Seeing the Unseen: Autism Involves Reduced Susceptibility toInattentional Blindness

John SwettenhamUniversity College London

Anna RemingtonUniversity College London and University of Oxford

Patrick Murphy, Maike Feuerstein, Kelly Grim, and Nilli LavieUniversity College London

Objective: Attention research in individuals with autism spectrum disorder (ASD) has produced con-flicting results. Some findings demonstrate greater distractibility while others suggest superior focusedattention. Applying Lavies load theory of attention to account for this discrepancy led us to hypothesizeincreased perceptual capacity in ASD. Preliminary support for our hypothesis has so far been found foradults with ASD with reaction time (RT) and signal detection sensitivity measures. Here we test the novelprediction we derived from this hypothesis that children with ASD should have lower rates of inatten-tional blindness than controls. Method: Twenty-four children with ASD (mean age 10 years 10months) and 39 typically developing children (age and IQ matched) took part in the study. We assessedthe effects of perceptual load on the rates of inattentional blindness in each group. Participantsperforming a line discrimination task in either a high load or low load condition were presented with anunexpected extra stimulus on a critical trial. Performance on the line judgment task and rates of detectionand stimulus identification were recorded. Results: Overall rates of detection and identification werehigher in the ASD group than in the controls. Moreover, whereas both detection and identification rateswere significantly lower in the high (compared with low) load conditions for the controls, these wereunaffected by load in the ASD group. Conclusion: Reduced inattentional blindness rates under load inASD suggests higher perceptual capacity is a core feature, present from childhood and leading to superiorperformance in various measures of perception and attention.

Keywords: autism spectrum disorder, inattentional blindness, attention, perceptual load, development

Research examining visual selective attention in individualswith autism spectrum disorder (ASD) has produced conflictingresults. One set of findings demonstrates increased levels of dis-tractibility in ASD relative to mental age-matched controls (Bu-rack, 1994; Christ, Holt, White, & Green, 2007; Ciesielski,Courchesne, & Elmasian, 1990), while other findings demonstratesuperior performance on visual search tasks. For example, indi-

viduals with ASD are faster and more accurate than mental-agematched controls at locating an embedded figure within a linedrawing (e.g., Jolliffe & Baron-Cohen, 1997; Shah & Frith, 1993),searching for local elements in a global figure (Plaisted, Swetten-ham, & Rees, 1999) and finding a target presented among multiplenontarget items (Joseph, Keehn, Connolly, Wolfe, & Horowitz,2009; ORiordan, 2004; ORiordan, Plaisted, Driver, & Baron-Cohen, 2001; Plaisted, ORiordan, & Baron-Cohen, 1998). So,while a number of studies suggest visual selective attention isimpaired in ASD leading to greater irrelevant distractor process-ing, performance on visual search tasks suggests superior visualselective attention.

Here we propose an account for this apparent discrepancy byapplying Lavies Load Theory of attention and cognitive control(e.g., Lavie, 2005) to ASD. According to Load Theory, the extentof distractor processing depends on the level of perceptual loadinvolved in the task. Perceptual load refers to the amount oftask-relevant information; for example, the number of differenttask stimuli or the perceptual requirements of a task (Lavie, 1995;Lavie & Tsal, 1994). A high load task might consist of a largenumber of task-relevant items in a visual search array, or a subtleperceptual discrimination. A lower load task would consist offewer items in a visual search array, or a more gross perceptualdiscrimination (e.g., Lavie, 2005, 2010). Central to the theory arethe assumptions that while perception is a limited-capacity pro-

This article was published Online First January 13, 2014.John Swettenham, Developmental Science, University College London;

Anna Remington, Institute of Cognitive Neuroscience, University CollegeLondon; Department of Experimental Psychology, University of Oxford;Patrick Murphy, Maike Feuerstein, and Kelly Grim, Developmental Sci-ence, University College London; Nilli Lavie, Institute of Cognitive Neu-roscience, University College London.

We thank Dr Eleni Matechou for her advice regarding statistical anal-yses. Preparation of this article was supported by the Department ofDevelopmental Science (JS) and the Scott Family Junior Research Fellow-ship (AR). We gratefully acknowledge the efforts of staff and pupils atschools in London who participated in the research, the National AutisticSociety (United Kingdom) and the charity Resources for Autism.

Correspondence concerning this article should be addressed to JohnSwettenham, Developmental Science, University College London, Chan-dler House, 2 Wakefield Street, London WC1N 1PF. E-mail:j.swettenham@ucl.ac.uk

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

Neuropsychology 2014 American Psychological Association2014, Vol. 28, No. 4, 563570 0894-4105/14/$12.00 DOI: 10.1037/neu0000042

563

cess, the processing of information that can be accommodatedwithin perceptual capacity is mandatory. As a consequence, per-ceptual processing proceeds on all items relevant to a current goal,as well as those that are irrelevant until capacity is exhausted.Thus, in tasks of low perceptual load (e.g., requiring participants toidentify a target stimulus from just two possible items) any sparecapacity left over from the processing of task-relevant stimuli willspill over into the processing of task-irrelevant distractor stimuli.In contrast, tasks involving high perceptual load (e.g., requiringsearch for a target stimulus in an array of four or more similaritems) engage full perceptual capacity, leaving little or no room fordistractor processing.

We reasoned that the pattern of reduced resistance to distractorson some selective visual attention tasks and enhanced performanceon the visual search tasks may reflect a higher perceptual capacityin individuals with ASD compared with controls. Because LoadTheory asserts that it is not possible to assign any less than totalcapacity, individuals with a higher perceptual capacity will neces-sarily process more information, whether relevant or irrelevant tothe task. This accounts for superior performance on tasks wherethat further information is relevant to the search (Joseph et al.,2009; ORiordan et al., 2001; ORiordan, 2004; Plaisted et al.,1998), and increased distractor processing on tasks where thefurther information consists of irrelevant distractors (e.g., Burack,1994; Remington, Swettenham, Campbell, & Coleman, 2009).

We have provided preliminary evidence for this hypothesis byshowing that increased levels of perceptual load in a search task(manipulated through the search set size) have led to reducedresponse competition effects produced by an irrelevant distractorin typical adults but not in adults with ASD. The latter continuedto show a response competition effect at higher load, indicatingthat they perceived the distractor (Remington et al., 2009). Thisinitial study, however, only provided indirect evidence of en-hanced perceptual capacity in ASD as it relied on measuring aninterference effect on RTs which could also be influenced byvarious other factors such as effects on response selection ratherthan perception, and overall susceptibility to distraction. Moredirect support has been recently provided in a study that used asignal detection method to assess the effects of load on perceptualsensitivity (Remington, Swettenham, & Lavie, 2012). The findingsshowed that increased levels of load in a visual search task led toreduced perceptual sensitivity (as measured with d=) for typicaladults, but left perceptual sensitivity unaffected in adults withASD, in line with our load hypothesis.

Our previous findings may have important implications for thephenomenon of inattentional blindness (when people fail to noticea conspicuous visual event, e.g., a man dressed in a gorilla suit) inindividuals with ASD. Because the level of perceptual load hasbeen shown to be a critical factor in inattentional blindness(Cartwright-Finch & Lavie, 2007), our hypothesis of enhancedcapacity leads to the prediction that individuals with ASD wouldbe less likely than matched controls to experience inattentionalblindness in tasks of higher load. That is, they would show en-hanced levels of awareness, being able to notice additional infor-mation when a task is taxing on capacity. The inattentional blind-ness paradigm is a well-established measure of awareness in themainstream cognitive literature of attention. To assess awarenessfor a stimulus that should be entirely unattended participants areasked to attend to a primary task (e.g., discriminating line length)

and an unexpected stimulus is then presented on the last trial in atask-irrelevant location (e.g., in the periphery for a task concerningthe display center). Awareness is then measured for this putativelyunattended stimulus on that trial, by asking participants if theynoticed anything extra on the last task display. The inattentionalblindness measure is only based on that single report for eachparticipant because once participants have been asked about theirawareness one cannot be sure that the stimulus will be entirelyunattended on subsequent trials, hence the single critical-trialnature of the paradigm.

In the present study, we aimed to establish further our loadhypothesis, testing the implications for perceptual awareness inchildren with ASD. Specifically, we asked whether higher percep-tual capacity is present already in childhood by assessing theeffects of perceptual load on the extent to which children withASD (compared with mental age-matched controls) suffer frominattentional blindness or are able to notice an extra item that is notpart of the task.

Method

Participants

A total of 39 typically developing children and 28 children witha diagnosis of ASD participated in the study. The typically devel-oping children attended an ordinary state local primary school. Thechildren with ASD attended a specialist autism primary school thatadmits only children who have received a diagnosis of ASD.Informed consent was received in compliance with APA ethicalguidelines. All participants with ASD had received a clinicaldiagnosis of ASD from a trained, independent clinician using theADOS and following criteria listed in the Diagnostic and Statis-tical Manual of Mental Disorders, Fourth Edition (AmericanPsychiatric Association, 1994). The diagnostic notes for the chil-dren with ASD indicated no other neurological or comorbid psy-chiatric condition. The study used a between-participants designand participants within both groups were randomly allocated toeither a high load or a low load condition. Participants wereexcluded if they obtained a score of less than five correct on the sixnoncritical trials, were incorrect on the critical trial, or wereincorrect on a final control trial. These rather stringent exclusioncriteria were used to ensure that all participants are engaging in thecentral task when the unexpected stimulus appears. Extensivepiloting was conducted before task administration in the schools toincrease the likelihood that most children would be able to performthe primary line-length discrimination task and so pass our strin-gent performance criterion. Overall 4 participants (all with ASD)scored less than five correct noncritical trials (3 from the low loadcondition and 1 from the high load condition) and therefore wereremoved from the sample. There may be a number of reasons whythese 4 children performed poorly on the line discrimination task.For example they may have lacked motivation to perform the taskor misunderstood the instructions. For the purposes of our exper-iment though, we only wished to include participants who hadclearly engaged in the line-length discrimination (of either high orlow load) at the same time that the irrelevant stimulus appeared.

The remaining participants were correct on the critical trial andwere able to correctly detect and identify the critical stimulus in afinal control trial, leading to no further exclusions. From the

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

564 SWETTENHAM ET AL.

remaining 24 children with ASD, 11 were in the low load condi-tion and 13 in the high load. From the 39 typically developingchildren 19 were in the low load condition and 20 in the highload condition. The ASD group was matched with the TD groupfor chronological age and for nonverbal IQ using the RavensStandard Progressive Matrices (Raven, Raven, & Court, 1998;Table 1).

Independent samples t tests indicated no significant differencebetween the groups overall or between subgroups in the low andhigh load conditions in chronological age or Ravens IQ score, (allp values .6).

Stimuli and ProcedureMicrosoft Visual Basic (version 6) was used to create computer-

based stimuli that were presented on a 15-in laptop screen (screenresolution 1024 768 pixels). Viewing distance was 60 cm.

In the high load condition, a target cross with a shorter armsubtending 3.35 and a longer arm subtending 3.9 appeared. Inthe low load condition, a target cross with a shorter arm subtending1.25 and a longer arm subtending 3.9 appeared. Each cross wasblack and appeared on a white background. The longer line waseither horizontal (H) or vertical (V) on each trial (3 H and 3 Vappeared randomly across the first 6 trials, and either H or Vcounterbalanced across participants on Trial 7; Figure 1a).

Each child was tested individually in a room at their school.They were first shown an example of each cross, presented on anA4 sheet of paper. The experimenter checked that the child couldcorrectly report which line was longest in each example, either theline going across or the line going up and down. They werethen told that crosses would be shown very briefly on the computerscreen and for each cross they would be asked to indicate whichline of the cross was longer.

A fixation circle subtending 0.15 appeared at the center of thescreen at the beginning of each trial (1,500 ms), followed by ablank display (96 ms), then a centrally located target cross (110ms) immediately followed by a visual mask (a mesh pattern ofblack lines on a white background, 496 ms). The sequence endedwith a blank screen which remained until the child had made aresponse. Participants in both conditions were required to judgewhether the longer line was vertical or horizontal. The responsecould either be verbal (e.g., the line going across) or the partic-ipant could point to one of the two illustrations of the cross placedalongside the computer. The experimenter then entered the re-sponses on the computer. Each participant completed seven trials.

On the seventh, critical trial, a black outline square shape (sidessubtending 0.3) appeared in addition to the cross. This criticalstimulus appeared in one of four peripheral locations, counterbal-anced between participants, in one of the four quadrants of thescreen as defined by the lines of the cross. Each one of theselocations was equidistant from fixation (at 3.2 eccentricity) andequidistant from the two closest lines of the cross (Figure 1b).

Following the line judgment response on this trial, participantswere asked whether they had noticed anything else appearing onthe screen that had not been present before. If the participantreported seeing something they were asked to indicate where onthe screen the critical stimulus had appeared. Participants werethen presented with a choice of 4 shapes (Figure 1c) and askedwhich one of the shapes they had seen.

The critical trial was then repeated in a final control trial.Participants were told to ignore the cross and instead to look to seeif they could see anything else on the screen. Reports of thequadrant where the critical stimulus appeared and the shape of thecritical stimulus were again noted. Only participants who reportedawareness of the critical stimulus and could identify location andshape of the critical stimulus were included in further analysis.

Results

Task Performance

A two-way between-subjects analysis of variance (ANOVA)was conducted on the number of correct responses on the crossjudgment task with the factors of diagnostic group (ASD vs. TD)and load condition (high vs. low). Mean and SD detection andidentification scores are shown in Table 1. There was a main effectof load, F(1, 59) 8.50, p .01, p2 0.13, with participants inthe low load condition scoring higher (M 6.50, SD 0.57) thanparticipants in the high load condition (M 6.03, SD 0.73).

Figure 1. (a) Illustration of stimuli used in the low load condition (eitherone of the two top crosses shown here appears on each trial in randomorder) and the high load condition (one of the two bottom crosses appearson each trial in random order); (b) Illustration of a high load critical trial,with the critical stimulus (small square) in the top right quadrant; (c) Fourshapes presented to participants who reported seeing the detection stimu-lus.

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

565REDUCED INATTENTIONAL BLINDNESS IN AUTISM

There was no main effect of group, F(1, 59) 2.74, p .10, p2 0.04, and no interaction between group and load, F(1, 59) 0.208,p .65, p2 0.004.

These findings indicate that both groups engaged in the task andthat the manipulation of load was effective in increasing the taskdifficulty. (Recall that the numerically small effect of load online-length discrimination is because of the restricted range en-forced by our exclusion criteria; see the Method section).

Awareness ReportsParticipants who reported that they had noticed something on the

critical trial, and correctly identified in which quadrant the critical stim-ulus appeared, were classified as having detected the criticalstimulus. Participants who were also able to identify the criticalstimulus either by pointing to the correct shape from the fourshapes presented, or by verbally describing a square, were classi-fied as having correctly identified the critical stimulus. Figure 2shows the percentage of participants from each group who wereable to detect and identify the critical stimulus.

DetectionBecause this was a prospective binomial study design, a satu-

rated logistic regression model was fitted (using IBM SPSS Sta-

tistics for Windows, Version 20.0) with the binary variable de-tection as the response, and group and load as factors.

The likelihood ratio test indicated a significant interaction be-tween group and load 2(1) 7.95, p .005. This confirms adifferent effect of load on detection between the groups. Whereasfor the controls detection rate in low load condition (42.1% ofparticipants) was significantly higher than in high load (0% ofparticipants), 2(1) 13.72, p .001; for the ASD group, therewas no significant association between load and detection, 2(1).001, p .97 (54.5% of participants in the low load condition, and53.8% of participants in the high load condition; Figure 2).

In all conditions, detection rates were higher in the ASD groupthan control group resulting in overall detection rates of 54.2% and20.5%, respectively.

The significant reduction for the control group in rates ofinattentional blindness from 42.1% in the low load conditions to 0in the high load conditions confirms that our manipulation of taskload was successful in determining rates of inattentional blindnessin the TD population.

The lack of effect of load on rates of detection in ASD (despiteno greater deterioration in the line-discrimination task perfor-mance with higher load, as described in the previous section) is aspredicted from our increased perceptual capacity hypothesis.

IdentificationThe identification results showed a similar pattern to the detec-

tion results. The likelihood ratio test indicated an associationbetween group and load that was approaching significance,2(1) 3.42, p .06. Individual group analyses revealed asignificant interaction between load and identification for thecontrol group, 2(1) 4.58, p .03, indicating a greater likeli-hood of correct identification in the low load (15.8% of partici-pants) than in the high load (0% of participants) conditions, and noassociation between load and identification for the ASD group(correct identification in 54.5% of participants in the low load and53.8% of participants in the high load conditions), 2(1) .001,p .973.

In all conditions, identification rates were higher in the ASDgroup than control group resulting in overall identification rates of54.2% and 7.7%, respectively.

Table 1Descriptive Statistics for Each Group

Group StatisticAge

(years:months) Ravens IQCross judgmentscore (out of 7)

Autism spectrum disorder (n 24)Low load (n 11) M 10:09 105.09 6.73

SD 1:5 11.57 0.46Range 9:213:5 90135

High load (n 13) M 10:11 104.78 6.15SD 1:6 8.81 0.80Range 9:913:10 90110

Control (n 39)Low load (n 19) M 10:10 104.32 6.37

SD 0:8 12.46 0.59Range 9:1011:11 90128

High load (n 20) M 10:8 109.16 5.95SD 0:8 11.48 0.69Range 9:1012:1 92126

Figure 2. Percentage detection and identification rates for low and highload subgroups.

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

566 SWETTENHAM ET AL.

DiscussionOur findings demonstrate that children with ASD have reduced

rates of inattentional blindness and reduced impact of the level ofperceptual load in a task on awareness, despite equally high levelsof task performance as those of controls. These findings suggestthat ASD involves increased visual perceptual capacity from child-hood as we detail below.

The reduced rates of inattentional blindness found in ASDoverall clearly show an advantage in perceptual awareness in theASD group. Critically, while both groups showed a similarly highlevel of task performance in both high and low load conditions, theawareness rate was reduced by high load for typically developingchildren, whereas for the children with ASD the awareness ratewas unaffected by the increased perceptual load of the primaryvisual task and remained high in the high load condition too.

Note that because the increased rate of awareness in ASD wasnot at the cost of performance on the primary line-length discrim-ination task itself, these findings suggest increased capacity. Thechildren with ASD were able to detect the presence of an unex-pected, task-irrelevant stimulus in addition to meaningfully engag-ing with the central task as shown by the finding of a main effectof load on the line discrimination performance for both groups. Forthe ASD group, the results were also equivalent across the mea-sures of detection and identification judgments, in further supportof enhanced perceptual capacity.

The current demonstration of reduced inattentional blindnessunder load in children with ASD is important because it suggeststhat higher perceptual capacity is already present in childhood,rather than emerging in adulthood as a result of atypical develop-ment. Previous demonstrations of increased perceptual capacity inASD have been exclusively with adults (Remington et al., 2009;2012). In addition, our group effect was found despite both groupsbeing matched for performance on the Ravens Progressive Ma-trices (RPM). This strengthens our findings still further as ASD-related perceptual peaks are potentially accounted for in the RPMmatching. If anything, one might expect a perceptual advantage tobe masked by this matching.

Our increased perceptual capacity hypothesis not only accountsfor the present findings but also may be able to account for theapparent discrepancy between previous reports of superior perfor-mance of individuals with ASD on visual attention (e.g., visualsearch and embedded figures), but reduced resistance to distractorson other selective attention tasks. According to load theory allstimuli in the visual field, regardless of their task relevance, will beprocessed automatically until perceptual capacity is exhausted.Applying load theory, we would therefore expect individuals witha higher perceptual capacity to process more information andtherefore be less prone to inattentional blindness as we haveshown. In cases where the information is task-relevant this wouldlead to superior performance, as is seen in visual search tasks andembedded figures tasks (ORiordan, 2004; ORiordan et al., 2001;Plaisted et al., 1998; Shah & Frith, 1983; Shah & Frith, 1993).Where the information is irrelevant, as in distractor tasks (Adams& Jarrold, 2012; Burack, 1994; Christ et al., 2007; Geurts et al.,2008; Remington et al., 2009), then this would result in increasedprocessing of distractor stimuli, leading some to propose thatautism involves a distractability deficit. However, note that in ourtask the irrelevant stimulus was not designed to be distracting (e.g.,

was not response competing, cf. the flanker tasks). Our findings ofan awareness advantage in this task, therefore, provides furthersupport for our interpretation that both bodies of data demonstrateenhanced perceptual capacity rather than showing a reduced abilityto control distraction.

Furthermore, our account leads to the specific prediction thatdifferences between ASD and controls will be most apparent onvisual search tasks when the perceptual load is high, allowing theASD group to take advantage of their ability to process moreinformation than controls. A close analysis of existing studiesusing the visual search paradigm shows that the data supports thisprediction. For example, participants with ASD display superiorperformance when target-distractor similarity is high (ORiordan& Plaisted, 2001), when feature search is more difficult (Kemner,van Ewijk, van Engeland, & Hooge, 2008; ORiordan et al., 2001)and on conjunctive versus feature search tasks (ORiordan, 2004;Plaisted et al., 1998). On all these tasks, the ASD group advantageis also more prominent when array sizes are high and when thetarget is absent, conditions that would favor individuals with a highperceptual capacity because they involve the scanning of a largenumber of items.

For visual attention tasks that require participants to ignoreirrelevant distracters, our account also makes the specific predic-tion that group differences will be most apparent when perceptualload is high, this time with the ASD group showing higher levelsof distraction than mental age matched controls. The logic here isthat a high level of perceptual load in the central task will exhaustcapacity in control participants, leaving no resources for distractorprocessing, but will not exhaust capacity in participants with ASDleaving spare capacity and allowing resources to spill over andautomatically process irrelevant distractors. Again, existing studiesconfirm this prediction (Remington et al., 2009; Remington et al.,2012), and the present findings show a similar pattern.

Our account and present results might also explain why previousstudies examining susceptibility to change blindness in ASD haveproduced equivocal results. Change blindness refers to the diffi-culty participants have in detecting the difference between twopictures when they are presented successively with a brief inter-ruption (Simons & Levin, 1997). Group differences have beenreported when complex scenes (high load) are used (Fletcher-Watson et al., 2012; Smith & Milne, 2009). Children with ASDdetect the change more quickly than controls under such condi-tions. However, when pairs of images of isolated objects (lowload) were used in a change blindness task, no group differencewas found (Burack et al., 2009). Again, our increased perceptualcapacity hypothesis of ASD would predict these findings, becausethe higher capacity in the ASD participants would lead to theprocessing of more stimuli in a high perceptual load scene leadingto faster detection of the changed item. This advantage would beabsent in a low perceptual load scene because only a limitednumber of items need to be processed and this can be achievedequally by both ASD children and controls.

The single trial nature of the inattentional blindness paradigmmeans that this study relies on a small data set. We shouldtherefore be cautious not to overstate the importance of the find-ings from this study alone. We note however that the convergenceof the present results with previous results in studies assessing bothdistractor processing and detection as a function of load in ASD,that have used longer versions of load tasks (Remington et al.,

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

567REDUCED INATTENTIONAL BLINDNESS IN AUTISM

2009; Remington, Swettenham & Lavie, 2012), further adds sup-port for our interpretation.

Currently, there are no established measures that directly assessvisual perceptual capacity in the mainstream attention and visualperception literature. In the absence of such measures researcherstypically rely on measures such as failure or success in dual taskperformance similar to our current study design (see also Macdon-ald & Lavie, 2008, and Remington et al., 2012, in the case ofASD). In line with this rationale, the results from the present study,as well as our previous studies, are most likely to be explained byan enhanced perceptual capacity. In each case continued process-ing of additional stimuli that are not part of the primary task isfound at higher levels of load in ASD compared with controls,while there are no group differences at low levels of load. Never-theless, the goal of establishing direct measures of perceptualcapacity is an important direction for future research.

Unfortunately, an increased perceptual capacity in ASD is un-likely to have an impact on the ability to learn what is relevant andwhat is not, even though it may enable the processing of moreinformation. In load theory, perceptual processing capacity and theability to prioritize the focus of attention on relevant rather thanirrelevant information are considered to be two distinct functions.The function of paying attention to what is relevant and what isirrelevant is considered a high-level cognitive control function thathas been shown to be dissociable from perceptual capacity (e.g.,Lavie, 2000; Lavie, Hirst, de Fockert, & Viding, 2004).

What neural mechanisms might underlie the increased percep-tual capacity in ASD? One clue comes from research with typicaladults on the neural basis of processing load. It is now wellestablished that perceptual load in the context of a visual searchtask elicits responses in the posterior parietal cortex, especially theinferior parietal sulcus (IPS; Wojciulik & Kanwisher, 1999). Morerecent studies have established a neural signature of gradual in-creases in perceptual load in the IPS, using tasks such as visualsearch or motion tracking (Jovivich et al., 2001; Mitchell &Cusack, 2008). Furthermore, this increase in parietal activity isaccompanied by reduced baseline level of visual cortex excitabilityin areas unrelated to the task (Carmel, Thorne, Rees & Lavie,2011; Muggleton, Lamb, Walsh, & Lavie, 2008) and a reduction invisual cortex activity in response to task-irrelevant stimuli (Bah-rami, Lavie, & Rees, 2007; Pinsk, Doniger, & Kastner, 2004;Schwartz et al., 2005; Yi, Woodman, Widders, Marois, & Chun,2004). These findings suggest that neural activity in parietal andvisual cortical areas mediates perceptual capacity in neurotypicalindividuals.

Given the likely neural mechanism in typical adults, we predictthat in ASD increased perceptual capacity may be related togreater availability of parietal and visual cortex resources. Consis-tent with this proposal are anatomical observations of increasedgray matter volume in the parietal cortex in ASD (Ashtari et al.,2007; Brieber et al., 2007) and evidence of overconnectivity in thevisual cortex (Just, Cherkassky, Keller, Kana, & Minshew, 2007;Kta, Mottron, & Bertone, 2010). Furthermore, functional neuro-imaging has revealed stronger activation in the visual cortex onvarious perceptual tasks in which individuals with ASD displaybehavioral superiorities, such as visual search (Keehn, Brenner,Palmer, Lincoln, & Muller, 2008), embedded figures (Manjaly etal., 2007; Ring et al., 1999), block design (Hubl et al., 2003), andmatrix reasoning (Soulires et al., 2009). It will be important to

directly test this proposal in future studies by measuring corticalactivity while manipulating load in ASD.

We have now demonstrated, through our behavioral studies,increased perceptual capacity in both adults and children withASD. Support for our account with data from children suggeststhat increased capacity may be a core feature of the disorder ratherthan a result of unusual environmental input. For example, a wellknown feature of ASD is a preference for processing local ratherthan global information (Happ & Frith, 2006). This preferencecould be considered a phenomenon of increased perceptual capac-ity. Alternatively the preference to attend to detail could contributeto the development of an increased perceptual capacity, being acause rather than consequence. Future studies testing our accountwith even younger populations may be useful in furthering ourunderstanding of the origins of increased perceptual capacity.

We also note that our studies with both adults and children withASD have so far only involved intellectually able individuals (IQ 90) and have not involved individuals with a dual diagnosis(e.g., ASD and attention deficit disorder). This might therefore beconsidered a select group. In addition, we have not yet includedother developmentally disordered groups as a comparison. Futureresearch may establish whether increased perceptual capacity isspecific to ASD and whether only intellectually able individualswith ASD have increased perceptual capacity.

Finally, a greater understanding of what characterizes this un-usual attentional and perceptual style may help in the design ofeducational and therapeutic programs which capitalize on therelative strengths associated with ASD.

ReferencesAdams, N. C., & Jarrold, C. (2012). Inhibition in autism: Children with

autism have difficulty inhibiting irrelevant distractors but not prepotentresponses. Journal of Autism and Developmental Disorders, 42, 10521063. doi:10.1007/s10803-011-1345-3

American Psychiatric Association. (1994). Diagnostic and Statistical Man-ual of Mental Disorders (4th ed.) Washington DC: Author.

Ashtari, M., Nichols, S., McIlree, C., Spritzer, L., Adesman, A., & Arde-kani, B. (2007, November). Novel imaging technique shows gray matterincrease in brains of autistic children. In Annual Meeting of the Radio-logical Society of North America. Chicago. IL.

Bahrami, B., Lavie, N., & Rees, G. (2007). Attentional load modulatesresponses of human primary visual cortex to invisible stimuli. CurrentBiology, 17, 509513. doi:10.1016/j.cub.2007.01.070

Brieber, S., Neufang, S., Bruning, N., Kamp-Becker, I., Remschmidt, H.,Herpertz-Dahlmann, B., . . . Konrad, K. (2007). Structural brain abnor-malities in adolescents with autism spectrum disorder and patients withattention deficit/hyperactivity disorder. Journal of Child Psychology andPsychiatry, 48, 12511258. doi:10.1111/j.1469-7610.2007.01799.x

Burack, J. A. (1994). Selective attention deficits in persons with autism:Preliminary evidence of an inefficient attentional lens. Journal of Ab-normal Psychology, 103, 535543. doi:10.1037/0021-843X.103.3.535

Burack, J. A., Joseph, S., Russo, N., Shore, D. I., Porporino, M., & Enns,J. (2009). Change detection in naturalistic pictures among children withautism. Journal of Autism and Developmental Disorders, 39, 471479.doi:10.1007/s10803-008-0647-6

Carmel, D., Thorne, J. D., Rees, G., & Lavie, N. (2011). Perceptual loadalters visual excitability. Journal of Experimental Psychology: HumanPerception and Performance, 37, 13501360. doi:10.1037/a0024320

Cartwright-Finch, U., & Lavie, N. (2007). The role of perceptual load ininattentional blindness. Cognition, 102, 321340. doi:10.1016/j.cognition.2006.01.002

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

568 SWETTENHAM ET AL.

Christ, S. E., Holt, D. D., White, D. A., & Green, L. (2007). Inhibitorycontrol in children with autism spectrum disorder. Journal of Autism andDevelopmental Disorders, 37, 1155. doi:10.1007/s10803-006-0259-y

Ciesielski, K. T., Courchesne, E., & Elmasian, R. (1990). Effects offocused selective attention tasks on event-related potentials in autisticand normal individuals. Electroencephalography and Clinical Neuro-physiology, 75, 207220. doi:10.1016/0013-4694(90)90174-I

Fletcher-Watson, S., Leekam, S. R., Connolly, B., Collis, J. M., Findlay,J. M., McConachie, H., . . . Rodgers, J. (2012). Attenuation of changeblindness in children with autism spectrum disorders. British Journal ofDevelopmental Psychology, 30, 446 458. doi:10.1111/j.2044-835X.2011.02054.x

Happ, F., & Frith, U. (2006). The weak central coherence account:Detail-focused cognitive style in autism spectrum disorders. Journal ofAutism and Developmental Disorders, 36, 525. doi:10.1007/s10803-005-0039-0

Hubl, D., Bolte, S., Feineis-Matthews, S., Lanfermann, H., Federspiel, A.,Strik, W., . . . Dierks, T. (2003). Functional imbalance of visual path-ways indicates alternative face processing strategies in autism. Neurol-ogy, 61, 12321237. doi:10.1212/01.WNL.0000091862.22033.1A

Jolliffe, T., & Baron-Cohen, S. (1997). Are people with autism andAsperger syndrome faster than normal on the Embedded Figures Test?Child Psychology & Psychiatry, 38, 527534. doi:10.1111/j.1469-7610.1997.tb01539.x

Joseph, R. M., Keehn, B., Connolly, C., Wolfe, J., & Horowitz, T. (2009).Why is visual search superior in autism spectrum disorder? Develop-mental Science, 12, 10831096. doi:10.1111/j.1467-7687.2009.00855.x

Jovivich, J., Peters, R. J., Koch, C., Braun, J., Chang, L., & Ernst, T.(2001). Brain areas specific for attentional load in a motion-trackingtask. Journal of Cognitive Neuroscience, 13, 10481058. doi:10.1162/089892901753294347

Just, M. A., Cherkassky, V. L., Keller, T. A., Kana, R. K., & Minshew,N. J. (2007). Functional and anatomical cortical underconnectivity inautism: Evidence from an FMRI study of an executive function task andcorpus callosum morphometry. Cerebral Cortex, 17, 951961. doi:10.1093/cercor/bhl006

Keehn, B., Brenner, L., Palmer, E., Lincoln, A. J., & Muller, R. A. (2008).Functional brain organization for visual search in ASD. Journal of theInternational Neuropsychological Society, 14, 9901003. doi:10.1017/S1355617708081356

Kta, L., Mottron, L., & Bertone, A. (2010). Far visual acuity is unre-markable in autism: Do we need to focus on crowding? Autism Re-search, 3, 333341. doi:10.1002/aur.164

Kemner, C., van Ewijk, L., van Engeland, H., & Hooge, I. (2008). Briefreport: Eye movements during visual search tasks indicate enhancedstimulus discriminability in subjects with PDD. Journal of Autism andDevelopmental Disorders, 38, 553557. doi:10.1007/s10803-007-0406-0

Lavie, N. (1995). Perceptual load as a necessary condition for selectiveattention. Journal of Experimental Psychology: Human Perception andPerformance, 21, 451468. doi:10.1037/0096-1523.21.3.451

Lavie, N. (2000). Selective attention and cognitive control: Dissociatingattentional functions through different types of load. In S. Monsell & J.Driver (Eds.), Attention and performance XVIII (pp. 175194). Cam-bridge, MA: MIT Press.

Lavie, N. (2005). Distracted and confused?: selective attention under load.Trends in Cognitive Science, 9, 7582. doi:10.1016/j.tics.2004.12.004

Lavie, N. (2010). Attention, distraction, and cognitive control under load.Current Directions in Psychological Science, 19, 143148.

Lavie, N., Hirst, A., de Fockert, J. W., & Viding, E. (2004). Load theoryof selective attention and cognitive control. Journal of ExperimentalPsychology: General, 133, 339354. doi:10.1037/0096-3445.133.3.339

Lavie, N., & Tsal, Y. (1994). Perceptual load as a major determinant of thelocus of selection in visual attention. Perception & Psychophysics, 56,183197. doi:10.3758/BF03213897

Macdonald, J. S., & Lavie, N. (2008). Load induced blindness. Journal ofExperimental Psychology: Human Perception and Performance, 34,10781091. doi:10.1037/0096-1523.34.5.1078

Manjaly, Z. M., Bruning, N., Neufang, S., Stephan, K. E., Brieber, S.,Marshall, J. C., . . . Fink, G. R. (2007). Neurophysiological correlates ofrelatively enhanced local visual search in autistic adolescents. Neuroim-age, 35, 283291. doi:10.1016/j.neuroimage.2006.11.036

Mitchell, D. J., & Cusack, R. (2008). Flexible, capacity-limited activity ofposterior parietal cortex in perceptual as well as visual short-termmemory tasks. Cerebral Cortex, 18, 17881798. doi:10.1093/cercor/bhm205

Muggleton, N., Lamb, R., Walsh, V., & Lavie, N. (2008). Perceptual loadmodulates visual cortex excitability to magnetic stimulation. Journal ofNeurophysiology, 100, 516519. doi:10.1152/jn.01287.2007

ORiordan, M. A. (2004). Superior visual search in adults with autism.Autism, 8, 229248. doi:10.1177/1362361304045219

ORiordan, M., & Plaisted, K. (2001). Enhanced discrimination in autism.Quarterly Journal of Experimental Psychology: A, 54, 961979.

ORiordan, M. A., Plaisted, K. C., Driver, J., & Baron-Cohen, S. (2001).Superior visual search in autism. Journal of Experimental PsychologyHuman Perception and Performance, 27, 719730. doi:10.1037/0096-1523.27.3.719

Pinsk, M. A., Doniger, G. M., & Kastner, S. (2004). Push-pull mechanismof selective attention in human extrastriate cortex. Journal of Neuro-physiology, 92, 622629. doi:10.1152/jn.00974.2003

Plaisted, K., ORiordan, M., & Baron-Cohen, S. (1998). Enhanced visualsearch for a conjunctive target in autism: A research note. Journal ofChild Psychology and Psychiatry, 39, 777783. doi:10.1017/S0021963098002613

Plaisted, K., Swettenham, J., & Rees, L. (1999). Children with autism showlocal precedence in a divided attention task and global precedence in aselective attention task. Journal of Child Psychology and Psychiatry, 40,733742. doi:10.1111/1469-7610.00489

Raven, J., Raven, J. C., & Court, J. H. (1998). Manual for Ravensprogressive matrices and vocabulary scalesSection 1: General over-view. Oxford, United Kingdom: Oxford Psychologists Press.

Remington, A., Swettenham, J., Campbell, R., & Coleman, M. (2009).Selective attention and perceptual load in autism spectrum disorder.Psychological Science, 20, 13881393. doi:10.1111/j.1467-9280.2009.02454.x

Remington, A. M., Swettenham, J. G., & Lavie, N. (2012). Lightening theload: Perceptual load impairs visual detection in typical adults but not inautism. Journal of Abnormal Psychology, 121, 544551. doi:10.1037/a0027670

Ring, H., Baron-Cohen, S., Wheelwright, S., Williams, S. C., Brammer,M., Andrew, C., . . . Bullmore, E. T. (1999). Cerebral correlates ofpreserved cognitive skills in autism. A functional MRI study of Embed-ded Figures Task performance. Brain, 122, 13051315. doi:10.1093/brain/122.7.1305

Schwartz, S., Vuilleumier, P., Hutton, C., Maravita, A., Dolan, R., &Driver, J. (2005). Attentional load and sensory competition in humanvision: Modulation of fMRI responses by load at fixation during task-irrelevant stimulation in the peripheral visual field. Cerebral Cortex, 15,770786. doi:10.1093/cercor/bhh178

Shah, A., & Frith, U. (1983). An islet of ability in autistic children: Aresearch note. Journal of Child Psychology and Psychiatry, 24, 613620. doi:10.1111/j.1469-7610.1983.tb00137.x

Shah, A., & Frith, U. (1993). Why do autistic individuals show superiorperformance on the block design task? Journal of Child Psychology andPsychiatry, 34, 13511364. doi:10.1111/j.1469-7610.1993.tb02095.x

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

569REDUCED INATTENTIONAL BLINDNESS IN AUTISM

Simons, D. J., & Levin, D. T. (1997). Change blindness. Trends inCognitive Science, 1, 261267. doi:10.1016/S1364-6613(97)01080-2

Smith, H., & Milne, E. (2009). Reduced change blindness suggests en-hanced attention to detail in individuals with autism. Journal of ChildPsychology and Psychiatry, 50, 300 306. doi:10.1111/j.1469-7610.2008.01957.x

Soulires, I., Dawson, M., Samson, F., Barbeau, E. B., Sahyoun, C. P.,Strangman, G. E., . . . Mottron, L. (2009). Enhanced visual processingcontributes to matrix reasoning in autism. Human Brain Mapping, 30,40824107. doi:10.1002/hbm.20831

Wojciulik, E., & Kanwisher, N. (1999). The generality of parietal involve-ment in visual attention. Neuron, 23, 747764. doi:10.1016/S0896-6273(01)80033-7

Yi, D. J., Woodman, G. F., Widders, D., Marois, R., & Chun, M. M.(2004). Neural fate of ignored stimuli: Dissociable effects of perceptualand working memory load. Nature Neuroscience, 7, 992996. doi:10.1038/nn1294

Received April 15, 2013Revision received September 10, 2013

Accepted September 12, 2013

This

docu

men

tis

copy

right

edby

the

Am

eric

anPs

ycho

logi

calA

ssoc

iatio

nor

one

ofi

tsal

lied

publ

isher

s.Th

isar

ticle

isin

tend

edso

lely

fort

hepe

rson

aluse

oft

hein

divi

dual

use

ran

dis

not

tobe

diss

emin

ated

broa

dly.

570 SWETTENHAM ET AL.

Seeing the Unseen: Autism Involves Reduced Susceptibility to Inattentional BlindnessMethodParticipantsStimuli and Procedure

ResultsTask PerformanceAwareness ReportsDetectionIdentification

DiscussionReferences