be concerned than the invalidly cued target location. In general, although
with our use of schematic face stimuli and our procedure of our present results strongly suggest that gaze cues rely on a
presenting these faces briefly (as opposed to leaving them voluntary orienting of attention, there are many unresolved
visible, as is done in many studies that examine cuing by issues regarding the status of gaze cues, and future studies
gaze). However, schematic faces can readily direct atten- might support a more ‘automatic’ status for gaze cues.
tion in younger participants (Driver et al., 1999; Friesen & Two recent studies (Kingstone et al., 2000; Ristic, Friesen,
Kingstone, 1998; Langton & Bruce, 1999), and our con- & Kingstone, 2002) appear at odds with our strong claim
trol participants could direct attention from the gaze cue. that there should be no double dissociation between orient-
Further, although our gaze and word cues were presented ing from gaze cues and from word cues. In these studies,
briefly, our control participants could orient from these cues; split-brain patients were able to use arrow cues to orient
we also have unpublished results from younger participants attention when those cues were presented in either visual
that replicate the gaze-cuing effects using schematic stimuli field (Ristic et al., 2002), but these patients could use gaze
that are presented briefly (Vecera, in preparation). cues to orient attention when those cues were presented in
Another possible concern with the present results is that a single visual field (Kingstone et al., 2000). These two
our gaze and word cues predicted the upcoming target’s lo- studies suggest that orienting attention (presumably volun-
cation on 75% of trials. Thus, these cues may have tapped en- tarily) from arrow cues occurs in both hemispheres whereas
dogenous (voluntary), not exogenous (automatic) attentional orienting attention (presumably reflexively) from gaze cues
processes. Our failure to find gaze- and word-cuing effects occurs in a single hemisphere only. Although these results
in EVR could result if he had an impairment in allocating at- are provocative, the hemispheric differences between arrow
tention voluntarily; gaze cues could orient attention automat- cues and gaze cues could be based upon perceptual process-
ically in general, but this effect is absent in EVR because of ing differences. If face processes are more lateralized than
a damaged endogenous control system. One difficulty with processes for other stimuli such as arrows, then the non-
this interpretation is that the automatic, exogenous compo- specialized hemisphere might be ‘blind’ to face and gaze
nent of gaze cues should have elicited attentional orienting in stimuli and thereby prevent any attentional orienting. The
EVR. Other studies have demonstrated that frontal-damaged results from split-brain patients are consistent with the con-
patients show impairments in voluntarily orienting atten- clusion that the perceptual inputs to a voluntary attentional
tion, even when valid cue appear as often as neutral cues orienting system are different for gaze cues and arrow cues.
(Alivisatos & Milner, 1989; Koski et al., 1998). Finally, These studies have not, however, demonstrated differences
in our unpublished results from EVR discussed above, we in attentional orienting per se between gaze and arrow cues.
found that he exhibits intact orienting from unpredictive pe- Our results and conclusions from EVR raise a question
ripheral cues but not from unpredictive gaze or word cues, regarding studies of gaze cuing in neurologically normal
suggesting that his impairment in orienting attention volun- participants. Why do studies of neurologically normal
tarily is not restricted to the use of predictive cues. participants appear to indicate that eye gaze cues orient
A further issue for discussion is if EVR’s lesions are attention automatically (Driver et al., 1999; Friesen &
extensive enough to disrupt two attentional processes, one Kingstone, 1998; Langton & Bruce, 1999)? Normal par-
for orienting from word cues and one for orienting from ticipants are faster to respond to validly cued targets than
gaze cues. Such an alternative explanation does not explain invalidly cued targets, even when the validly cued targets
why EVR could orient from peripheral cues, which tap appear on a minority of trials. However, in the studies just
automatic orienting. Of course, one could propose separate cited, only one cue (a small face) was presented. Neurolog-
mechanisms for automatic orienting to peripheral cues and ically normal participants may voluntarily orient attention
gaze cues. However, our proposal that gaze cues tap volun- from the gaze cue because there is only one cue and no
tary attentional processes is more parsimonious than an ac- competing information (e.g., another cue, such as a word or
count that proposes a unique automatic attentional process arrow). Participants might assume that they are expected to
for gaze cues. Our proposal makes the strong prediction orient based on the gaze cue, and they adjust their behavior
that orienting to symbolic cues and gaze cues should not accordingly. One method for circumventing participants’
be doubly dissociable. Studies of other patients will be re- strategies would be to present two central symbolic cues,
quired to test this prediction. Evidence against our position such as a face and an arrow. When the two cues point in
would need to show that attentional deficits are not caused different directions, participants should orient toward the
by perceptual impairments. That is, to reflect an attentional gazed-at location if gaze cues orient attention automatically.
impairment in using gaze cues, the perception of these gaze In a recent study (Vecera, in preparation), we displayed
cues must be intact. Future work must also be sensitive to face and arrows cues together and found that approximately
the possibility that some symbolic cues, such as arrows, half of the participants use the gaze cue to orient attention
might tap reflexive orienting: An arrow centered on fixation and half use the arrow cue to orient attention, suggesting
with an arrowhead to the left or right of fixation might that gaze cues do not have a special status compared to
capture attention in the direction the arrow points because other symbolic cues (e.g., words or arrows). These findings
1664 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665

with normal participants provide converging evidence Alivisatos, B., & Milner, B. (1989). Effects of frontal or temporal
that suggests gaze cues may involve voluntary attentional lobectomy on the use of advance information in a choice reaction time
orienting. task. Neuropsychologia, 27, 495–503.
Baddeley, A. (1996). Exploring the central executive. Quarterly Journal
Although single-case studies can yield important neuro- of Experimental Psychology. A, Human Experimental Psychology, 49A,
psychological information (Shallice, 1988), an ever-present 5–28.
concern is the generalizability of results. In EVR’s case, the Baron-Cohen, S. (1995). Mindblindness: An essay on autism and theory
brain regions affected by his lesions are similar to those of of mind. Cambridge, MA: MIT Press.
other frontal-lobe damaged patients who have been stud- Bechara, A., Damasio, H. D., Tranel, D., & Anderson, S. W. (1998).
Dissociation of working memory from decision making within the
ied in spatial cuing tasks (Alivisatos & Milner, 1989; Koski
human prefrontal cortex. Journal of Neuroscience, 18, 428–437.
et al., 1998), and EVR’s results replicate previous studies Channon, S., & Crawford, S. (2000). The effects of anterior lesions on
that have investigated word cues in larger patient groups performance of a story comprehension test: Left anterior impairment
(Alivisatos & Milner, 1989). Thus, EVR exhibits reduced cu- on a theory of mind-type task. Neuropsychologia, 38, 1006–1017.
ing effects following symbolic cues as do other frontal-lobe Chao, L. L., & Knight, R. T. (1995). Human prefrontal lesions increase
patients, suggesting that he is not an isolated case. distractability to irrelevant sensory inputs. Neuroreport, 6, 1605–1610.
Cheney, D. L., & Seyfarth, R. M. (1990). How monkeys see the world.
Some patients with frontal-lobe damage have impair- Chicago: University of Chicago Press.
ments in theory-of-mind tasks (Channon & Crawford, Damasio, H., & Damasio, A. R. (1989). Lesion analysis in
2000; Rowe, Bullock, Polkey, & Morris, 2001; Stone, neuropsychology. Oxford: Oxford University Press.
Baron-Cohen, & Knight, 1998; Stuss, Gallup, & Driver, J., Davis, G., Ricciardelli, P., Kidd, P., Maxwell, E., &
Alexander, 2001). EVR’s impairment in orienting attention Baron-Cohen, S. (1999). Gaze perception triggers reflexive visuospatial
orienting. Visual Cognition, 6, 509–540.
from eye gaze cues could be explained by such an impair-
Eslinger, P. J., & Damasio, A. R. (1985). Severe disturbance of higher
ment. However, a theory-of-mind impairment does not read- cognition after bilateral frontal lobe ablation: Patient EVR. Neurology,
ily explain EVR’s inability to orient from word cues. Based 25, 1731–1741.
on EVR’s results, and the results from other frontal patients, Friesen, C. K., & Kingstone, A. (1998). The eyes have it: Reflexive
it appears that voluntary attentional control processes may orienting is triggered by nonpredictive gaze. Psychonomic Bulletin and
be disrupted following frontal-lobe damage. Such attentional Review, 5, 490–493.
Frith, C. D., & Frith, U. (1999). Interacting minds—biological basis.
control processes are necessary for a wide range of cogni- Science, 286, 1692–1695.
tive performance, including spatial orienting from symbolic Gagliardi, J. L., Kirkpatrick-Steger, K. K., Thomas, J., Allen, G. J., &
cues and, perhaps, making predictions about others’ behav- Blumberg, M. S. (1995). Seeing and knowing: Knowledge attribution
ior that some refer to as a ‘theory-of-mind.’ Given the prox- versus stimulus control in adult humans (Homo sapiens). Journal of
imity between brain mechanisms for voluntary attentional Comparative Psychology, 109, 107–114.
Hommel, B., Pratt, J., Colzato, L., & Godijn, R. (2001). Symbolic control
control and for ‘theory-of-mind’ tasks (Frith & Frith, 1999;
of visual attention. Psychological Science, 12, 360–365.
Shallice, 2001), at least some ‘theory-of-mind’ impairments Jonides, J. (1981). Voluntary versus automatic control over the mind’s
might reflect impairments in the voluntary control of selec- eye’s movement. In J. B. Long & A. D. Baddeley (Eds.), Attention
tive attention. The present results indicate that sensitivity to and performance (Vol. IX, pp. 187–203). Hillsdale, NJ: Erlbaum.
social information, such as gaze direction, and sensitivity to Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform
evolutionarily-recent symbolic information, such as words, face area: A module in human extrastriate cortex specialized for face
perception. Journal of Neuroscience, 17, 4302–4311.
appear to require voluntary attentional processes. Kimberg, D. Y., D’Esposito, M., & Farah, M. J. (2000). Frontal lobes II:
Cognitive issues. In M. J. Farah & T. E. Feinberg (Eds.), Patient-based
approaches to cognitive neuroscience (pp. 317–326). Cambridge, MA:
Acknowledgements MIT Press.
Kingstone, A., Friesen, C. K., & Gazzaniga, M. S. (2000). Reflexive joint
attention depends on lateralized cortical connections. Psychological
The research in this paper was supported in part by grants
Science, 11, 159–166.
from the National Science Foundation (BCS 99-10727), Koski, L. M., Paus, T., & Petrides, M. (1998). Directed attention after
the National Institute of Mental Health (MH60636), the unilaterial frontal excisions in humans. Neuropsychologia, 36, 1363–
National Institute of Neurological Disease and Stroke 1371.
(P01 NS19632), and the National Institute of Aging Langton, S. R. H., & Bruce, V. (1999). Reflexive visual orienting in
(AG/NS15071). The authors wish to thank Daniel Tranel response to the social attention of others. Visual Cognition, 6, 541–568.
Langton, S. R. H., & Bruce, V. (2000). Do the eyes have it? Cues to the
for sharing data on patient EVR and Steve Luck, Maureen
direction of social attention. Trends in Cognitive Sciences, 4, 50–59.
Marron, Morris Moscovitch, and two anonymous referees Leslie, A. M. (1991). The theory of mind impairment in autism: Evidence
for useful comments and discussion. for a modular mechanism of development? In A. Whiten (Ed.), Natural
theories of mind (pp. 63–78). Oxford: Basil Blackwell.
Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal
References cortex function. Annual Review of Neuroscience, 24, 167–202.
Müller, H. J., & Rabbit, P. M. A. (1989). Reflexive and voluntary
Alivisatos, B. (1992). The role of the frontal cortex in the use of advance orienting of visual attention: Time course of activation and resistance to
information in a mental rotation paradigm. Neuropsychologia, 30, 145– interruption. Journal of Experimental Psychology: Human Perception
159. and Performance, 15, 315–330.
 S.P. Vecera, M. Rizzo / Neuropsychologia 42 (2004) 1657–1665 1665

Perrett, D. I., Hietanen, J. K., Oram, M. W., & Benson, P. J. (1992). Shallice, T. (2001). ‘Theory of mind’ and prefrontal cortex. Brain, 124,
Organization and functions of cells responsive to faces in the temporal 247–248.
cortex. Philosophical Transactions of the Royal Society of London, Stone, V. E., Baron-Cohen, S., & Knight, R. T. (1998). Frontal lobe
B335, 23–30. contributions to theory of mind. Journal of Cognitive Neuroscience,
Posner, M. I. (1980). Orienting of attention. Quarterly Journal of 10, 640–656.
Experimental Psychology, 32, 3–25. Stuss, D. T., Gallup, G. G., & Alexander, M. P. (2001). The frontal lobes
Posner, M. I., Snyder, C. R. R., & Davidson, B. J. (1980). Attention and are necessary for ‘theory of mind’. Brain, 124, 279–286.
the detection of signals. Journal of Experimental Psychology: General, Tipples, J. (2002). Eye gaze is not unique: Automatic orienting in response
109, 160–174. to uninformative arrows. Psychonomic Bulletin and Review, 9, 314–
Povinelli, D. J., & Eddy, T. J. (1996). Chimpanzees: Joint visual attention. 318.
Psychological Science, 7, 129–135. Tranel, D., Damasio, H., & Damasio, A. (1995). Double dissociation
Povinelli, D. J., Nelson, K. E., & Boysen, S. T. (1990). Inferences about between overt and covert face recognition. Journal of Cognitive
guessing and knowing by chimpanzees (Pan troglodytes). Journal of Neuroscience, 7, 425–432.
Comparative Psychology, 104, 203–210. Vecera, S. P. (in preparation). The eyes may not have it: Interrupting the
Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory allocation of ‘social attention’.
of mind? Behavioral and Brain Sciences, 4, 515–526. Vecera, S. P., & Johnson, M. H. (1995). Gaze detection and the cortical
Reaux, J. E., Theall, L. A., & Povinelli, D. J. (1999). A longitudinal processing of faces: Evidence from infants and adults. Visual Cognition,
investigation of chimpanzees’ understanding of visual perception. Child 2, 59–87.
Development, 70, 275–290. Wilkins, A. J., Shallice, T., & McCarthy, R. (1987). Frontal lesions and
Ristic, J., Friesen, C. K., & Kingstone, A. (2002). Are the eyes special? sustained attention. Neuropsychologia, 25, 359–365.
It depends on how you look at it. Psychonomic Bulletin and Review, Yantis, S., & Hillstrom, A. P. (1994). Stimulus-driven attentional
9, 507–513. capture: Evidence from equiluminant visual objects. Journal of Experi-
Rowe, A. D., Bullock, P. R., Polkey, C. E., & Morris, R. G. (2001). ‘Theory mental Psychology: Human Perception and Performance, 20, 95–
of mind’ impairments and their relationship to executive functioning 107.
following frontal lobe excisions. Brain, 124, 600–616. Yantis, S., & Jonides, J. (1984). Abrupt visual onsets and selective
Shallice, T. (1988). From neuropsychology to mental structure. New York: attention: Evidence from visual search. Journal of Experimental
Cambridge. Psychology: Human Perception and Performance, 10, 601–621.