Inhibition of return in response to gaze cues: The roles of time course and fixation cue

Recent studies have demonstrated that orienting of attention in response to nonpredictive gaze cues arises rapidly and automatically, in a similar fashion to peripheral sudden onset cueing. However, while peripheral cues consistently elicit inhibition of return (IOR) at about 300 ms following cue onset, very little is known about inhibition effects in response to gaze cues. The present experiments systematically examined the conditions under which IOR arises with such cues. Reliable inhibition effects were obtained. Importantly, IOR emerged only at long cue-target intervals and only when a second cue actively triggered attention away from the cued location. This suggests that compared to sudden onset cueing, gaze cueing results in both prolonged facilitation and a delayed onset of inhibition processes. Thus, although both types of cues elicit very similar orienting effects in terms of their basic behavioural outcomes, there are more subtle differences between gaze and peripheral cues with respect to the maintenance and quality of those cueing effects across time.     

Resolving conflicting views: Gaze and arrow cues do not trigger rapid reflexive shifts of attention

It has become widely accepted that the direction of another individual's eye gaze induces rapid, automatic, attentional orienting, due to it being such a vital cue as to where in our environment we should attend. This automatic orienting has also been associated with the directional-arrow cues used in studies of spatial attention. Here, we present evidence that the response-time cueing effects reported for spatially nonpredictive gaze and arrow cues are not the result of rapid, automatic shifts of attention. For both cue types, response-time effects were observed only for long-duration cue and target stimuli that overlapped temporally, were largest when the cues were presented simultaneously with the response-relevant target, and were driven by a slowing of responses for invalidly cued targets rather than speeding for validly cued ones. These results argue against automatic attention-orienting accounts and support a novel spatial-incongruency explanation for a whole class of rapid behavioural cueing effects.     

Spatial cueing by social versus nonsocial directional signals

A cost-benefit analysis was used to investigate whether reflexive effects in a spatial cueing task are stronger when target location is cued by another person's gaze rather than arrows because the relative contribution of attentional shifts versus automatic priming is greater in the case of gaze cues. Across four experiments, nonpredictive arrows triggered rapid facilitatory, inhibition-less priming that peaked at 300-500ms SOA and then died away; across three experiments, nonpredictive gaze cues triggered facilitation-plus-costs at SOAs of 300-400ms or more, suggesting that gaze cues trigger stronger (and longer) attentional effects. At 200 ms SOA, gaze cues triggered facilitation-without-cost, consistent with the view that facilitatory effects accrue more rapidly due to earlier automatic priming, whereas costs are manifest slightly later, when attentional effects come online. There was some evidence that nonpredictive gaze cues trigger long-lasting congruency effects so long as observers maintain their preparedness to respond. Findings support the view that gaze is a unique symbolic directional cue.     

The influence of social and symbolic cues on observers' gaze behaviour

Research has shown that social and symbolic cues presented in isolation and at fixation have strong effects on observers, but it is unclear how cues compare when they are presented away from fixation and embedded in natural scenes. We here compare the effects of two types of social cue (gaze and pointing gestures) and one type of symbolic cue (arrow signs) on eye movements of observers under two viewing conditions (free viewing vs. a memory task). The results suggest that social cues are looked at more quickly, for longer and more frequently than the symbolic arrow cues. An analysis of saccades initiated from the cue suggests that the pointing cue leads to stronger cueing than the gaze and the arrow cue. While the task had only a weak influence on gaze orienting to the cues, stronger cue following was found for free viewing compared to the memory task.     

On the uniqueness of attentional capture by uninformative gaze cues: facilitation interacts with the Simon effect and is rarely followed by IOR.

Orienting to an uninformative peripheral cue is characterized by a brief facilitation followed by a long-lasting inhibition once attention is removed from the cued location. Although central gaze cues cause reflexive orienting, the inhibitory effect that is relatively ubiquitous following exogenous orienting to uninformative peripheral cues has been relatively rare. We hypothesized that IOR might be seen following gaze-induced orienting if attention were effectively returned to centre by a return gaze or return flash. The timecourse of gaze-directed orienting was measured by varying the interval between the gaze cue and a peripheral target requiring an orientation discrimination (permitting measurement of the Simon effect). Significant facilitation was observed at all but the longest SOA tested, 2,880 ms, by which time the facilitation had disappeared with no evidence of IOR. Gaze-induced cuing (which was unaffected by return cue condition) interacted with the Simon effect, decreasing it at the gazed-at location, a pattern that is not seen with more typical endogenous and exogenous cuing.     

A direct link between gaze perception and social attention

How information is exchanged between the cognitive mechanisms responsible for gaze perception and social attention is unclear. These systems could be independent; the "gaze cueing" effect could emerge from the activation of a general-purpose attentional mechanism that is ignorant of the social nature of the gaze cue. Alternatively, orienting to social gaze direction might be directly determined by the operation of cognitive mechanisms specifically dedicated to gaze perception. This second notion is the dominant assumption in the literature, but there is little direct support for this account. Here, we systematically manipulated observers' perception of gaze direction by implementing a gaze adaptation paradigm. Gaze cueing was reduced only in conditions where perception of specific averted gaze stimuli was impaired (Experiment 1). Adaptation to a pointing stimulus failed to impact gaze cueing (Experiment 2). Overall, these data suggest a direct link between the specific operation of gaze perception mechanisms and the consequential orienting of attention.     

A biased-competition approach to spatial cueing: Combining empirical studies and computational modelling

We examined the effects of cue luminance on visual orienting. Experiment 1 established that the commonly-found early facilitation and late inhibition of return (IOR) effects were independent of cue luminance with single cues in terms of their amplitude, although IOR was delayed in the low compared to the high luminance cue condition. In contrast, Experiment 2 revealed that, with dual cues of mixed luminance, both facilitation and IOR effects were found only with bright cues. When cues had equal luminance, however, there were cueing effects for two cued locations but only when the cues were bright. The data were accommodated in a neural network model of biased competition in which cueing effects emerge at more than one location provided input activation is sufficient to overcome competitive damping of the selection system.     

Automatic attentional orienting to other people’s gaze in schizophrenia

Explicit tests of social cognition have revealed pervasive deficits in schizophrenia. Less is known of automatic social cognition in schizophrenia. We used a spatial orienting task to investigate automatic shifts of attention cued by another person’s eye gaze in 29 patients and 28 controls. Central photographic images of a face with eyes shifted left or right, or looking straight ahead, preceded targets that appeared left or right of the cue. To examine automatic effects, cue direction was non-predictive of target location. Cue–target intervals were 100, 300, and 800?ms. In non-social control trials, arrows replaced eye-gaze cues. Both groups showed automatic attentional orienting indexed by faster reaction times (RTs) when arrows were congruent with target location across all cue–target intervals. Similar congruency effects were seen for eye-shift cues at 300 and 800?ms intervals, but patients showed significantly larger congruency effects at 800?ms, which were driven by delayed responses to incongruent target locations. At short 100-ms cue–target intervals, neither group showed faster RTs for congruent than for incongruent eye-shift cues, but patients were significantly slower to detect targets after direct-gaze cues. These findings conflict with previous studies using schematic line drawings of eye-shifts that have found automatic attentional orienting to be reduced in schizophrenia. Instead, our data indicate that patients display abnormalities in responding to gaze direction at various stages of gaze processing—reflected by a stronger preferential capture of attention by another person’s direct eye contact at initial stages of gaze processing and difficulties disengaging from a gazed-at location once shared attention is established.     

Spatiotemporal distribution of facilitation and inhibition of return arising from the reflexive orienting of covert attention

Currently, there is debate regarding both the spatial and temporal relationship between facilitation and inhibition of return (IOR) components of attention, as observed on the covert orienting of visual attention task (COVAT). These issues were addressed in a series of experiments where the spatial and temporal relationships between cue and target were manipulated. Facilitation occurred only when the stimulus onset asynchrony (SOA) was short and there was temporal overlap between cue and target. IOR occurred only when SOA was long and there was no temporal overlap between cue and target. Facilitation encompassed the cued location and all locations between the cue and fixation, whereas IOR arose for the entire cued hemifield. These findings suggest that the facilitation and IOR found on COVATs that use noninformative peripheral cues are separable and stimulus-driven processes.     

The role of the oculomotor system in covert social attention

Observing a change in gaze direction triggers a reflexive shift of attention and appears to engage the eye-movement system. However, the functional relationship between social attention and this oculomotor activation is unclear. One extremely influential hypothesis is that the preparation of a saccadic eye movement is necessary and sufficient for a covert, reflexive shift of attention (the premotor theory of attention; Rizzolatti et al., 1994). Surprisingly, this theory has not been directly tested with respect to reflexive gaze cueing. In order to address this issue, gaze cueing, peripheral cueing, and arrow cueing were examined under conditions in which some stimuli appeared at locations that could not become the goal of a saccadic eye movement. It was observed that peripheral cues failed to elicit reflexive attentional orienting when targets appeared beyond the range of eye movements. Similarly, nonpredictive arrow cues were ineffective when targets could not become the goal of a saccade. In contrast, significant gaze-cueing effects were still observed when targets were beyond the range of eye movements. These data demonstrate that the mechanisms involved in gaze cueing are dissociated from those involved in exogenous orienting to peripheral or arrow cues. Furthermore, the findings suggest that, unlike peripheral cueing and reflexive arrow cueing, gaze cueing is independent of oculomotor control. We conclude that the premotor theory does not offer a compelling explanation for gaze cueing.     

The effects of onsets and offsets on visual attention

 Two experiments are presented which examine the effect of onset and offset cues on early occurring attentional cueing effects and later occurring inhibition of return (IOR). The first experiment compared the effects of single onset cues, single offset cues, and simultaneous onset and offset cues (at opposite locations) at a 100-ms stimulus-onset-asynchrony (SOA) and IOR at a 900-ms SOA. Whereas the first experiment examined these conditions with choice localization keypress responses, the second experiment used simple detection keypress responses. Both experiments found that onset and offset cues presented in isolation produce early facilitation and late IOR. When onset and offset cues were simultaneously presented, facilitation but not IOR was found with localization responses, while neither facilitation or IOR was found with detection responses. Overall, these findings suggest that offset cues can be treated in the same manner as onset cues by the attentional system, although the onset cues may have priority in orienting attention when targets must be localized in space. Received: 26 May 2000 / Accepted: 31 January 2001     

Automatic Mechanisms for Social Attention Are Culturally Penetrable

Are mechanisms for social attention influenced by culture? Evidence that social attention is triggered automatically by bottom‐up gaze cues and is uninfluenced by top‐down verbal instructions may suggest it operates in the same way everywhere. Yet considerations from evolutionary and cultural psychology suggest that specific aspects of one's cultural background may have consequence for the way mechanisms for social attention develop and operate. In more interdependent cultures, the scope of social attention may be broader, focusing on more individuals and relations between those individuals. We administered a multi‐gaze cueing task requiring participants to fixate a foreground face flanked by background faces and measured shifts in attention using eye tracking. For European Americans, gaze cueing did not depend on the direction of background gaze cues, suggesting foreground gaze alone drives automatic attention shifting; for East Asians, cueing patterns differed depending on whether the foreground cue matched or mismatched background cues, suggesting foreground and background gaze information were integrated. These results demonstrate that cultural background influences the social attention system by shifting it into a narrow or broad mode of operation and, importantly, provides evidence challenging the assumption that mechanisms underlying automatic social attention are necessarily rigid and impenetrable to culture.     

Eye gaze versus arrows as spatial cues: two qualitatively different modes of attentional selection

This study aimed to evaluate the type of attentional selection (location- and/or object-based) triggered by two different types of central noninformative cues: eye gaze and arrows. Two rectangular objects were presented in the visual field, and subjects' attention was directed to the end of a rectangle via the observation of noninformative directional arrows or eye gaze. Similar experiments with peripheral cues have shown an object-based effect: faster target identification when the target is presented on the cued object as compared to the uncued object, even when the distance between target and cue was the same. The three reported experiments aimed to compare the location- and object-based attentional orienting observed with arrows and eye gaze, in order to dissociate the orienting mechanisms underlying the two types of orienting cues. Results showed similar cueing effects on the cued versus oppositely cued locations for the two cue types, replicating several studies with nonpredictive gaze and arrow cues. However, a pure object-based effect occurred only when an arrow cue was presented, whereas a pure location-based effect was only found for eye-gaze cues. It is suggested that attention is nonspecifically directed to nearby objects when a noninformative arrow is used as cue, whereas it is selectively directed to a specific cued location when noninformative eye gaze is used. This may be mediated by theory of mind mechanisms.     

Why does the gaze of others direct visual attention?

Viewing another person directing his or her gaze can produce automatic shifts of covert visual attention in the same direction. This holds true even when the task‐relevant target is much more likely to occur at the uncued location. These findings, along with other evidence, have been taken to suggest that gaze represents a “special” stimulus—the foundation of a social cognition system that can make inferences about the mental states of other people. However, gaze‐driven cueing effects could simply be due to spatial compatibility between cue and target. We compared the attentional effects of gaze shifts to a face with the tongue extended laterally to the left or right. When tongue direction was a nonpredictive cue, we found cueing effects from tongues that were indistinguishable from those produced by gaze. However, in contrast to previous findings with gaze, tongue cues did not overcome a validity manipulation in which targets were four times more likely to appear at the uncued location. We conclude that simple attentional cueing effects from gaze may be better explained by spatial compatibility, and that more complex, unique features of cueing from gaze may be better indices into perceptual systems specialized for social cognition.     

Abrupt onsets and gaze direction cues trigger independent reflexive attentional effects

The present study investigated whether the direction of visual signals is influenced independently by two automatic visual orienting phenomena: orienting to a gazed-at location and inhibition of return (IOR) to the location of an abrupt onset. A schematic face served as both a nonpredictive gaze direction cue and an abrupt onset cue. Results indicated that target detection was facilitated at the gazed-at location, and that it was inhibited at the abrupt onset location. Importantly, these two different reflexive attention effects were triggered by the same event and exhibited overlapping time courses. Moreover, the IOR effect did not vary as a function of the facilitatory effect of gaze. These findings strongly suggest that reflexive attention to gaze direction and reflexive inhibition to an abrupt onset are independent processes, and that gaze direction does not produce IOR at the gazed-at location.     

Modulation of reflexive orienting to gaze direction by facial expressions

Facial expression and gaze perception are thought to share brain mechanisms but behavioural interactions, especially from gaze-cueing paradigms, are inconsistent. We conducted a series of gaze-cueing studies using dynamic facial cues to examine orienting across different emotional expression and task conditions, including face inversion. Across experiments, at a short stimulus–onset asynchrony (SOA) we observed both an expression effect (i.e., faster responses when the face was emotional versus neutral) and a cue validity effect (i.e., faster responses when the target was gazed-at), but no interaction between validity and emotion. Results from face inversion suggest that the emotion effect may have been due to both facial expression and stimulus motion. At longer SOAs, validity and emotion interacted such that cueing by emotional faces, fearful faces in particular, was enhanced relative to neutral faces. These results converge with a growing body of evidence that suggests that gaze and expression are initially processed independently and interact at later stages to direct attentional orienting.     

Is “Inhibition of Return” due to the inhibition of the return of attention?

Inhibition of Return (IOR) is usually explained in terms of orienting–reorienting of attention, emphasizing an underlying mechanism that inhibits the return of attention to previously selected locations. Recent data challenge this explanation to the extent that the IOR effect is observed at the location where attention is oriented to, where no reorienting of attention is needed. To date, these studies have involved endogenous attentional selection of attention and thus indicate a dissociation between the voluntary attention of spatial attention and the IOR effect. The present work demonstrates a dissociation between the involuntary orienting of spatial attention and the IOR effect. We combined nonpredictive peripheral cues with nonpredictive central orienting cues (either arrows or gaze). The IOR effect was observed to operate independent of involuntary spatial orienting. These data speak against the “reorienting hypothesis” of IOR. We suggest an alternative explanation whereby the IOR effect reflects a cost in detecting a new event (the target) at the location where another event (a cue) was coded before.     

The role of attention in the facilitation effect and another "inhibition of return"

The present study investigated how anticipation of a target's appearance affects human attention to gaze cues provided by a schematic face. Subjects in a 'catch' group received a high number of 'catch' trials, in which no target stimulus appeared. Subjects in the control group did not receive any catch trials. As in previous studies, both groups showed a facilitation effect to the cued location during shorter stimulus onset asynchrony (SOA). In both groups, an analysis of eye movements confirmed that subjects' eyes remained on the fixation point, ruling out the possibility that the facilitation effect was due to shifting eye movements (saccades) as opposed to a shift in covert attention. But while the control group's response time (RT) decreased as SOA increased, the catch group's RT had a U-shaped pattern and the facilitation effect to the cued location was reversed at the longest SOA (1005 ms). These results suggest that subjects in the catch group disengaged their attention during long SOAs because they expected the trial to be a catch trial. This disengagement of attention during long SOAs results in a delay before attention could be re-focused to the previous location regardless of the cue validity ["IOR (inhibition of return)"-like-phenomenon]. Unlike the conventional IOR, we suggest that this "IOR"-like phenomenon caused by an unpredictive central gaze cue is likely to be mediated by an endogenous mechanism.     

Facilitation and inhibition caused by the orienting of attention in propositional reasoning tasks

In an attempt to study the orienting of attention in reasoning, we developed a set of propositional reasoning tasks structurally similar to Posner's (1980) spatial cueing paradigm, widely used to study the orienting of attention in perceptual tasks. We cued the representation in working memory of a reasoning premise, observing whether inferences drawn using that premise or a different, uncued one were facilitated, hindered, or unaffected. The results of Experiments 1a, 1b, 1c, and 1d, using semantically (1a–1c) or statistically (1d) informative cues, showed a robust, long-lasting facilitation for drawing inferences from the cued rule. In Experiment 2, using uninformative cues, inferences from the cued rule were facilitated with a short stimulus onset asynchrony (SOA), whereas they were delayed when the SOA was longer, an effect that is similar to the “inhibition of return” (IOR) in perceptual tasks. Experiment 3 used uninformative cues, three different SOAs, and inferential rules with disjunctive antecedents, replicating the IOR-like effect with the long SOAs and, at the short SOA, finding evidence of a gradient-like behaviour of the facilitation effect. Our findings show qualitative similarities to some effects typically observed in the orienting of visual attention, although the tasks did not involve spatial orienting.     

Facilitation and inhibition caused by the orienting of attention in propositional reasoning tasks

In an attempt to study the orienting of attention in reasoning, we developed a set of propositional reasoning tasks structurally similar to Posner's (1980) spatial cueing paradigm, widely used to study the orienting of attention in perceptual tasks. We cued the representation in working memory of a reasoning premise, observing whether inferences drawn using that premise or a different, uncued one were facilitated, hindered, or unaffected. The results of Experiments 1a, 1b, 1c, and 1d, using semantically (1a-1c) or statistically (1d) informative cues, showed a robust, long-lasting facilitation for drawing inferences from the cued rule. In Experiment 2, using uninformative cues, inferences from the cued rule were facilitated with a short stimulus onset asynchrony (SOA), whereas they were delayed when the SOA was longer, an effect that is similar to the “inhibition of return” (IOR) in perceptual tasks. Experiment 3 used uninformative cues, three different SOAs, and inferential rules with disjunctive antecedents, replicating the IOR-like effect with the long SOAs and, at the short SOA, finding evidence of a gradient-like behaviour of the facilitation effect. Our findings show qualitative similarities to some effects typically observed in the orienting of visual attention, although the tasks did not involve spatial orienting.