Eyewitness Testimony in Autism Spectrum Disorder

ORIGINAL PAPER

Eyewitness Testimony in Autism Spectrum Disorder: A Review

Katie L. Maras • Dermot M. Bowler

Published online: 10 March 2012

� Springer Science+Business Media, LLC 2012

Abstract Autism spectrum disorder (ASD) is estimated

to affect around 1% of the population, and is characterised

by impairments in social interaction, communication, and

behavioural flexibility. A number of risk factors indicate

that individuals with ASD may become victims or wit-

nesses of crimes. In addition to their social and commu-

nication deficits, people with ASD also have very specific

memory problems, which impacts on their abilities to recall

eyewitnessed events. We begin this review with an over-

view of the memory difficulties that are experienced by

individuals with ASD, before discussing the studies that

have specifically examined eyewitness testimony in this

group and the implications for investigative practice.

Finally, we outline related areas that would be particularly

fruitful for future research to explore.

Keywords Autism spectrum disorder � Eyewitness � Memory � Suggestibility � Interviewing � Credibility

Introduction

Eyewitness testimony is central to the criminal justice

system, and often includes that given by individuals with

autism spectrum disorder (ASD). People with ASD com-

prise approximately 1% of the population (e.g. Baird et al.

2006), however, research identifying a number of ‘risk’

factors, such as social naivety, diminished insight into what

others are thinking (leading to exploitation by others) and

repetitive and stereotyped interests, suggests that they may

be over-represented within the criminal justice system as

victims, witnesses or even perpetrators of crime (e.g.

Browning and Caulfield 2011; Hall et al. 2007; Petersilia

2001; Scragg and Shah 1994; Siponmaa et al. 2001;

Woodbury-Smith et al. 2005). In addition to their poten-

tially inflated representation in the criminal justice system,

people with ASD also have rather specific memory diffi-

culties (see Boucher and Bowler 2008). Understanding

their eyewitness capabilities and how best to interview

them is, therefore, essential. This article begins by

reviewing some of the literature on memory in ASD to

consider how the memory difficulties associated with the

disorder might impact on their abilities to recall an eye-

witnessed event, before discussing the research to date that

has examined how such memory impairments actually

translate in eyewitness scenarios (relevant literature sear-

ches were performed using ISI Web of Knowledge and

PsychINFO databases, to December 2011). Finally impli-

cations for policy and future research directions are

discussed.

Memory in ASD

ASD is characterised by impairments in the areas of social

functioning and communication, and by the presence of

stereotypic and repetitive behaviours (American Psychiat-

ric Association 2000). Consistent evidence has also accu-

mulated over the last 50 years showing that individuals

with ASD experience specific difficulties with their mem-

ory, impacting on the ways in which they perceive,

understand, interpret, and reconstruct the world around

them. Some have argued that these difficulties may even

account for some of the diversity of behavioural features

K. L. Maras (&) � D. M. Bowler Autism Research Group, Department of Psychology,

City University London, Social Sciences Building,

Northampton Square, London EC1V 0HB, UK

e-mail: Katie.Maras.1@city.ac.uk

123

J Autism Dev Disord (2014) 44:2682–2697

DOI 10.1007/s10803-012-1502-3

that characterise the disorder (see Boucher and Bowler

2008). Individuals with ASD have a unique memory pro-

file, with peaks and troughs in their abilities. Some memory

processes such as cued recall (e.g. Bennetto et al. 1996),

priming (e.g. Gardiner et al. 2003), recognition (e.g.

Bowler et al. 2008; Minshew and Goldstein 1993;

Minshew and Goldstein 1993) and memory for facts (e.g.

Bowler and Gaigg 2008) are consistently reported to be

intact, whilst others such as source monitoring (e.g. Bowler

et al. 2004) episodic recollection and the recall of per-

sonally experienced events (e.g. Bowler et al. 2007; Russell

and Jarrold 1999) tend be impaired. These memory diffi-

culties provide empirical and practical motivation for

examining how eyewitnessed events are encoded, stored

and retrieved by individuals with ASD. We next briefly

consider some of the memory processes that are known to

be impaired in ASD, and how these may, theoretically,

affect their eyewitness testimony.

Episodic Memory and Personally Experienced Events

Episodic memory involves engaging in mental time travel

in order to re-experience the spatio-temporal context of the

event in question. Episodes in one’s memory are charac-

terised by the co-occurrence of elements of experience (e.g.

having dinner in a particular place with a particular friend

at a particular time), and are defined individually by the

specific combination of these attributes that are unique to

that episode. For an episode to be retrieved, its components

need to be marked in such a way that their retrieval is in a

bound unit. Early accounts of memory in ASD suggested

impaired episodic memory in the disorder (e.g. Boucher

1981; Boucher and Warrington 1976) and these accounts

still hold today (see Lind and Bowler 2008 for a review).

For example, Goddard et al. (2007) found that adults with

ASD recalled fewer specific memories from their past than

their matched comparison participants, and took signifi-

cantly longer to retrieve the ones that they could remember.

Similarly, over two experiments Bruck et al. (2007)

reported that children with ASD also recalled fewer epi-

sodes from their past, and fewer details than typically

developing children for a previously participated in staged

event.

Findings of impaired episodic recollection in ASD also

indicate that people with the disorder have particular dif-

ficulties retrieving specific events. Indeed, Bowler et al.

(2007) have shown that individuals with ASD place a

greater reliance on ‘knowing’ or semantic memory—which

is relatively unimpaired in ASD (e.g. Crane and Goddard

2008), and are less likely to experience the type of con-

scious recollection—known as autonoetic awareness—that

is the hallmark of episodic remembering (Tulving 1985).

As such, when prompted about an event that occurred in

their past people with ASD tend to report knowledge the

event, but fail to demonstrate the autonoetic awareness of

‘reliving’ it in its full spatio-temporal context in a manner

that involves the self as the centre of the experience (see

Lind and Bowler 2008 for a review). Indeed, it has been

argued that these episodic memory impairments reflect a

failure in ASD to use self-involvement to facilitate their

memory (e.g. Crane et al. 2009; Goddard et al. 2007; Klein

et al. 1999; Toichi et al. 2002). This leads to deficits in

recalling events that were personally experienced (e.g.

Hare et al. 2007). For example, in contrast to typical

individuals who are better able to recall events that were

self-performed than events that were performed by another

person, children with ASD have been shown to recall

events that they themselves performed less well than events

that they observed being performed by a peer (e.g. Boucher

and Lewis 1989; Farrant et al. 1998; Russell and Jarrold

1999; but see Lind and Bowler 2009b, and Williams and

Happé 2009). These findings suggest that if an individual

with ASD finds themselves as a participant in a crime, be it

as an active witness, victim or perpetrator, they may find it

difficult to recall what happened. Moreover, as we discuss

next, a number of facets of memory that contribute to this

episodic deficit in ASD might also specifically shape the

eyewitness testimony that they provide.

Source Monitoring

As mentioned in the previous section, episodic events

comprise a number of perceptual, temporal, spatial,

semantic and affective elements (Johnson et al. 1993).

These elements need to be linked together at encoding in

order to form a bound coherent representation that makes

that episode distinct from other episodes (Schacter et al.

1998). However, if these components are not sufficiently

bound then source monitoring failures can occur, where

one aspect or feature of the episode is retrieved but without

the context of the rest of the episode. Thus, one may recall

an element of the experience, but not which experience it

was from.

In order to recall a specific experience, one also needs to

be able to access individual elements of the episode to

trigger the broader memory (e.g. Squire 1995). Given that

individuals with ASD often perform poorly on tests of

episodic memory, it comes as no surprise that they also

show source monitoring impairments in a number of areas.

For example, they show impairments in recollecting whe-

ther they had performed an action or generated a word

themselves or whether an experimenter had performed it

(e.g. Farrant et al. 1998; Russell and Jarrold 1999) and in

recalling which of two stimuli were presented more

recently (Bennetto et al. 1996). They also make more

intrusion errors on recall trials on the California Verbal

J Autism Dev Disord (2014) 44:2682–2697 2683

123

Learning Test (Bennetto et al. 1996) and show source

monitoring failures for the format in which words were

previously presented (Bowler et al. 2004).

Based on these empirical observations of impaired

source memory in ASD one could tentatively predict that

their eyewitness testimony might be affected in a number

of ways. First, if an individual with ASD finds it difficult to

remember where or when they learnt something, they

might be more susceptible to confuse post-event details

that, for example, they heard from a co-witness or read in a

newspaper account, as being details that they actually

witnessed themselves. Second, for the same reason they

might be more suggestible to incorporate into their reports

erroneous details that are gained through leading questions

(that contain misinformation in the form of the desired

answer in the question). Third, if the witness has trouble

remembering where or when they learnt something then

recall of a specific event might be enmeshed with details

from other events. Fourth, if a witness has difficulty pin-

pointing the source of their memories they may have dif-

ficulty in recalling a specific episode of an event that has

occurred more than once or is embedded in daily activities,

such as a commute into work. Fifth, they may have diffi-

culty recalling the temporal order in which details of an

event occurred (e.g. whether the criminal act occurred

before or after the suspect had left the scene). In a criminal

case this can mean the difference between convincing

testimony versus diminished witness credibility.

Task Support Hypothesis

Despite the memory difficulties experienced by individuals

with ASD when tested using unsupported recall procedures

(e.g. Bowler et al. 2008; Smith et al. 2007), an accumu-

lating body of evidence suggests that they can perform at a

similar level to their typical counterparts if they are pro-

vided with appropriate support during the task. Bowler

et al. (2004) coined the term task support hypothesis to

account for findings from research utilising priming, rec-

ognition and cued recall paradigms showing that such

memories are, at least in individuals without accompanying

severe intellectual disability, implicitly intact in ASD (e.g.

Bennetto et al. 1996; Boucher and Warrington 1976;

Bowler et al. 1997, 2004; Minshew et al. 1992). It has been

suggested that difficulties in deploying flexible strategies to

recall details—caused by impairments in executive func-

tioning (see Hill 2004 for a review)—mean that there are

fewer strategies available to access the information nec-

essary to trigger remembering of the event (e.g. Hughes

and Russell 1993). Therefore the provision of more support

for such strategies increases remembering. These findings

are important from an eyewitness perspective because they

suggest that recall impairments in ASD are more related to

retrieval rather than encoding mechanisms, implying that

more supportive retrieval mechanisms may help witnesses

with ASD to recall more.

Memory Organisation and Relational Processing

Recent research has found that whilst individuals with ASD

demonstrate intact or even enhanced item-specific pro-

cessing, they experience difficulties in processing relations

among elements of an experience. For example, they

demonstrate difficulties in processing a stimulus in relation

to its context such as the time of day or location (Gaigg

et al. 2008), with recalling items in their correct temporal

order (e.g. Bennetto et al. 1996; Poirier et al. 2011), and

fail to spontaneously use categorical relations among items

to aid their recall (e.g. Hermelin and O’Connor 1967;

Bowler et al. 1997; Volkmar et al. 1996). Recollective

experiences require that information is encoded and stored

in relation to spatial and temporal contextual information

(e.g. Peters et al. 2009), which might explain why indi-

viduals with ASD have problems recollecting episodic

events.

Nevertheless, this item-specific and impaired relational

processing style might actually be a positive feature of the

disorder and enhance their eyewitness testimony if, for

example, they are less susceptible to ‘filling in the gaps’ in

their memory with highly plausible but inaccurate details.

Indeed, Mottron and colleagues have suggested an

enhanced perceptual processing account of ASD (e.g.

Mottron et al. 2006), whereby individuals with ASD have

enhanced low-level processing. A related account is that

individuals with ASD have weak central coherence, where

their superior focus on details is counterbalanced by a

reduced drive to extract overall meaning (see Happé and

Frith 2006 for a review). This increased perceptual

expertise might even mean superior eyewitness perfor-

mance for small but largely unrelated details that typical

individuals would simply fail to perceive (see also Loth

et al. 2008; Shah and Frith 1983). On the other hand,

findings of diminished relational processing might mean

that witnesses with ASD have difficulty comprehending

and remembering the causal chain of events and relation-

ships between persons and agents, and the order in which

these details occurred. However, as is the case for other

memory processes and in line with the task support

hypothesis, if more support is provided, individuals with

ASD can exploit the relations amongst items to enhance

their recall to a similar level as that of their typical coun-

terparts. Indeed, whilst early work demonstrated that

individuals with ASD do not make use of semantic rela-

tions among items to aid their memory recall, when cued

recall, more support for context, or superordinate category

cues are provided their recall is undiminished (e.g. Boucher

2684 J Autism Dev Disord (2014) 44:2682–2697

123

and Warrington 1976; Gaigg et al. 2008; Minshew et al.

1992). This has important implications for police inter-

viewing techniques, which we discuss later in this article.

Emotion and Memory

Individuals with ASD demonstrate marked abnormalities in

emotional behaviours and do not process emotional stimuli

such as faces and social scenes in the same way that typical

individuals do (e.g. Norbury et al. 2009; Spezio et al.

2007). It has been argued that that people with ASD are

relatively insensitive and inattentive to their social envi-

ronment because of an abnormality of the amygdala—a

limbic structure that plays a central role in responses to

affective or emotionally charged stimuli (e.g. Baron-Cohen

et al. 2000; Schultz 2005). The amygdala is involved in the

modulation of memory consolidation (e.g. Cahill and

McGaugh 1995, 1998; Canli et al. 2000) and in typical

individuals, emotionally arousing events are both better

remembered and forgotten less than neutral, non-arousing

events (Bradley et al. 1992; Burke et al. 1992; Cahill and

McGaugh 1998; Heuer and Reisberg 1990; Kensinger and

Corkin 2003). Despite the role of the amygdala in ASD,

only four empirical studies to date have specifically

examined whether arousing events are also better remem-

bered by individuals with ASD. Three of these studies have

reported reduced enhancement effects for emotionally

arousing words or visual scenes on memory in this group

(Beversdorf et al. 1998; Deruelle et al. 2008; Gaigg and

Bowler 2008), and the other has reported typical modula-

tion of arousing words to enhance recall (South et al.

2008). Given these findings, it might be tentatively pre-

dicted that individuals with ASD may not show memory

enhancement effect for, or attenuated forgetting of, emo-

tionally arousing events. Given that most criminal events

will be, in some part at least, emotionally arousing, this is

an important implication.

Implications of the Memory Profile of ASD

for Eyewitness Testimony

Taking into account the findings on memory in ASD, it is

often the case that two contrasting predictions can be made

as to how individuals with ASD will fare as eyewitnesses.

On the one hand, a number of findings would suggest that

their testimony might be less complete and less accurate

than that of their typical counterparts. Take, for example, a

personally experienced event that involves a strong social

element, and the recall of which requires understanding of

the actions that occurred between people in a specific

temporal order, in addition to being emotionally arousing.

Most ASD researchers would agree that any of these

elements could cause problems in remembering for an

individual with ASD. Indeed, memory difficulties aside,

sensory differences such as heightened sensitivity to noise,

touch and light (e.g. Crane et al. 2009; Dawson and

Watling 2000) might mean that a witness with ASD will

have difficulty screening out sensory stimuli, particularly in

new situations. Therefore if either the witnessed event itself

(at encoding) or the retrieval environment such as the

police suite is noisy, echoes, or has fluorescent or buzzing

strip lighting (as is often the case with police stations), a

witness with ASD may find it difficult to attend to the

speaker and give testimony to the best of their ability.

On the other hand, if an individual with ASD witnesses

an event as part of their obsessive interests and where the

event is non-social in nature (e.g. involving online activi-

ties such as IT fraud), with arbitrary details (as is the case

with a lot of crimes that are briefly witnessed where the

‘bigger picture’ is not always available), they may in fact

make an excellent witness, over and above that of their

typical counterparts. Similarly, if individuals with ASD

rely less on context and follow more of an item-specific

processing style they may be less likely to substitute gaps

in their memory with details that fit with their ‘schemas’

for that type of event. On the same basis they might also be

less susceptible to post-event misinformation, and if they

have a diminished theory of mind then they may not pick

up on the implicit demands of a questioner’s suggestive

questions.

Either way, the specific and distinctive memory profile

of individuals with ASD suggests that they may make a

rather different type of witness than their typical counter-

parts. Moreover, if their memories are encoded, stored and/

or retrieved in a different way from those of typical indi-

viduals, the psychological principles on which current

police interviewing techniques are based may simply not

be effective for witnesses with ASD. We now turn our

attention now to work that has specifically examined this.

A summary of these studies can be found in Table 1.

Research on Eyewitness Testimony in ASD

How well do Witnesses with ASD Recall a Previously

Witnessed Event?

Based on the studies that have explored eyewitness recall

in ASD to date, it seems that witnesses with ASD can recall

as much and/or as accurately as their typical counterparts,

if they are interviewed appropriately. McCrory et al. (2007)

used a live classroom event to compare eyewitness recall in

11–14 year-old children with ASD and their IQ-matched

typically developing counterparts. McCrory et al. reported

that whilst the children with ASD freely recalled around

a third less information than the typically developing

J Autism Dev Disord (2014) 44:2682–2697 2685

123

Table 1 Summary of studies exploring eyewitness testimony in ASD to date, including samples, measures, and main findings

Author Sample Event Interview paradigm/recall

measures

Main findings

McCrory

et al.

(2007)

24 ASD (VIQ = 103) and 27 TD

(VIQ = 102) children;

11–14 years

Live classroom

event with neutral

and socially salient

sub-scenes

Free recall followed by general

and specific questions with

misleading questions at the end

(1 day after event)

Free recall by ASD group less

complete (with fewer gist

elements), but no less accurate

than TD group

No group differences in amount of

new details elicited by

questions, and no group

differences in suggestibility

Bruck

et al.

(2007)

30 ASD (FIQ = 96) and 38 TD

(FIQ = 105); 5–10 years

Staged event (magic

show), where child

was recipient of

some of the magic

activities

Participants given true and false

reminders about event, and

leading and misleading

questions (8 days after event).

Free recall and yes/no questions

asked further 4 days later

Autobiographical memory

questionnaire also administered

in earlier session

ASD group recalled fewer details

from both staged event and

autobiographical questionnaire

No groups differences in errors or

suggestibility to false reminders,

but ASD group more likely to

assent to false control questions

about show. ASD group also less

likely to reject never-before

heard ‘silly’ items on

autobiographical questionnaire

Maras

and

Bowler

(2010)

26 ASD (VIQ = 108) and 26 TD

(VIQ = 112) adults

Video of car park

stabbing incident

viewed on a large

projector screen

Interviewed with either full

cognitive interview or a

comparison structured interview

(30–60 min after event)

No group differences in correct

details reported or accuracy

when interviewed with

Structured Interview. Cognitive

Interview failed to increase

amount of correct details

reported by ASD group, who

were less accurate than TD

group, particularly for person

and action details

Maras

and

Bowler

(2012)

28 ASD (VIQ = 111) and 28 TD

(VIQ = 110) adults

Photographs of

everyday scenes

Context reinstatement instructions

followed by free recall, either in

same or different room in which

photographs were viewed (1 h

after event)

ASD group recalled fewer correct

details and were less accurate

than TD group when

interviewed in different room,

but no group differences when

interviewed in same room. ASD

group impaired on reporting of

person and action details overall

North

et al.

(2008)

26 ASD and 27 TD adults (IQs not

stated but sample excluded

participants with IQ \70)

Gudjonsson

Suggestibility

Scales (GSS 2);

Gudjonsson

Compliance Scale

(GCS)

Free recall (immediately and

again after 1 h delay) followed

by misleading questions and

negative feedback on GSS, and

questionnaire measuring

compliance on GCS

No differences between groups on

any of GSS measures (free recall

or suggestibility)

ASD group scored as significantly

more compliant on GCS

Maras

and

Bowler

(2011)

16 ASD (VIQ = 110) and 16 TD

(VIQ = 111) adults

Slide sequence of

photographs of

bank robbery

Exposed to schema-typical and

atypical post-event

misinformation 20 min after

event. Provided written free

recall and answered questions

relating to misinformation

20 min later

Free recall by ASD group less

complete and less accurate than

TD group

Both groups reported more

schema typical than atypical

misinformation, and no group

differences in amount of

misinformation (typical or

atypical) reported

2686 J Autism Dev Disord (2014) 44:2682–2697

123

children did, they were no less accurate with regards to the

proportion of errors or incorrect details that they reported at

this stage. Nevertheless, the ASD group were significantly

less likely to mention the most salient or gist elements of

the event, indicating that they may be less aware of

information that is socially salient in the context of an

event. However, the use of guided and specific questioning

effectively reduced group differences to the extent that

both groups reported a comparable number of event and

socially salient details.

Similarly, Bruck et al. (2007) reported that children with

ASD recalled fewer details than typically developing

children in response to autobiographical (life event) ques-

tions. In a second experiment, Bruck et al. set up a staged

event in which ASD and typically developing children

participated. Again, the ASD group recalled fewer details

from the staged event than the comparison group. How-

ever, in both experiments the details that they did report

were predominantly accurate. This implies that whilst

children with ASD are more likely to either forget or fail to

retrieve memories of personally experienced events, the

details that they do report are just as accurate as those

reported by typically developing children. Findings from

these studies with children suggest that they are capable of

providing valuable eyewitness testimony, although results

from other studies with adults have reported that witnesses

with ASD make more errors than their typical counterparts

in their free-recall of a previously witnessed event (e.g.

Maras and Bowler 2011; Maras et al. 2012). It is also

important at this point to note that these findings, which are

based on research with high-functioning individuals,

should be interpreted with caution when formulating con-

clusions for the wider autism spectrum. Low-functioning

people with ASD, who have accompanying intellectual

disability, will have broader memory difficulties on top of

their ASD-specific memory impairments (Boucher et al.

2008). As a result, these individuals are likely to have

poorer memory of an event. Research has yet to specifi-

cally explore this; therefore, any conclusions that we make

about witnesses with ASD at present can apply only to

high-functioning people with the disorder.

The question arises as to how effective existing police

interviewing techniques are for interviewing witnesses with

ASD. The ‘Cognitive Interview’ (CI) is currently the most

widely recommended research-based police interviewing

technique. However, in practice police officers often feel ill

equipped or under too much time pressure to adequately

apply it (Dando et al. 2008). When it is used appropriately,

it increases the number of correct details reported without

compromising accuracy for most witnesses, including

children, the elderly and typical adults (see Memon et al.

2010 for a review) and increases the reporting of correct

details by witnesses with intellectual disabilities (e.g.

Milne et al. 1999). Despite the substantial amount of

research on the CI with numerous different populations,

only one study to date has explored how effective it is for

witnesses with ASD (Maras and Bowler 2010).

The CI is based on two basic principles of how memory

typically operates; that retrieval of an event will be

enhanced if the context experienced at recall matches that

experienced during encoding (Fisher and Geiselman 1992;

Roediger et al. 1989; Tulving and Thomson 1973), and that

memories are stored as interconnected nodes that provide

multiple retrieval routes (Tulving 1974). On the basis of

these principles the full CI was constructed to comprise

four stages: (a) context reinstatement, (b) imagery-guided

questioning, (c) change the order of recall, and (d) change

the perspective of recall. In context reinstatement witnesses

are encouraged, in a series of verbal instructions by the

interviewer, to mentally reconstruct the external (physical)

and internal (subjective) states that they experienced during

the witnessed event before freely reporting as many details

Table 1 continued

Author Sample Event Interview paradigm/recall

measures

Main findings

Maras

et al.

(2012)

19 ASD (VIQ = 109) and 19 TD

(VIQ = 109) adults

(Experiment 1); 24 ASD

(VIQ = 113) and 24 TD

(VIQ = 111) adults

(Experiment 2)

Arousing or neutral

versions of a

narrated slide

sequence (Exp. 1)

or video clip

(Exp. 2)

Written free recall and forced

choice recognition approx.

12 days later (Exp. 1) and

written free recall immediately,

1-h, and 1-day later (Exp. 2).

Physiological measures of

arousal also taken during

viewing of event

In both experiments, arousing

story versions elicited

heightened physiological

responses and attenuated

forgetting rates (more correct

details) than neutral story

versions, and did so similarly in

both groups

Overall, ASD group freely

recalled more incorrect details

(Exp. 1) and fewer correct

details (Exp. 2) than TD group

Key ASD Autism Spectrum Disorders, TD typically developed comparisons, VIQ Mean Verbal IQ, FIQ Mean Full-Scale IQ

J Autism Dev Disord (2014) 44:2682–2697 2687

123

of the event as possible. Recall of trivial or incomplete

details is encouraged (under the ‘report all’ instruction),

since important facts may be elicited that co-occurred with

seemingly unimportant events (Geiselman et al. 1986).

Context reinstatement is followed by imagery-guided

questioning, in which witnesses are asked open-ended

questions based on what they said during their first free

recall attempt. Further details are elicited by asking wit-

nesses to summon and describe mental images of the event,

for example focusing on the best image they have of the

victim in order to describe their clothing. For the change

order stage witnesses are then asked to recall the events in a

different order, for example starting with the last thing they

witnessed and working backwards in detail until they

report the first thing they witnessed. Finally, the witness is

asked to recall the event from a different perspective. For

example, from the perspective of another person or imag-

ining that they were positioned in a different location

(Fisher and Geiselman 1992). All four of these mnemonic

strategies can elicit more detailed descriptions of a recalled

event because witnesses are encouraged to access their

memory through different routes (e.g. Schank and Abelson

1977). The effectiveness of this strategy, however, depends

on how a person stores and retrieves a memory in the first-

place, and a substantial amount of evidence indicates that

individuals with ASD may do so rather differently than

typical individuals (Bowler and Gaigg 2008).

Maras and Bowler (2010) compared recall performance

of witnesses with ASD and their age- and IQ-matched

typical comparisons under either the CI, or a Structured

Interview, which had four recall attempts so that recall

could be compared to the CI, but without the CI’s cognitive

mnemonics. Encouragingly, the ASD and comparison

groups did not differ in terms of the quantity (number of

correct details) or quality (accuracy) of their reports when

interviewed with a Structured Interview. However, unlike

the comparison group, the CI failed to increase the number

of correct details that they reported, and actually made

them less accurate relative to their typical counterparts, and

this was so across all four stages of the CI. Difficulties with

the change order and change perspective stages were

expected, given that individuals with ASD have well-

documented difficulties with temporal order memory (e.g.

Bennetto et al. 1996), in adopting a frame of reference

other than their own and on spatial working memory tasks

(Minshew et al. 1999; Morris et al. 1999; Williams et al.

2005, 2006). Moreover, these last two stages of the CI are

rarely used in practice by police officers in any case

(Clarke and Milne 2001; Dando et al. 2008).

What was at first glance surprising considering the task

support hypothesis (Bowler et al. 1997, 2004), was that the

context reinstatement stage was problematic and failed to

elicit an increase in reporting of correct details. The

question then arose of whether this stage is ineffective

because people with ASD fail to encode and bind an event

with its contextual detail in the first place, or whether the

difficulties lie in retrieving it. For example, context rein-

statement is based on the exploitation of the relations

between context and event details to trigger more details

from memory, and individuals with ASD perform poorly

on relational processing tasks (Gaigg et al. 2008). How-

ever, a number of lines of evidence suggest that the

problem that individuals with ASD have with the tradi-

tional context reinstatement procedure is a retrieval one.

For example, they have difficulties with a number of the

cognitive demands of this procedure, including mental time

travel (e.g. Lind and Bowler 2008), following complex

linguistic instructions (e.g. Goldstein et al. 1994), and

integrating these verbal instructions with their visuo-spatial

memory for the event (e.g. Kana et al. 2006). Moreover,

Bowler et al. (2008) reported that whilst ASD participants

failed to make use of context to aid their memory on tests

of recall, on recognition tests they were able to utilise

context words that were presented at study to enhance their

memory performance to a similar degree as typical indi-

viduals. Taken together, these findings suggest that, by

reducing verbal-to-visual integration demands with more

visual support, individuals with ASD may be able to draw

on contextual details of an event to aid their memory.

In a follow-up study Maras and Bowler (2012) tested

this notion by interviewing witnesses with ASD with the

context reinstatement procedure either in a different room

in which they witnessed the event, or in the same room. In

line with the previous study (Maras and Bowler 2010), the

ASD group were significantly less accurate and recalled

fewer correct details than their typical counterparts when

interviewed with the context reinstatement procedure in a

different room from which they witnessed the event. When

they were interviewed with context reinstatement but back

in the same place in which they had witnessed the event,

however, their recall was enhanced to the level of their

typical counterparts, suggesting that they can utilise con-

text to facilitate their eyewitness recall if more support for

context is provided. This is consistent with an under-con-

nectivity account of ASD (Just et al. 2004), where diffi-

culties in ASD are proposed to lie in integrating

information from different domains (e.g. verbal and

visual); the traditional context reinstatement procedure

requires participants to follow a series of verbal instruc-

tions to trigger their visual memory for contextual details

of the event. Moreover, individuals with ASD are thought

to rely more on visual rather than verbal styles of pro-

cessing, and use their perceptual representation system to

facilitate their recall (see Ben Shalom 2003). This goes

some way in explaining why the verbal mental context

reinstatement instructions are problematic for people with

2688 J Autism Dev Disord (2014) 44:2682–2697

123

ASD, and has implications for enhancing the quality and

completeness of their eyewitness reports in police inter-

views. To say that returning to the scene of the crime

would enhance the recall of witnesses with ASD would be

somewhat of an overstatement, given that this conjecture is

based on one study in a laboratory, but these findings do

provide a platform for future work to build upon, which we

discuss in more detail in the future research directions

section below.

In eyewitness research, particularly that utilising the CI,

each detail that a participant reports is often broken down

into whether it pertains to a person, action, surrounding or

object. A consistent finding across Maras and Bowler

(2010) and Maras and Bowler (2012) is that participants

with ASD report fewer person and action details. This is

not surprising when one considers the social impairments

that characterise ASD, coupled with previous findings of

diminished attention to social cues by individuals with

ASD when observing social situations (e.g. Klin et al.

2002a, b; Norbury et al. 2009). ASD and comparison

participants do not, however, differ in their reporting of

details relating to surroundings and objects. These findings

are again not surprising; individuals with ASD would not

be expected to have difficulty in recalling non-social

aspects of an event, particularly given that these can be

recalled using more of a rote or item-specific strategy. The

practical implications of these findings are that when

relying on an ASD witness’s report for evidence, details

that relate to surroundings and objects are likely to be more

reliable than details that pertain to persons and actions. The

type of interview did not differentially affect the reporting

of these details in Maras and Bowler (2010, 2012), how-

ever, and future work should examine whether there are

more supportive interview techniques that can specifically

help to increase the quantity and accuracy of person and

action details.

How Suggestible are Witnesses with ASD?

Four studies to date have explored suggestibility in ASD,

two with children and two with adults. Following free

recall and specific questioning for a previously witnessed

staged event, McCrory et al. (2007) asked their child par-

ticipants a series of leading questions, each of which

entailed an incorrect assumption (e.g. ‘‘what colour was the

man’s scarf?’’ when in fact there was no scarf). McCrory

et al. found no difference between groups for suggestibility

to misleading information. That is, the use of leading

questions increased the reporting of details that did not

occur in line with the suggested answer by witnesses with

ASD, but no more than was the case for the typically

developing children. However, McCrory et al. caution

against generalising these findings regarding comparable

suggestibility between groups to having comparable com-

pliance, and highlight the importance for further research

to investigate whether individuals with ASD might be more

likely to go along with propositions, whilst not necessarily

accepting them as true.

Bruck et al. (2007) also reported no difference in sug-

gestibility between ASD and comparison groups of chil-

dren. However, in a second experiment where participants

completed an autobiographical questionnaire, Bruck et al.

included three ‘‘silly’’ items (e.g. ‘‘Have you ever helped a

lady find a monkey in the park’’). These were mixed in

with the 12 life event questions in their questionnaire in

order to ascertain that answers were reliable. As expected,

the typically developing children were less suggestible to

the silly questions than to the life event questions. The

ASD children, however, were equally as suggestible to

both types of questions. Bruck et al. have argued that

because the ASD children were only more suggestible than

the typical children for the silly questions, but not for the

12 more plausible life event questions, that this effect does

not simply reflect a greater compliance to leading or sug-

gestive questioning. Instead, it appears to reflect a constant

pattern of compliance across suggestion type, whether it is

related to what actually happened or not. Thus, whilst the

ASD and typical children were equally as suggestible to

questions that were familiar to what actually happened, the

typically developing children appeared to use their com-

plete lack of unfamiliarity to never before heard false items

to reject suggestions by the interviewer. By contrast, the

children with ASD failed to use a lack of familiarity to

identify the interviewer’s suggestions as a whole different

version of events, meaning they were as suggestible to

these questions as they were the more plausible questions.

These findings have important implications for legal

questioning in real-life cases whereby children with ASD

may be more susceptible to acquiesce to biased inter-

viewers who either do not believe the child’s version of

events, or wish to elicit an entirely different version of

events from them in order to defend or acquit a suspect.

Whereas typical individuals appear to be resistant to such

an outright change in versions of events, it is possible that

children with ASD might be more malleable in their

testimony.

In the third study to assess suggestibility in ASD, North

et al. (2008) used the Gudjonsson Suggestibility Scales

(Gudjonsson 1997) to measure suggestibility and compli-

ance in high-functioning adults with ASD. North et al.

reported that, in line with McCrory et al. (2007) and Bruck

et al. (2007), the ASD group were no more or less sug-

gestible to leading questions and negative feedback than

their typical counterparts. However, North et al. did report

that the ASD group scored higher on a compliance ques-

tionnaire than their matched comparison participants,

J Autism Dev Disord (2014) 44:2682–2697 2689

123

which indicates an increased tendency to accede to prop-

ositions put forward by another person, even though pri-

vately they disagree with them. This finding of increased

compliance in ASD is important because it suggests that in

forensic interviewing contexts witnesses or suspects with

ASD might be more prone than typical individuals to

respond compliantly to the requests and demands of the

interviewer, even if they do not actually hold this infor-

mation as being true. Moreover, heightened compliance

might also mean greater susceptibility to exploitation by

others, leading to increased victimisation and bowing to

pressure to commit offenses. This has important practical

implications and warrants further examination using more

ecologically valid compliance scenarios.

In the fourth study examining suggestibility in ASD,

Maras and Bowler (2011) explored whether high-func-

tioning individuals with ASD might actually be less sug-

gestible to schema-related misinformation effects. Because

individuals with ASD have more of a item-specific pro-

cessing style (Gaigg et al. 2008), show reduced generalised

event knowledge (Loveland and Tunali 1993) and are

impaired in their ability to spontaneously generate core

elements defining everyday events, such as going to a

restaurant or the cinema (Volden and Johnston 1999), it

was predicted that they may also be less susceptible to

incorporating schema-typical post-event misinformation

into their subsequent reports than their typical counterparts.

Participants were presented with a mock newspaper extract

about a previously witnessed slide sequence of a bank

robbery. The extract contained some inaccurate items of

misinformation that were either typical (e.g. that a cus-

tomer was forced to lie on the floor) or atypical (e.g. that

the robbers held the door open for a customer before

entering the bank) with bank robbery schema. Contrary to

predictions, ASD and comparison witnesses were equally

suggestible, and both groups incorporated more schema

typical than atypical post-event misinformation into their

subsequent reports. This suggests that high-functioning

individuals with ASD do use their existing schemas to aid

their memory, leading them to erroneously report schema-

consistent but inaccurate details. The ASD group did,

however, recall fewer details and made more errors in their

free recall than the comparison group overall.

From these four existing studies exploring eyewitness

suggestibility in ASD, one might tentatively suggest that

individuals with ASD may freely recall less information

from an event, particularly with regards to gist or social

salience (e.g. McCrory et al. 2007), but that, high-func-

tioning individuals at least, are no more or less suggestible

than their typical counterparts. Future research would be

valuable in exploring whether these findings still stand for

low-functioning individuals, and when other forms of

suggestive influences are encountered. Given the social

difficulties that characterise ASD it would be interesting to

see if they are as susceptible to co-witness conformity

effects if they discuss the event with a co-witness who

reports a slightly different version of events (e.g. Gabbert

et al. 2003).

Given North et al’s findings of heightened compliance in

ASD, it might also be interesting to explore whether wit-

nesses with ASD can be made to change their reports as

easily as their typical counterparts are under adversarial

styles of questioning, such as that used in cross-examina-

tion (e.g. Valentine and Maras 2011). There are a number

of factors that might predict that under such circumstances

individuals with ASD would be more suggestible. Execu-

tive functioning impairments and difficulties in following

complex verbal dialogue may mean difficulties for wit-

nesses with ASD in comprehending the sort of long-winded

multiple part questions with complex syntax that barristers

tend to favour, even when they are questioning witnesses

with intellectual disabilities (e.g. Kebbell et al. 2004). It

may also be difficult for individuals with ASD to com-

prehend why they are being challenged on details to which

they know the barrister already knows the answer. Even

higher-functioning individuals who have ‘bootstrapped’ a

theory of mind (e.g. Happé 1995) are likely to struggle with

double negative questions (e.g. ‘‘is it not the case that the

weapon was not visible before the attack?’’), and accusa-

tory styles of questioning for details that they have already

clarified in previous interviews and they know that the

barrister also knows. This would be an interesting area for

future research to explore.

How well do Witnesses with ASD Recall Emotionally

Arousing Events?

As noted in the memory section above, a substantial body

of research shows that ASD is characterised by difficulties

in emotional processing domains (e.g. Dawson et al. 1990;

Hobson 1991; Kamio et al. 2006; Kasari et al. 1990;

Yirmiya et al. 1992). Coupled with findings of reduced

enhancement effects for emotionally arousing words or

visual scenes on memory in ASD (Beversdorf et al. 1998;

Deruelle et al. 2008; Gaigg and Bowler 2008), this suggests

that individuals with ASD may have difficulties recalling a

previously witnessed emotionally arousing event. How-

ever, over two experiments using more dynamic eyewit-

ness stimuli (a slide sequence with accompanying narrative

in Experiment 1, and a videoed event in Experiment 2),

Maras et al. (2012) found that individuals with ASD

recalled more details from arousing versions of the events

than they did from neutral versions of the same events.

Moreover, in contrast to previous findings with word lists

(Gaigg and Bowler 2008), both ASD and comparison

groups showed attenuated forgetting of the arousing

2690 J Autism Dev Disord (2014) 44:2682–2697

123

versions over a 1-day and 1-week delay, whilst details from

the neutral versions were forgotten more over increasing

delays. Both groups also appeared to demonstrate enhanced

physiological arousal for the arousing over the neutral story

versions, suggesting that the event may have elicited an

orienting response in arousal (with a decrease in heart rate)

and indicating that arousal may typically modulate memory

for individuals with ASD for this type of event. A possible

explanation for the discrepant findings between Gaigg and

Bowler (2008) and Maras et al. (2012) is that the former

used word lists whilst the latter used more dynamic slide

sequence/video stimuli, which may have formed more of

an interesting narrative than lists of unrelated words. This

might have led to more of an orienting response (as

opposed to a defensive response) to the stimuli (see

Christianson 1992 for a review). The implication from this

is that high-functioning individuals with ASD are similarly

affected by arousing events, remembering them equally as

well as their typical counterparts and forgetting them less

than neutral events. However, this conclusion is tentative

given that it is based on the findings from two experiments

in the laboratory. Future work should explore this using

more real-life eyewitness events that are experienced in

real time rather than viewed on a video or in slides. It might

also be worthwhile for future work to vary the valence of

the arousal by comparing memory for positively versus

negatively arousing events, thus manipulating the type of

responses (i.e. orienting, which decreases heart rate com-

pared to defensive, which increases heart rate) with dif-

ferent types of eyewitness stimuli.

Implications for Practice

From the rather sparse work that has explored eyewitness

testimony in ASD to date, it seems that high-functioning

witnesses with ASD are capable of providing reliable tes-

timony and are no more suggestible than their typical

counterparts, but that the currently recommended police

interviewing technique (the CI) is unsuitable for them.

Once additional research has replicated and extended this

work, it will be important to ensure that findings appro-

priately inform investigative practice. It has been reported

that police officers often feel that they do not receive

Place this order or similar order and get an amazing discount. USE Discount code “GET20” for 20% discount