Eagle-Eyed Visual Acuity: An Experimental
Investigation of Enhanced Perception in Autism
Emma Ashwin, Chris Ashwin, Danielle Rhydderch, Jessica Howells, and Simon Baron-Cohen
the first reports of the condition. Over time, empirical evidence has supported the notion that those with ASC have superior visual abilities
compared with control subjects. However, it remains unclear whether these abilities are specifically the result of differences in sensory
thresholds (low-level processing), rather than higher-level cognitive processes.
optometric test, the Freiburg Visual Acuity and Contrast Test, to investigate basic low-level visual acuity.
Results: Individuals with ASC have significantly better visual acuity (20:7) compared with control subjects (20:13)—acuity so superior that
it lies in the region reported for birds of prey.
and that basic standardized tests of sensory thresholds may inform causal theories of ASC.
Key Words: Autism, enhanced perception, optometry, sensory hy-
persensitivity, sensory threshold, visual acuity
interests and repetitive behavior (1). Although these core deficits
are necessary for a diagnosis of ASC, reflecting a focus on social
impairments, first reports of the condition by Kanner (2) in the
1940s also noted that individuals with ASC show enhanced
perception of details. Individuals with ASC often report sensory
abnormalities (3,4), and one recent study by Leekam and McGeer
(5) found that 90% of the autistic children whom they tested
showed sensory abnormalities according to the Diagnostic Inter-
view for Social and Communication Disorder. Sensory hypersen-
sitivity previously formed part of the DSM-III diagnostic criteria of
autism (6), although it was excluded from the most recent
diagnostic manual, DSM-IV (1).
Sensory abnormalities may underpin a number of behaviors
such as stereotypies, aversions or preferences for certain stimuli
(7), as well as resistance to change and need for sameness,
narrow interests and obsessions, and even hypersystemizing (8).
Thus research in the field of ASC should aim to investigate
differences in sensory perception and their origins to better
understand the etiology of and possible interventions for ASC.
Different decades have seen varied interest in empirical
sensory research in ASC. In the 1950s, there was little sensory
research in ASC, with more emphasis on psychodynamic theories
(9). The 1960s and 1970s saw a focus on relatively high-level
aspects of perception such as coding of meaning, pattern-
recognition, and cognitive integration (10,11). Over the last 3
decades, research has been predominantly concerned with as-
sessing the proximal causes of social dysfunction, such as face
and voice processing (12,13), social-cognitive causes such as
utism spectrum conditions (ASC) are neurodevelopmen-
tal and are diagnosed on the basis of difficulties with
social interaction and communication, along with narrow
mindblindness (14), or high-level cognitive abnormalities such as
executive dysfunction (15). Only recently has empirical sensory
research received renewed interest, with studies of perceptual
Enhanced visual perception in ASC was first illustrated em-
pirically by the Block Design and Embedded Figures subtests of
the Wechsler intelligence scales, on which individuals with ASC
score significantly higher, relative to performance on other
subtests and to age-matched typical control subjects (18–20).
These tasks require the identification of a target feature in the
presence of complex distractors. The higher scores of individuals
with ASC suggests that either their visual perception at the local
level is superior to control subjects or that processing at the
global level is impaired, causing less interference and resulting in
faster and more accurate performance in tests of perception at
the local level. Plaisted, O’Riordan, and Baron-Cohen (16) also
found that the superior primary visuospatial perceptual skills in
ASC were illustrated by the findings of faster reaction times and
greater accuracy in conjunctive search tasks in children with ASC.
Systemizing is the ability to predict lawful events, changes, and
patterns in stimuli, and findings of “hypersystemizing” in ASC (8)
may be secondary to an underlying superiority in attention to
Despite the recent resurgence in interest in sensory percep-
tion research in ASC, it is surprising that no studies to date have
used basic and standardized optometric tests of visual acuity
(threshold). Visual acuity can be measured in the general popu-
lation using various standardized methods. A test commonly
used by optometrists is the Snellen chart (21), which comprises
11 lines of letters of different sizes that the patient is asked to read
aloud from the top. Subsequent rows have increasing numbers of
letters at decreasing size and the smallest row a patient can
identify indicates his or her visual acuity, expressed in the form
20:X, where “average” vision is denoted as 20:20, or the detail
one sees accurately from 20 feet away. Poorer vision would be in
the region of 20:40. This equates to being able to see detail
accurately from 20 feet away that an average individual could see
accurately from 40 feet away. Acuity of 20:10 represents excellent
vision and is generally thought to be the upper limit of typical
human visual acuity.
Although Snellen charts are still the most widely used mea-
From the Autism Research Centre, Department of Psychiatry, University of
Cambridge, Douglas House, Cambridge, United Kingdom.
Address reprint requests to Emma Ashwin, M.A., Autism Research Centre,
University of Cambridge Douglas House, 18b Trumpington Road, Cam-
bridge, CB2 8AH, United Kingdom; E-mail: email@example.com.
Received March 23, 2008; revised June 2, 2008; accepted June 14, 2008.
BIOL PSYCHIATRY 2008;xx:xxx
© 2008 Society of Biological Psychiatry
ARTICLE IN PRESS
sure of visual acuity, they have received criticism because of
other factors that may influence results (22). For example, the
number of letters increases while their size decreases. This
introduces two variables, rather than just one, so that one cannot
dissect whether difficulties are due to one factor or the other.
Furthermore, some studies suggest that crowding of letters in
lower rows makes them inherently more difficult to read (23).
Also, because the charts are freely available, some people may
simply memorize the Snellen chart before being tested with it to
give the impression that their vision is better than it is. Although
this is unlikely to be a common source of error, it is yet another
shortcoming of the test. A final issue is that there are fairly large,
and uneven, jumps in acuity level between the rows.
Subsequent visual acuity tests have sought to remedy such
issues so that they can provide a more accurate measure of an
individual’s visual abilities. One example of such a test is the
Freiburg Visual Acuity and Contrast Test (FrACT), which has
shown high test-retest reliability (24). Here we report the first
study to use this measure in ASC not only to fill a gap in vision
science but also to test the possibility that higher-order cognitive
differences (e.g., in visual discrimination, attention to detail, and
systemizing) may be secondary to lower low-level visual thresh-
Methods and Materials
Thirty participants were invited to our laboratory in Cam-
bridge to take part in the experiment. The ASC group comprised
n ? 15 adult males, 8 of whom were diagnosed with high-
functioning autism (HFA) and 7 with Asperger syndrome (AS).
All the participants with ASC were recruited from a volunteer
database (http://www.autismresearchcentre.com) and had been
diagnosed according to international criteria (1) by a professional
clinician in a recognized center. Each participant was asked to
provide documentary evidence of diagnosis to take part. Fifteen
adult males from the general population with no prior history of
any psychiatric diagnosis were recruited through local advertise-
ments and served as a typical control group. All participants were
screened for any preexisting optical conditions and all had
normal or corrected vision.
Participants were administered the Freiburg Visual Acuity and
Contrast Test (FrACT), a standardized optometric test that uses
the Landholt-Coptotype (24)1. The gaps in the C-shape range
from .4 mm to 25 mm and appear in one of four positions: up,
down, left, or right. Participants sat at a fixed distance of 60 cm
from the computer screen and were instructed to identify the
location of the “missing” part of the C-shaped stimulus by
selecting one of the 4 arrow keys of the keyboard, according to
the position of the gap (top, bottom, left, right).
The C-shape started at a fixed size and decreased in size with
each correct answer. After each incorrect response, the figure
appeared at a slightly larger size in the subsequent trials until a
correct response was recorded (method of limits approach).
Participants had 3 sec to respond on each of the 150 trials, which
equates to a maximum test time of just under 6 min. The protocol
required that the task should be stopped if any participant
appeared to be suffering fatigue or lack of concentration, and the
task program was set to abort if they did not respond on three or
more consecutive trials. Participants were encouraged to go as
quickly and accurately through the test as possible. The results
generated a Snellen decimal, a measure of visual acuity, in
which a value of 1.00 represents “average” 20:20 vision.
Snellen values above 1.00 represent increasingly accurate
vision (20:10 vision ? 2.00), and values below 1.00 represent
worse vision (20:40 vision ? .50).
All the participants were also administered the Wechsler
Abbreviated Scales of Intelligence (25) to determine IQ score and
the ASC group was also asked to complete the Autism Spectrum
Quotient (AQ) (26) as a validation check on diagnosis.
All participants were able to complete the task and were
included in the final analyses. Kolmogorov-Smirnov tests indi-
cated that none of the variable ranges were significantly skewed.
Table 1 shows descriptive characteristics for the control and ASC
1The visual acuity task was part of a set of tasks that took place over one
and a half hours, but task order was randomized and matched for the
control and ASC groups to account for any order effects.
Figure 1. Boxplot showing the data for the autism spectrum conditions
(ASC) and control groups.
Table 1. Descriptive Characteristics of the Control and Autism Spectrum Conditions (ASC) Groups
Control Group (n ? 15) ASC Group (n ? 15)
VariableMeanSD Range MeanSDRange Cohen’s d
2 BIOL PSYCHIATRY 2008;xx:xxx
E. Ashwin et al.
ARTICLE IN PRESS
groups. There were no significant differences between the ASC
and control groups for full-scale IQ [t(30) ? .28; p ? .42] or for
verbal IQ [t(30) ? .31; p ? .79] or for performance IQ [t(30) ?
1.10; p ? .12]. There were also no differences between the
groups for socioeconomic status (both groups came from a mix
of occupations) or educational level (both groups had 60% with
university level education and were matched for qualifications
according to the following 5-point scale: 1 ? no formal qualifi-
cations, 2 ? “O” Level/GCSE or equivalent, 3 ? “A” Level, HND or
vocational qualification, 4 ? university degree, 5 ? postgradu-
ate qualification). The groups were also matched for handed-
ness (all right-handed according to self-description) and for
numbers of normal versus corrected vision (6/15 corrected in the
control group vs. 5/15 for the ASC group). The ASC group were
significantly older than the control group [t(30) ? 2.38; p ? .03;
d ? .86]. AQ scores for the ASC group (mean ? 36.6, SD ? 7.1,
82.4% scoring 32?) were similar to the findings from previously
published studies (mean AQ score ? 35.8, SD ? 6.5, 80% scoring
Results showed that the ASC group scored a mean acuity
measure of 2.79 (SD ? ? .37), which was significantly better than
the control group mean of 1.44 [SD ? ? .26; t(30) ? 4.63; p ?
.001; see Figures 1 and 2].
There were no significant correlations between visual acuity
score and any other control variables included in the study (see
Table 2). For both the control and ASC groups, there was no
significant difference in acuity scores for normal versus corrected
vision [control group: t(15) ? –.17; p ? .66; ASC group: t(15) ?
.29; p ? .87]. Additionally for the ASC group there was no
significant difference in visual acuity score according to specific
diagnosis [HFA vs. AS; t(15) ? .11; p ? .93] and no correlation
between AQ and visual acuity score (r ? .147).
The Snellen score of 2.79 for the ASC group represents acuity
2.79 times better than “average” and translates to vision of 20:7.
Expressed differently, the ASC group could discriminate the same
detail of an object at 20 feet as a person with “average” vision
would see from 7 feet away. To put this in perspective, birds of
prey have visual acuity approximately 2 times better than that of
humans (27,28), which approaches the results seen with the ASC
group. A visual acuity score of 1.44 for the control group
translates to a ratio of 20:13, which is better than “average” 20:20
vision but still lies within the typically found limits of human
vision, which can be as great as 20:10 (29). The ASC group, at
20:7, indicates superior visual acuity that lies outside this typical
The results cannot be attributed to the differences in age
between the groups because the ASC group was older than the
control group, and visual acuity declines with age (30,31) and
because the significant difference was seen even after covarying
for age. If anything, this suggests that, had the groups been age
matched, there might have been an even greater difference in
visual acuity between the ASC and control groups.
There was no significant difference in visual acuity score
between the HFA and AS participants within the ASC group and
no significant correlation between acuity and AQ scores. We did
not make any specific predictions about this before the study, but
in line with comparisons to the control group, we might have
expected there to be a positive correlation between AQ and
visual acuity scores and perhaps for the HFA subgroup to attain
higher visual acuity scores. Our results suggest that increased
visual acuity applies to individuals across the autistic spectrum.
Because the sample size is quite small, we cannot rule out the
possibility that differences would be found between subgroups
or that acuity might correlate with AQ, in a larger sample of
The results of this study call for a search for possible
mechanisms underlying enhanced perceptual functioning and
attention to detail seen in ASC. This may be at the level of neural
hyperexcitation and is likely governed by factors that give rise to
normal variation in visual acuity in typically developing individ-
uals. For example, the fovea of the human eye is responsible for
central high resolution vision, as is measured by visual acuity
tests, and research has shown there to be a direct correlation
between foveal cone cell density and the level of visual acuity
(29). However, the decline in visual acuity that occurs over the
course of normal (nonpathological) aging cannot be attributed to
cone density because numbers remain stable over the second to
ninth decades of life (32). In this case pre- and postreceptor
processes have been suggested as mediating the age-related
decline in acuity (33).
Methylation of neurons increases in aging animals and has
been shown to cause depletion of levels of dopamine (DA) (34).
In animals, the controlled depletion of retinal DA has been
correlated with a decline in visual acuity that mimics the effects
of aging in human studies (35). Human studies of neurologic
disorders have also provided evidence for the role of DA in visual
acuity. Parkinson’s disease (PD) is both age-related and associ-
s l or t noCCSA
Figure 2. Graph showing the range of data for the autism spectrum condi-
tions (ASC) and control groups.
Table 2. Group Correlations Between Visual Acuity Score and Covariates
CovariateControl GroupASC Group
Correlations are Pearson correlations and all are nonsignificant.
ASC, autism spectrum conditions.
E. Ashwin et al.
BIOL PSYCHIATRY 2008;xx:xxx 3
ARTICLE IN PRESS
ated with a relatively higher incidence of visual dysfunction (36).
Furthermore, the symptoms of PD are thought to be caused by
the degeneration of DA neurons in the substantia nigra and a
deficiency of DA in the neostriatal DA terminals (37), both of
which may be influenced by methylation processes (38). An
implication of this is therefore that regional levels of methylation,
DA neuron density, and striatal DA levels might also mediate
typical variation in visual acuity.
From these lines of evidence we may postulate that our
findings of significantly enhanced visual acuity in ASC may be
due to atypically high numbers of foveal cone cells or to
dopamine receptors at the retinal or neural level (and perhaps
increased levels of dopamine in these areas), which may be
caused by hypomethylation. Other neurotransmitters, such as
?-aminobutyric acid (GABA), that play a role in neural excitation
may also be relevant (39), especially because genes controlling
GABA have been implicated in ASC (40). However, although the
mechanisms outlined earlier have been evidenced in typically
developing populations and neurologic disorders, they remain
purely speculative in the case of increased visual acuity in ASC.
Further studies of ASC are needed to pinpoint the underlying
neurobiological mechanism that gives rise to superior visual
acuity, and because ASC is genetic (41), genes involved in
sensory neurophysiology may also play a key role.
Exactly how significantly enhanced visual acuity translates to
higher cognitive and diagnostic features of ASC needs to be
further clarified. We suggest that the results presented here
provide further evidence for enhanced local processing (42) at an
increasingly basic level. Visual acuity is defined as the spatial-
resolving capacity of the visual system, such that higher visual
acuity equates to higher spatial-frequency abilities (the ability to
detect local detail). Low-level spatial frequencies provide infor-
mation about the global properties of a stimulus (43). Faces are
made up of distinct features of high-spatial frequency but require
us to integrate features at the global level so that we extract all
the necessary information from them. Consistent with this,
event-related potential and functional magnetic resonance imag-
ing studies have shown that low-level spatial frequency is more
effective than high spatial frequency for face processing (44) and
emotion recognition (45). Individuals with ASC, who have
atypically high spatial-frequency abilities, may not be able to
“overcome” this ability to process the global information in a
dynamic stimulus. This could contribute to the face-processing
abnormalities (46) and deficits in emotional perception seen in
ASC (47–49). Finally, these findings may underlie the cognitive
strengths observed in ASC, such as superior perceptual discrim-
ination (16) and hypersystemizing abilities (8).
The use of standardized tests such as the FrACT is favorable in
terms of comparability and reliability of results. However, we
acknowledge that it does not necessarily represent the most
ecologically valid way of testing the impact of basic visual
abilities in everyday life. Future studies should use more socially
The potential significance of the results necessitates replica-
tion of the study with larger numbers of age-matched partici-
pants. It would also be prudent to extend investigation to females
and to children with ASC to test for sexually dimorphic or
developmentally sensitive differences. Such a disparity in visual
acuity in childhood, for example, would support the existence of
an innate difference in ASC that is not explained by compensa-
tory processes that may have taken place by adulthood. It will
also be important to examine sensory thresholds in other sensory
modalities in ASC to test whether the apparent hypersensitivity is
specific to vision or general to all the senses.
Sensory hypersensitivity formed part of the earlier diagnostic
criteria of autism (DSM-III), although it is not included in the
current criteria (DSM-IV). The results of this study suggest that,
with further investigation, inclusion of such criteria may be
warranted and that basic standardized tests of sensory thresholds
may inform causal theories of ASC.
The authors were supported by the Medical Research Council
(United Kingdom) during the period of this work. We thank
Professor John Mollon, Dr. Bhismadev Chakrabarti, and Teresa
Tavassoli for valuable discussions.
The authors report no biomedical financial interests or po-
tential conflicts of interest.
ual of Mental Disorders, 4th ed. Washington DC: American Psychiatric
3. Grandin T (1996). Centre for the Study of Autism. Available at: http://
www.autism.org/temple/visual.html. Accessed May 23, 2008.
4. Grandin T (2006): Thinking in pictures. New York: Vintage Press.
5. Leekam SR, McGeer V (in press). Sensory-perceptual dysfunction and
the development of autism. Trends Cogn Sci.
ual of Mental Disorders, 3rd ed. Washington, DC: American Psychiatric
7. Emmons P, Anderson L (2005). Understanding Sensory Dysfunction:
8. Baron-Cohen S (2006): The hyper-systemizing, assortative mating the-
ory of autism. Neuropsychopharmacol Biol Psychiatry 30:865–872.
9. Bettelheim B (1955). Truants from Life; The Rehabilitation of Emotionally
Disturbed Children. Glencoe, IL: Free Press.
10. Hermelin B, O’Connor N (1967): Remembering of words by psychotic
and subnormal children. Br J Psychol 58:213–218.
11. Frith U (1970). Studies in pattern detection in normal and autistic chil-
dren. II. Reproduction and production of color sequences. J Exp Child
13. Pelphrey KA, Sasson NJ, Reznick JS, Paul G, Goldman BD, Piven J (2002):
Visual scanning of faces in autism. J Autism Dev Disord 32:249–261.
14. Baron-Cohen S (1995): Mindblindness: An Essay on Autism and Theory of
Mind. Boston: MIT Press/Bradford Books.
15. Ozonoff S, Pennington BF, Rogers SJ (1991): Executive function deficits
in high-functioning autistic individuals: Relationship to theory of mind.
16. Plaisted K, O’Riordan M, Baron-Cohen S (1998): Enhanced visual search
for a conjunctive target in autism: A research note. J Child Psychol Psy-
17. Bertone A, Mottron L, Jelenic P, Faubert J (2005): Enhanced and dimin-
ished visuo-spatial information processing in autism depends on stim-
ulus complexity. Brain 128:2430–2441.
18. Jolliffe T, Baron-Cohen S (1997): Are people with autism or Asperger’s
Syndrome faster than normal on the Embedded Figures Task? J Child
19. Shah A, Frith U (1983): An islet of ability in autistic children: A research
mance on the block design task? J Child Psychol Psychiatry 34:1351–
21. Snellen H (1965). Cited in Riggs LA. Visual acuity. In: Graham CH, editor.
23. Liu L, Arditi A (2001). How crowding affects letter confusion. Optom
4 BIOL PSYCHIATRY 2008;xx:xxx
E. Ashwin et al.
ARTICLE IN PRESS
24. Bach M (1996). Freiburg Visual Acuity and Contrast Test. Available at: Download full-text
http://www.michaelbach.de/fract/index.html. Accessed May 23, 2008.
25. Wechsler D (1999): Wechsler Abbreviated Scale of Intelligence. San Anto-
nio, TX: Psychological Corporation.
autism-spectrum quotient (AQ): Evidence from Asperger syndrome/
high- functioning autism, males and females, scientists and mathema-
27. Fox R, Lehmkuhle SW, Westendorf DH (1976): Falcon visual acuity. Sci-
28. Reymond L (1987): Spatial visual acuity of the falcon, Falco berigora: A
behavioural, optical and anatomical investigation. Vision Res 27:1859–
29. Beynon J (1985). Visual acuity and the eye. Physics Educ 20:234–237.
30. McGrath C, Morrison JD (1981). The effects of age on spatial frequency
31. Lee FS, Matthews LJ, Dubno JR, Mills JH (2005). Longitudinal study of
pure-tone thresholds in older persons. Ear Hearing 26:1–11.
32. Gao H, Hollyfield JG (1992): Aging of the human retina. Differential loss
of neurons and retinal pigment epithelial cells. Invest Ophthal Vis Sci
“stable” optic neuropathy. Arch Ophthalmol 123:785–788.
34. Crews FT, Hirata F, Axelrod J (1980): Identification and properties of
methyltransferases that synthesize phosphatidylcholine in rat brain
synaptosomes. J Neurochem 34:1491–1498.
mimics some of the effects of ageing on visual function. Vision Res
36. Uc EY, Rizzo M, Anderson SW, Qian S, Rodnitzky RL, Dawson JD (2005):
Visual dysfunction in Parkinson disease without dementia. Neurology
37. Koller WC, Rueda MG (1998): Mechanism of action of dopaminergic
agents in Parkinson’s disease. Neurology 50:11–14.
38. Charlton CG, Crowell B Jr (1995): Striatal dopamine depletion, tremors,
and hypokinesia following the intracranial injection of S-adenosylme-
thionine: A possible role of hypermethylation in parkinsonism. Mol
39. Deisseroth K, Malenka RC (2005): GABA excitation in the adult brain: A
mechanism for excitation-neurogenesis coupling. Neuron 47:775–777.
40. Ma DQ, Whitehead PL, Menold MM, Martin ER, Ashley-Koch AE, Mei H,
et al. (2005): Identification of significant association and gene-gene
interaction of GABA receptor subunit genes in autism. Am J Hum Genet
41. Santangelo SL, Tsatsanis K (2005): What is known about autism: Genes,
brain, and behavior. Am J Pharmacogenomics 5:71–92.
42. Frith U (1989). Autism: Explaining the Enigma. Oxford, UK: Blackwell.
43. Badcock JC, Whitworth FA, Badcock DR, Lovegrove WJ (1990). Low-
45. Katsyri J, Saalasti S, Tiippana K, von Wendt L, Sams M (2008). Impaired
syndrome. Neuropsychologia 46:1888–1897.
46. Klin A, Jones W, Schultz R, Volkmar F, Cohen D (2002). Visual fixation
patterns during viewing of naturalistic social situations as predictors of
social competence in individuals with autism. Arch Gen Psychiatry 59:
recognition in high functioning autism. Social Neurosci 1:349–363.
48. Golan O, Baron-Cohen S (2006): Systemizing empathy: Teaching adults
plex emotions using interactive multimedia. Dev Psychopathol 18:591–
49. Golan O, Baron-Cohen S, Golan Y (2008). The “Reading the Mind in
Films” Task [Child Version]: Complex emotion and mental state recog-
nition in children with and without autism spectrum conditions [pub-
lished online ahead of print February 29]. J Autism Dev Disord.
E. Ashwin et al.
BIOL PSYCHIATRY 2008;xx:xxx 5
ARTICLE IN PRESS