Exp Brain Res (2011) 211:429–438

DOI 10.1007/s00221-011-2698-8

RESEARCH ARTICLE

Interpersonal memory-based guidance of attention is reduced

for ingroup members
Xun He • Anne G. Lever • Glyn W. Humphreys

Received: 8 December 2010 / Accepted: 13 April 2011 / Published online: 26 April 2011
Ó Springer-Verlag 2011

Abstract Participants jointly engaged in common tasks Keywords Working memory
Visual attention

with co-actors can be influenced in guiding their own Memory-based guidance
Interpersonal processing

attention by representations of what the co-actor also holds Ingroup
in memory (He et al. under review). This demonstrates an
effect of interpersonal memory on attention. Here, we
tested how this interpersonal memory effect is affected by Introduction
the relationship between the actors. Participants searched
for targets while maintaining images in working memory In everyday social life, people frequently have to perform
or after previewed images that co-actors had to memorise. tasks together. In those socially interactive scenarios,
We examined three groups: Caucasian strangers (low individuals can direct attention to the objects and events
ingroup relations) and two other groups with likely higher attended by others (Eilan et al. 2004; Schuch and Tipper
ingroup relations (Caucasian friends and Chinese partici- 2007), prevent attention from being allocated to locations
pants living in Britain). In all three groups, attention was inhibited by a testing partner (Frischen et al. 2009; Welsh
directed to stimuli that matched the item the individual had et al. 2005, 2007) and adjust actions to others’ behaviour
to memorise. However, images that had to be memorised (Chartrand and Bargh 1999; Knoblich and Jordan 2003).
by co-actors only attracted the attention of Caucasian For instance, Sebanz et al. (2003, 2005) studied the spatial
strangers but not the Caucasian friends and Chinese par- compatibility effect (Simon 1969) when tasks were divided
ticipants. We suggest that interpersonal memory-based among people. The spatial compatibility effect reflects the
guidance of attention is modulated by the nature of the conflict in response selection that occurs in a two-choice
relationship between individuals and reduces when indi- reaction time (RT) task when both response alternatives are
viduals have higher ingroup relations. at disposal of a single participant. The effect can be
eliminated when a single participant performs a go/nogo
version of the task, but it is restored when two participants
each perform a go/nogo version, when they appear jointly
Part of this study was reported in the 7th International Conference on
to form a complete representation of the two-choice task.
Cognitive Science (August 2010) and in a British Neuropsychological Recently, we (He et al. under review) have provided
Society meeting (October 2010). additional evidence that individuals performing a task
together also memorise the same stimuli (even if a stimulus
X. He (&)
G. W. Humphreys
is relevant only to one participant), and this interpersonal
Behavioural Brain Sciences, School of Psychology,
College of Life and Environmental Sciences, University memory can guide attentional allocation. Visual attention,
of Birmingham, Edgbaston, Birmingham B15 2TT, UK in an information-rich world, is critical to select task-rel-
e-mail: x.he@bham.ac.uk evant stimuli (Eriksen and Yeh 1985; Pashler et al. 2001;
Yantis 1998) and to prioritise their processing (Cave and
A. G. Lever
Department of Psychology and Centre for Cognitive Science, Bichot 1999). There is evidence that visual attention is
University of Turin, Turin, Italy guided in a top-down manner from stimuli held in working

123
430 Exp Brain Res (2011) 211:429–438

memory (WM) (Chelazzi et al. 1993), and this can happen stronger ingroup relationship than pairs of Caucasian
even when the stimulus is irrelevant to the task being strangers. It can also be argued that pairs of Chinese
performed. For example, when a stimulus is held in WM strangers form stronger ingroup pairs than Caucasian
while participants search for another item, search is strangers, given that Asian cultures tend to reward group
affected if the WM stimulus is re-presented in the search rather than individual success (Markus et al. 2007; Triandis
display (Downing 2000; Olivers et al. 2006; Soto et al. 1995), and, for studies run in the UK, the ingroup status of
2005, 2006b, 2008; see also Han and Kim 2009). This Chinese individuals may be increased further as the indi-
interaction between WM and attention, however, is not viduals form a minority set, emphasising their ingroup
necessarily confined to a single person and may also be status. To strongly emphasise their ingroup here, we also
found in interpersonal task-sharing setups (Böckler et al. conducted the study with Chinese students in Mandarin
2010). He et al. (under review) adapted the WM and visual Chinese as the working language. Because language is a
search paradigm by asking a pair of participants to sit carrier of social identity (Tong et al. 1999), our use of
alongside each other and perform the memory and search Mandarin Chinese should activate the Chinese culture
tasks across alternating trials. Each participant was asked system (Bond and Yang 1982; Hong et al. 2000; Kemm-
only to memorise items from a particular category (if the elmeier and Cheng 2004; Ross et al. 2002), increasing any
memory was from category A, participant A memorised it; effects of perceived similarities between the Chinese par-
if from category B, then participant B remembered it; if it ticipants (see also Briley and Wyer 2002; Hogg 2004).
was from category C, then neither participant memorised Hence, we expect that, as is the case for British friends, the
it—this provided a priming baseline to test for the effects ingroup status will be higher for the Chinese participants
of mere re-appearance of the participant’s own or the other than the Caucasian strangers.
participant’s memory in the search task). When the mem- The question then is how this ingroup status influences
ory item belonged to the participant’s category and WM and the effect of WM on the allocation of attention.
re-appeared in the search display, there was an effect of the Interpersonal WM-based guidance may be enhanced or
memory item on search (the own memory effect). More reduced by increased ingroup status, in a manner similar to
interestingly, participants were affected by re-appearance social facilitation and social inhibition effects (e.g., Bond
in the search display of the WM cue from their co-actor’s and Titus 1983). An enhancement account can be derived
category. There were minimal effects of re-presenting an from the finding that ingroup members usually have a
initial cue that belonged to the category that neither par- tendency to cooperate and to intervene more strongly in
ticipant had to memorise (the priming baseline condition). their group’s behaviour (Kramer and Brewer 1984; Tajfel
These data suggest that participants code in memory 1982). For instance, people having a more positive rela-
information that is held in a co-actor’s memory, when tionship (as is typically the case for ingroup members)
participants engage in the same task. This interpersonal show stronger joint action effects (Hommel et al. 2009). It
memory modulates the subsequent allocation of attention. follows that, in our study, ingroup members may take
These effects observed in joint action scenarios reflect particular interest in their partner’s memory, generating
interpersonal interactions in forming a common coding of strong effects of the other’s memory on their own alloca-
the task (Knoblich and Sebanz 2006, 2008; Sebanz et al. tion of attention. In contrast, an alternative prediction can
2006). For instance, joint action effects are observed when be made, namely that high ingroup relations may lead to
participants believe that they interact with another human reduced effects of the other’s memory stimulus on one’s
being, but not when interactions are with a non-human own attention. This might come about if there is greater
agent (Tsai and Brass 2007). Given the importance of trust between ingroup members (Brewer and Kramer 1985;
interpersonal interactions, then, it seems very likely that the Brewer and Yuki 2007; Williams 2001). With higher
effects of joint action and interpersonal memory may be mutual trust, participants may recruit less information from
modulated by the social relationship between the partici- WM items relevant to the co-actor’s tasks, and so there will
pants. This was investigated here. We assessed the role of be less effect of the other’s memory on the guidance of
interpersonal memory in directing attention when the each participant’s attention. The present study tested these
co-actors were British Caucasian strangers, British two predictions.
Caucasian friends and Chinese strangers. It is also important to understand that in studies of
Friendship is crucial for social grouping (Tajfel et al. interpersonal memory on attention, we can separate out two
1971). A friend can be included in one’s own identity effects: one on the memory representation itself and one on
(Aron et al. 1991) and closely related with our self-repre- attentional guidance. For example, if ingroup members take
sentation via emotional and motivational links (forming a more note of their partner’s memory items, then they should
relational self-concept; Andersen and Chen 2002). Hence, show better longer-term memory for those items than for
pairs of Caucasian friends may be expected to have a items they did not attend to, in addition to attending to those

123
Exp Brain Res (2011) 211:429–438 431

items when they are subsequently presented. It is also direct exposure to the UK society for a period of 0–6 years
possible that any effects of memory to a partner’s actions (M = 1.7 years, SE = .3 years).
are transient and do not lead to better longer-term consoli- All participants came from the University of Birming-
dation of the items, when memory is tested over the longer ham and had normal or corrected-to-normal eyesight.
term. These possibilities for effects on longer-term memory Three participants in the Caucasian stranger group, one in
as well as on attention were evaluated here. the Caucasian friends group, and two in the Chinese group
The present experiment used a modified version of the were left-handed, while all others were right-handed. Par-
WM and attention procedure (Downing 2000). As in our ticipants were tested in pairs seated next to each other. In
previous study (He et al. under review), participants were the conditions with strangers, the two participants were
tested in pairs and, depending on the item initially shown, randomly assigned and they did not know each other before
one or neither participant had to hold a previewed image in the experiment. The tasks were spelled out to each par-
WM as they performed a visual search task for a target ticipant, one instruction at a time, with the other participant
shape. The previewed image belonged either to (1) a cat- present. The participants viewed simultaneously a single
egory that the participant was told to remember (own screen that was of equal distance to each participant.
memory), (2) a category that their co-actor was told to English was used as the working language for the
remember (other’s memory) or (3) a priming baseline (a Caucasian groups. In order to reduce any priming effect
category that neither participant was told to remember). from British culture upon the Chinese participants, and to
This initial stimulus could re-appear in the search display at activate the Chinese culture system, we used Chinese as the
the location of either the search target (on valid trials) or a working language for the Chinese participants. The
distractor (invalid trials). Following He et al. (under experimenter was Chinese, and he communicated with
review), we assessed whether participants were affected by the Chinese participants in Mandarin Chinese (including
the re-presentation of their own memory item or their the instructions); all documents were translated into
co-actor’s memory item in the search display, compared Chinese for testing purposes.
with the priming baseline. Following the main experiment,
participants were also given a surprise recall task for the Stimuli and procedure
items they had been presented with. If attention to a part-
ner’s actions influences their long-term memory (LTM), The experiment was programmed and run with E-Prime
then there should also be better recall for the partner’s 1.1, and it consisted of a memory task and a subsequent
items than for neutral items, which do not have to be visual search task. There were three image categories, each
attended by either the participant or their partner. of which consisted of forty images, which were counter-
balanced across participant pairs. One category (e.g. fruit)
was requested to be remembered by one participant, and
Method another category (e.g. four-footed animal) was designated
for the memory task for the other participant. The third
Participants category (e.g., musical instrument) was not relevant to any
participant’s memory task and trials where this occurred
Three groups of age- and sex-matched volunteers partici- formed the priming baseline.
pated in the present study. The first group (Caucasian The experiment consisted of four 60-trial blocks. Each
strangers) consisted of 24 native British Caucasian college trial started with a fixation cross presented along with a
students aged between 18 and 27 years (M = 20.1 years, peripheral cue (‘X’) in the left or right field with equal
SE = .4 years), including 20 women and 4 men. The probabilities. The cue indicated which participant was to
second group (Caucasian friends) were 32 native British perform in this trial. The cue lasted for 500 ms. The fixation
Caucasian students of 18–26 years of age (M = 20.4 years, was presented 500 ms longer than the cue, followed by a
SE = .3 years). They were 23 women and 9 men and were 500-ms preview image (2.1 9 2.1°) at the centre of the
included if they self-reported as being close friends who screen, randomly selected from one of the three categories
met each other frequently (at least twice a week). The third (see above). The cued participant had to keep this image in
group (Chinese) consisted of 24 Chinese students and WM if it belonged to his/her category. After a 2,000-ms
postdoctoral fellows with an age range of 21–35 years interval, the visual search display appeared, which lasted for
(M = 24.6 years, SE = .6 years), five of whom were men 1,000 ms or until response. This display contained two
and the rest were women. All the Chinese participants were images (2.1 9 2.1°), which appeared randomly at two out of
born and raised in China and were living in UK at the time four positions (2.9° to fixation) at the corners of a virtual
of the study. At the time of the experiment, the Chinese square. One stimulus was always the previewed image on
participants had been away from China and experiencing that trial; the other was a new image that was not from any of

123
432 Exp Brain Res (2011) 211:429–438

the categories used for the preview. Each image in the search Results
display was flanked by a pair of circles or squares (.6 9 .6°).
The circles flanked the previewed image (valid) or the other Search RTs
image (invalid) on 40% of total trials, respectively, and they
did not appear (replaced by squares; catch trials) on the RTs more than three standard deviations away from the
remaining 20% of the trials. The cued participant had to mean values for any participant were removed. A three-
respond as quickly as possible to the circles by pressing ‘c’ way (3 Memory 9 2 Validity 9 3 Group) mixed-design
on the keyboard (the participant on the left) or by clicking analysis of variance (ANOVA) showed a significant effect
the left mouse button (the participant on the right); the of Group, F(2, 77) = 8.40, P \ .0006. Overall RTs were
response had to be withheld if only squares were presented. longer for the Caucasian strangers (526 ms) than for the
If the previewed image had to be memorised by the Caucasian friends (468 ms) and Chinese (461 ms) partici-
responding participant (due to it belonging to the appropriate pants, Ps \ .0008, who did not differ, P [ .6. There was
category), there was another 500-ms interval after the search also a significant main effect of Memory, F(2,
display and then a memory test. The memory test contained 154) = 18.67, P \ .0001. RTs were shortest in the other’s
two images, one of which was always the memorised image memory condition (476 ms) than the priming condition
and the other was a different image from the same image (484 ms) and the own memory condition (496 ms), all
category. These stimuli appeared side by side for 3,000 ms pairwise Ps \ .003. We also found a significant Validity
or until response. The responding participant indicated effect, F(1, 77) = 40.61, P \ .0001. This effect of Validity
which picture matched the image in memory by pressing interacted with Memory, F(2, 154) = 4.72, P \ .011, and
‘c’/’v’ (left participant) or by clicking the left/right mouse there was a Memory 9 Validity 9 Group interaction, F(4,
button (right participant) for the left/right image. The next 154) = 3.09, P \ .018. This interaction was broken down
trial followed after 2,000 ms (Fig. 1). in two ways to reflect (1) the distinct effects for ingroup
At the end of the experiment, there was a surprise and outgroup members and (2) the contrast between the
memory test in which participants were asked to write Validity effects in the other’s memory and priming
down all the previewed items they could. conditions.

Fig. 1 Example of a trial. A

peripheral cue (‘X’) at the
beginning of each trial indicated
which participant was to
perform the memory and search
tasks. The cued participant kept
the preview image in memory if
it was from his/her category or
did nothing for the preview if
the cue was from another
category. There then followed a
search display, and the task was
to make a speeded response to
the presence of circles. After the
search display, there was a
memory test where the
participant had to judge which
of two images was presented as
the preview, but only when the
previewed image was to be
memorised

123
Exp Brain Res (2011) 211:429–438 433

Contrasting ingroup and outgroup partners

An analysis of the data for the Caucasian strangers (Fig. 2)

revealed a significant effect of Memory, F(2, 46) = 4.00,
P \ .026. This effect of Memory was due to the RT dif-
ference between the conditions of own memory (537 ms)
and other’s memory (517 ms), P \ .017 (other pairwise
Ps [ .1). The main effect of Validity was significant as
well, F(1, 23) = 11.42, P \ .003. A significant Mem-
ory 9 Validity interaction, F(2, 46) = 4.48, P \ .017,
demonstrated that the Validity effect varied across the
different memory conditions. Further, separate one-way
ANOVAs for the three memory conditions revealed a
significant Validity effect for own memory, (valid 519 ms
Fig. 3 RT results of the Caucasian friends group. Significant validity
vs. invalid 555 ms), F(1, 23) = 15.74, P \ .0007, and for
effects were found in the own memory and priming conditions, but
other’s memory, (502 ms vs. 532 ms), F(1, 23) = 10.79, there was no validity effect in the other’s memory condition. Asterisks
P \ .004, but not for the priming condition, (524 ms vs. indicate significant validity effects
524 ms), F(1, 23) = .003, P [ .9. The size of the Validity
effect did not differ across the own memory and other’s
memory conditions (F(1, 23) = .16, P [ .6, for the inter-
action of Memory 9 Validity when only the own memory
and other’s memory conditions were included).
An analysis of the data for the two ingroups (Figs. 3, 4)
showed significant main effects of Memory, F(2,
108) = 16.18, P \ .0001 (own memory 475 ms, other’s
memory 456 ms, priming 464 ms, all pairwise Ps \ .004),
and Validity, F(1, 54) = 29.55, P \ .0001, but no differ-
ences between the two ingroups, (468 ms vs. 461 ms), F(1,
54) = .17, P [ .6. There was also a significant Valid-
ity 9 Memory interaction, F(2, 108) = 5.94, P \ .004,
not qualified by the ingroup type, Fs \ 2.01, Ps [ .1.
There were significant effects of Validity in the own
memory and priming conditions, F(1, 54) = 24.10, and Fig. 4 RT results for the Chinese group. A significant validity effect
F(1, 54) = 18.11, both Ps \ .0001 (mean benefits for valid was found in the own memory and priming conditions, but the effect
over invalid trials of 29 and 23 ms, respectively). But this was absent in other’s memory condition. Asterisks indicate significant
validity effects

effect was not reliable in the other’s memory condition

(a difference of 5 ms between valid and invalid trials;
F(1, 54) = 1.37, P [ .2).

Contrasting the other’s memory and priming conditions

The specific contrast between the other’s memory and

priming conditions, suggested above, was confirmed in two
further separate Validity 9 Ingroup (ingroup vs. outgroup)
ANOVAs for each memory condition (other, priming). For
both ANOVAs, there were main effects of Validity,
Fs [ 4.95, Ps \ .029, and Ingroup, Fs [ 15.27, Ps \ .0001,
and Validity 9 Ingroup interactions, Fs [ 4.47, Ps \ .038.
However, the interactions reflected different patterns for the
Fig. 2 Reaction time (RT) results for the Caucasian strangers group.
two memory conditions. In the other’s memory condition, a
A significant validity effect was found in the own and other’s memory
conditions. Asterisks indicate significant validity effects Validity effect was found for individuals with low ingroup

123
434 Exp Brain Res (2011) 211:429–438

relations (502 ms vs. 532 ms), F(1, 23) = 10.78, P \ .004,

but not for individuals with high ingroup relations (453 ms
vs. 459 ms), F(1, 55) = 1.89, P [ .1. In contrast, in the
priming condition, a Validity effect was found for ingroup
members (453 ms vs. 476 ms), F(1, 55) = 19.59, P \ .0001,
but not for outgroup participants (524 ms vs. 524 ms), F(1,
23) = .003, P [ .9.

Search accuracy

False alarm rates did not differ across conditions,

Fs \ 1.49, Ps [ .2, and had a mean value of 7.3%. A
three-way (3 Memory 9 2 Validity 9 3 Group) ANOVA
for hit rates showed a significant effect of Group, F(2,
77) = 8.46, P \ .0005, with Caucasian strangers being
less accurate (92.1%) than Caucasian friends (95.4%) and
Fig. 5 Memory recall performance. Memory was best in the own
Chinese (96.6%), Ps \ .003 (the latter two groups did not
memory condition, while memory for the other’s memory condition
differ, P [ .2). There was also a Validity effect, F(1, was better than that for the priming condition
77) = 4.57, P \ .036, which interacted with Group, F(2,
77) = 7.19, P \ .002 (all other effects were not significant, partner’s memory category (28.8%), compared with items
Fs \ 2.83, Ps [ .097). Further ANOVAs suggested that from the category defining the priming baseline (20.6%),
this Validity effect was reliable for the outgroup partici- all Ps \ .0001 (Fig. 5).
pants (valid 93.5% vs. invalid 90.6%), F(1, 23) = 8.83,
P \ .007, but not for the ingroup members (Caucasian
friends: 94.6% vs. 95.7%; Chinese: 97.2% vs. 95.9%), Discussion
Fs \ 4.15, Ps [ .053.
WM and attention
Immediate and long-term memory performance
The present study showed that visual attention was directed
Participants showed good performance in the immediate by images that had been memorised and then re-appeared
memory test (at the end of each trial), confirming that they in a visual search display, as performance was affected by
followed the instructions to keep images for the corre- whether the memorised cue fell in the same or a different
sponding category in memory. A one-way ANOVA location to the search target. This result in the own memory
showed that memory test accuracy differed significantly condition replicated previous findings in showing faster
across groups, F(2, 77) = 3.17, P \ .048. A Tukey post RTs on valid than invalid trials, consistent with attention
hoc comparisons further revealed that Chinese participants being directed to the location of the cue when it
had slightly better performance overall than the Caucasian re-appeared (Downing 2000; Olivers et al. 2006; Soto et al.
strangers (98.8% vs. 96.8%), P \ .037. Memory perfor- 2005, 2006b). In addition, we demonstrated that visual
mance for the Caucasian friends did not differ from the attention can be guided interpersonally by images that only
other two groups, Ps [ .3. a co-actor is instructed to memorise (i.e., there is an
Performance in the surprise (longer term) memory recall ‘interpersonal memory effect’). Strikingly, this result
task was analysed with a two-way (3 Memory cate- occurred only for pairs of Caucasian strangers, who
gory 9 3 Group) mixed-design ANOVA with Group as the showed no differences in memory-based guidance between
between-subject factor. There was a significant main effect the category they were instructed to memorise (the own
of Group, F(2, 77) = 23.27, P \ .0001; Caucasian friends memory condition) and the category their partner was
had better performance (37.0%) than the other two groups, instructed to memorise (the other’s memory conditions).
Ps \ .0005, and the Chinese group had better performance The validity of the initial cue did not affect the perfor-
than the Caucasian strangers (29.2% vs. 22.7%), P \ .006. mance of these participants in the priming condition,
The effect of Memory category also reached significance, confirming that the cue effects were linked to holding a
F(2, 154) = 121.90, P \ .0001, but it did not interact with representation in WM. For these participants, then, there
Group, F(4, 154) = .61, P [ .6. Further pairwise tests for was clear evidence of their attention being affected by the
Memory category showed that report was best for the items their partner was meant to memorise, even though
participant’s own memory category (39.5%) then for their items in this category were never relevant to their memory

123
Exp Brain Res (2011) 211:429–438 435

and visual search tasks. Because the other’s memory cat- LTM. There was evidence for this in LTM, where the two
egory was always irrelevant, the data indicate that inter- sets of ingroup participants scored more highly than the
personal memory guidance is involuntary, and such outgroup participants (Caucasian friend and Chinese
memories are formed and guide attention even when never strangers [ Caucasian strangers). It can also be argued
beneficial to memory. that there was stronger top-down control of search in the
It is also interesting that, in the unexpected test of ingroup participants, particularly in the other’s memory
explicit LTM, the Caucasian strangers (similarly to the condition. Top-down control in search can be brought
participants in other groups) showed better memory for about by participants partitioning the cue in WM from the
items from their own memorised category relative to items ‘template’ for the search target, so that search is directed
from their partner’s category, with memory for the latter to the search target irrespective of whether the memory
items also being better than for items in the priming con- item re-appears in the search display. Neuropsychological
dition. This indicates a dissociation between explicit LTM evidence indicates that this process is disrupted after
(affected by the assignment of stimuli to own or other’s damage to frontal lobe structures associated with execu-
memory conditions) and short-term guidance of attention tive control of cognition, so that frontal patients are
from WM (not affected by stimulus assignment to own or strongly affected by irrelevant cues in WM (Soto et al.
other’s memory groups, for these participants). It may be, 2006a). Executive processes have also been reported to be
for instance, that longer-term memory is affected by dif- enhanced in bilingual participants, and, consistent with
ferential rehearsal for own memory items, operating in this, bilingual participants are less affected than monol-
addition to whether the items are initially maintained in inguals by irrelevant cues in WM (Hernández et al. in
WM (which takes place for both the own and other press). Here, enhanced top-down control in ingroup
memory conditions). members would lead to reduced effects of WM on search
In contrast to the Caucasian strangers, though, there was generally. This would impact on any effects of the other’s
no evidence for an interpersonal memory guidance effect memory category, perhaps due to partitioning of these
with Caucasian friends or with Chinese strangers – both of items from the template used for search or even due to
whom are likely to have higher ingroup relations with their suppression of the memory stimulus. This notion is
partner (e.g., Tong et al. 1999; Triandis 1995). This result illustrated in Fig. 6. In contrast, there may remain effects
runs counter to the idea that ingroup members might take of bottom-up priming; note that Hernández et al. (in
greater account of their partner’s task representation, press) reported that bottom-up priming was the same in
thereby generating larger rather than smaller effects of the bilingual and monolingual participants, though top-down
other’s memory on attentional guidance. For this result, guidance to irrelevant WM stimuli decreased. We
there are several possible explanations since in- and out- observed positive effects of validity for ingroup members
group participants may differ along a number of dimen- in the own memory and priming conditions, which may
sions—for instance in self-presentation, arousal/anxiety reflect both (a) a lack of suppression of the priming item
levels, motivation for competition and levels of mutual in WM and (b) stronger bottom-up priming in the own
trust. The possible effects of these factors are discussed
below.
Self-presentation is the process by which individuals
socially present their self-identities, which is typically
more modest to friends (ingroup members) than to
strangers (Baumeister 1982; Tice et al. 1995). It has been
found that self-presentation competes for resources with
cognitive control processes. When a counter-normative
presentation is adopted (being modest to strangers and
self-enhancing to friends), it will impair cognitive control
(Tice et al. 1995; Vohs et al. 2005). In the present study,
without any self-presentation manipulation, the partici-
pants would likely have adopted normative presentation Fig. 6 Memory representation (from left to right) and WM guidance
styles (being modest to ingroup members and self- on attention (from top to bottom). The stimuli from the participant’s
enhancing to outgroup members), which in prior experi- own and other’s categories are represented in WM in different
ments benefit memory and action for ingroup members compartments (at different encoding strength) and further encoded
into LTM. Attention is involuntarily guided by one’s own WM
(see Tice et al. 1995; Vohs et al. 2005). It follows that content. However, the attentional guidance from interpersonal WM
there should be generally enhanced performance for representations is dependent on ingroup status—high ingroup rela-
ingroup members, in terms of both attentional control and tions reduce effects of interpersonal WM on visual attention

123
436 Exp Brain Res (2011) 211:429–438

memory condition, overriding any suppression of the own membership does not always have a negative effect on joint
memory cue in WM. action performance. For instance, Hommel et al. (2009)
A second possibility is that the in- and outgroup dif- found that the ‘interactive’ Simon effect in joint action is
ferences reflect variations in trust. Greater trust is com- enhanced when participants share a more positive rela-
monly experienced among ingroup members because of tionship. The social Simon task they used consists of a
shared experiences, views and goals (Brewer and Kramer single stimulus to be responded to by one or two partici-
1985; Brewer and Yuki 2007; Williams 2001). Higher pants according to its colour. In this task, there is stimulus-
levels of trust may reduce the required amount of attention generated prompting of who responds, contrasting with the
or neural resource deployed to the actions and memory sets present study in which the respondent was cued at the start
of others. As a consequence, ingroup members may not of each trial. The net result would be greater uncertainty in
encode items from their partner’s category into WM, and agency in the interactive Simon task (Brass et al. 2005;
hence these items do not guide attention. However, this Hommel et al. 2009; Ruby and Decety 2001), leading to
interpretation should be considered with caution, as we did stronger effects of the other’s actions on ingroup members.
not assess or manipulate the trust level between partici- Against this, we employed a long interval between the
pants, while our sole measure of attention to the other’s preview image and the search display, giving ample time
actions is based on whether items from the other’s memory for participants to decide who performed the memory task.
category influenced selection. In addition, it is not clear Under these conditions, the benefits of increased attentional
why variation in trust should lead to the overall differences control in ingroup members should lead to reduced effects
in LTM performance we observed. of shared irrelevant memories on search.
A third possibility is that the differences between in- and
outgroup member reflects contrasting arousal levels. Chal- Joint memory
lenge and threat are more generally experienced between
outgroup than ingroup members, and this may lead to higher There have now been several studies addressing the issue
arousal in out- than ingroup pairs (Stephan and Stephan of interpersonal memory representation. For instance,
1985; see also Mendes et al. 2008). Conditions of high interpersonally shared memories, so-called transactive
arousal are associated with narrowed attention, a reduced memories (Hollingshead 2000; Wegner 1986), are better
attentional capacity (Kahneman 1973; Mueller and between individuals with a close relationship (Wegner
Thompson 1984), and a tendency for people to generate et al. 1991). Similarly, memory recall is better among
overlearned/dominant responses (Zajonc 1965). The con- friends than non-friends (Andersson and Rönnberg 1995)
sequence of this will be that less attention may be paid to the and memory is more strongly biased by socially tuned
tasks of an outgroup partner. This does not fit with the messages delivered to an ingroup audience than to an
reduced effect of interpersonal memory for ingroup mem- outgroup audience (Echterhoff et al. 2005, 2008). Shteyn-
bers, and it also is difficult to explain the overall enhance- berg (2010) also suggested that stimuli experienced across
ment in memory performance. ingroup members are more accessible to individuals. Our
Finally, differences in competition between group finding of a raised overall level of memory performance for
members could be a factor. Outgroup members may have a ingroup members fits with this. Our results also indicate
greater tendency to compete than ingroup members, who that joint LTMs can dissociate from information held in
tend to generate cooperation (Kramer and Brewer 1984; WM, since all participants showed an own memory
Tajfel 1982)1. Stronger competition could again lead to an advantage in LTM, while outgroup participants showed
increased focus on the participant’s own memory items and equal effects of other’s and own memory on WM-based
not on their partner’s (e.g., de Bruijn et al. 2008), pre- modulation of attention. We have suggested that LTM may
dicting a reduced effect of shared memory on attention for be affected by additional rehearsal processes, biased to own
outgroup partners. In contrast to this, there was reduced memory items. It should also be noted that ingroup par-
influence of the other’s memory items for in- rather than ticipants showed no effect of the other’s memory stimuli on
outgroup members. It is also not clear why memory per- attention, but did recognise these items better in the longer
formance should increase overall. term. This in turn suggests that ingroup members did dif-
Whichever account of the in- and outgroup behaviours is ferentially code the other’s memory stimuli in WM, but
put forward, it is important to note that increased ingroup were able to successfully partition these stimuli from the
search template, so that search was impervious to stimuli
1
Friends may show greater involvement (competition) in a compet- from their partner’s category.
itive game. However there is overwhelming prior evidence indicating
Overall, these arguments highlight that the effects of
that increased cooperation is more likely in non-competitive situa-
tions as in the present study (where no performance feedback was WM on attention are modulated by the group membership
given). of the participants, with effects differing according to

123
Exp Brain Res (2011) 211:429–438 437

whether co-acting participants have relatively high or low Eilan N, Hoerl C, McCormack T, Roessler J (2004) Joint attention:
ingroup membership. The control of WM effects on communication and other minds. Oxford University Press, Oxford
Eriksen CW, Yeh Y (1985) Allocation of attention in the visual field.
attention are modulated by group context. J Exp Psychol Hum Percept Perform 11:583–597
Frischen A, Loach D, Tipper SP (2009) Seeing the world through
Acknowledgments This work was supported by a grant from the another person’s eyes: simulating selective attention via action
Economic and Social Research Council, UK. observation. Cognition 111:212–218
Han SW, Kim M-S (2009) Do the contents of working memory
Ethical statements The study has been approved by the Ethics capture attention? Yes, but cognitive control matters. J Exp
Committee in University of Birmingham and hence been performed Psychol Hum Percept Perform 35:1292–1302
in accordance with the Declaration of Helsinki standards. All par- Hernández M, Humphreys GW, Costa A (in press) Escaping capture:
ticipants gave informed consent prior to their participation in the bilingualism modulates distraction from working memory.
study. The authors do not have any financial relationship with the Cognition
organisation that sponsored the research. Hogg MA (2004) Social identity, self-categorization, and communi-
cation in small groups. In: Ng S-H, Candlin CN, Chiu C-y (eds)
Language matters: communication, culture, and identity. City
University of Hong Kong Press, Hong Kong, pp 221–243
References Hollingshead AB (2000) Perceptions of expertise and transactive
memory in work relationships. Group Proc Intergr Relat
Andersen SM, Chen S (2002) The relational self: an interpersonal 3:257–267
social-cognitive theory. Psychol Rev 109:619–645 Hommel B, Colzato LS, van den Wildenberg WPM (2009) How
Andersson J, Rönnberg J (1995) Recall suffers from collaboration: social are task representations? Psychol Sci 20:794–798
joint recall effects of friendship and task complexity. Appl Cogn Hong Y-y, Morris MW, Chiu C-y, Martı́nez V (2000) Multicultural
Psychol 9:199–211 minds: a dynamic constructivist approach to culture and
Aron A, Aron EN, Tudor M, Nelson G (1991) Close relationships as cognition. Am Psychol 55:709–720
including other in the self. J Personal Soc Psychol 60:241–253 Kahneman D (1973) Attention and effort. Prentice Hall, Englewood
Baumeister RF (1982) A self-presentational view of social phenom- Cliffs
ena. Psychol Bull 91:3–26 Kemmelmeier M, Cheng BY-M (2004) Language and self-construal
Böckler A, Knoblich G, Sebanz N (2010) Socializing cognition. In: priming: a replication and extension in a Hong Kong sample.
Glatzeder BM, Goel V, Müller A (eds) Towards a theory of J Cross Cult Psychol 35:705–712
thinking. Springer, Berlin, pp 233–250 Knoblich G, Jordan S (2003) Action coordination in individuals and
Bond CF Jr, Titus LJ (1983) Social facilitation: a meta-analysis of 241 groups: learning anticipatory control. J Exp Psychol Learn Mem
studies. Psychol Bull 94:265–292 Cogn 29:1006–1016
Bond MH, Yang KS (1982) Ethnic affirmation versus cross-cultural Knoblich G, Sebanz N (2006) The social nature of perception and
accommodation: the variable impact of questionnaire language action. Curr Dir Psychol Sci 15:99–104
on Chinese bilinguals from Hong Kong. J Cross Cult Psychol Knoblich G, Sebanz N (2008) Evolving intentions for social
13:169–185 interaction: from entrainment to joint action. Philos Trans R
Brass M, Derrfuss J, von Cramon DY (2005) The inhibition of Soc Lon B Biol Sci 363:2021–2031
imitative and overlearned responses: a functional double disso- Kramer RM, Brewer MB (1984) Effects of group identity on resource
ciation. Neuropsychologia 43:89–98 use in a simulated commons dilemma. J Personal Soc Psychol
Brewer MB, Kramer RM (1985) The psychology of intergroup 46:1044–1057
attitudes and behavior. Annu Rev Psychol 36:219–243 Markus HR, Uchida Y, Omoregie H, Townsend SSM, Kitayama S
Brewer MB, Yuki M (2007) Culture and social identity. In: Kitayama (2007) Going for the gold: models of agency in Japanese and
S, Cohen D (eds) Handbook of cultural psychology. Guilford American contexts. Psychol Sci 17:103–112
Press, London, pp 307–322 Mendes WB, Major B, McCoy S, Blascovich J (2008) How
Briley DA, Wyer RS Jr (2002) The effect of group membership attributional ambiguity shapes physiological and emotional
salience on the avoidance of negative outcomes: implications for responses to social rejection and acceptance. J Personal Soc
social and consumer decisions. J Consum Res 27:157–178 Psychol 94:278–291
Cave KR, Bichot NP (1999) Visuospatial attention: beyond a Mueller JH, Thompson WB (1984) Test anxiety and distinctiveness of
spotlight model. Psychonom Bull Rev 6:204–223 personal information. In: Ploeg HM, Schwarzer R, Spielberger
Chartrand TL, Bargh JA (1999) The chameleon effect: the perception- CD (eds) Advances in test anxiety research, vol 3. Erlbaum,
behavior link and social interaction. J Personal Soc Psychol Hillsdale, pp 21–37
76:893–910 Olivers CNL, Meijer F, Theeuwes J (2006) Feature-based memory-
Chelazzi L, Miller EK, Duncan J, Desimone R (1993) A neural basis for driven attentional capture: visual working memory content
visual search in inferior temporal cortex. Nature 363:345–347 affects visual attention. J Exp Psycho Hum Percept Perform
de Bruijn ERA, Miedl SF, Bekkering H (2008) Fast responders have 32:1243–1265
blinders on: ERP correlates of response inhibition in competi- Pashler H, Johnston JC, Ruthruff E (2001) Attention and perfor-
tion. Cortex 44:580–586 mance. Annu Rev Psychol 52:629–651
Downing PE (2000) Interactions between visual working memory and Ross M, Xun WQE, Wilson AE (2002) Language and the bicultural
selective attention. Psychol Sci 11:467–473 self. Personal Soc Psychol Bull 28:1040–1050
Echterhoff G, Higgins ET, Groll S (2005) Audience-tuning effects on Ruby P, Decety J (2001) Effect of subjective perspective taking
memory: the role of shared reality. J Personal Soc Psychol during simulation of action: A PET investigation of agency. Nat
89:257–276 Neurosci 4:546–550
Echterhoff G, Higgins ET, Kopietz R, Groll S (2008) How Schuch S, Tipper SP (2007) On observing another person’s actions:
communication goals determine when audience tuning biases influences of observed inhibition and errors. Percept Psychophys
memory. J Exp Psychol Gen 137:3–21 69:828–837

123
438 Exp Brain Res (2011) 211:429–438

Sebanz N, Knoblich G, Prinz W (2003) Representing others’ actions: Tong Y-Y, Hong Y-Y, Lee S-L, Chiu C-Y (1999) Language use as a
just like one’s own? Cognition 88:B11–B21 carrier of social identity. Int J Intercult Relat 23:281–296
Sebanz N, Knoblich G, Prinz W (2005) How two share a task: Triandis HC (1995) Individualism and collectivism. Westview,
corepresenting stimulus-response mappings. J Exp Psychol Hum Boulder
Percept Perform 31:1234–1246 Tsai C-C, Brass M (2007) Does the human motor system simulate
Sebanz N, Bekkering H, Knoblich G (2006) Joint action: bodies and Pinocchio’s action? Coacting with a human hand versus a
minds moving together. Trends Cogn Sci 10:70–76 wooden hand in a dyadic interaction. Psychol Sci 18:1058–1062
Shteynberg G (2010) A silent emergence of culture: the social tuning Vohs KD, Baumeister RF, Ciarocco NJ (2005) Self-regulation and
effect. J Personal Soc Psychol 99:683–689 self-presentation: regulatory resource depletion impairs impres-
Simon JR (1969) Reactions towards the source of stimulation. J Exp sion management and effortful self-presentation depletes regu-
Psychol 81:174–176 latory resources. J Personal Soc Psychol 88:632–657
Soto D, Heinke D, Humphreys GW, Blanco MJ (2005) Early, Wegner DM (1986) Transactive memory: a contemporary analysis of
involuntary top-down guidance of attention from working the group mind. In: Mullen B, Goethals GR (eds) Theories of
memory. J Exp Psychol Hum Percept Perform 31:248–261 group behaviour. Springer, New York, pp 185–208
Soto D, Humphreys GW, Heinke D (2006a) Dividing the mind: the Wegner DM, Erber R, Raymond P (1991) Transactive memory in
necessary role of the frontal lobes in separating memory from close relationships. J Personal Soc Psychol 61:923–929
search. Neuropsychologia 44:1282–1289 Welsh TN, Elliott D, Anson JG, Dhillon V, Weeks DJ, Lyons JL,
Soto D, Humphreys GW, Heinke D (2006b) Working memory can Chua R (2005) Does Joe influence Fred’s action? Inhibition of
guide pop-out search. Vis Res 46:1010–1018 return across different nervous systems. Neurosci Lett
Soto D, Hodsoll J, Rotshtein P, Humphreys GW (2008) Automatic 385:99–104
guidance of attention from working memory. Trends Cogn Sci Welsh TN, Lyons J, Weeks DJ, Anson JG, Chua R, Mendoza J, Elliott
12:342–348 D (2007) Within- and between-nervous-system inhibition of
Stephan WG, Stephan CW (1985) Intergroup anxiety. J Soc Issues return: observation is as good as performance. Psychonom Bull
41:157–175 Rev 14:950–956
Tajfel H (1982) Social psychology of intergroup relations. Annu Rev Williams M (2001) In whom we trust: group membership as an
Psychol 33:1–39 affective context for trust development. Acad Manag Rev
Tajfel H, Billig M, Bundy R, Flament C (1971) Social categorization 26:377–396
and intergroup behaviour. Eur J Soc Psychol 1:149–178 Yantis S (1998) Control of visual attention. In: Pashler H (ed)
Tice DM, Butler JL, Muraven MB, Stillwell AM (1995) When Attention. Psychology Press, Hove, pp 223–256
modesty prevails: differential favorability of self-presentation to Zajonc R (1965) Social facilitation. Science 149:269–274
friends and strangers. J Personal Soc Psychol 69:1120–1138

123