Content uploaded by Nancy Minshew
Author content
All content in this area was uploaded by Nancy Minshew on Sep 23, 2014
Content may be subject to copyright.
Neuropsychologic functioning in autism:
Profile of a complex information processing disorder
NANCY J. MINSHEW,
1
GERALD GOLDSTEIN,
2
and DON J. SIEGEL
3
1
Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA
2
Highland Drive VA Medical Center, Pittsburgh, PA
3
Western Psychiatric Institute and Clinic, Pittsburgh, PA
(Received November 27, 1995; Revised August 20, 1996; Accepted September 13, 1996)
Abstract
Neurobehavioral theories of autism have hypothesized core deficits in sensory input or perception, basic attentional
abilities or generalized attention to extrapersonal space, anterograde memory, auditory information processing,
higher order memory abilities, conceptual reasoning abilities, executive function, control mechanisms of attention,
and higher order abilities across domains. A neuropsychologic battery designed to investigate these hypotheses was
administered to 33 rigorously diagnosed autistic individuals with IQ scores greater than 80, and 33 individually
matched normal controls. Stepwise discriminant function was used to define the profile of neuropsychologic
functioning across domains. The neuropsychologic profile in these autistic individuals was defined by impairments
in skilled motor, complex memory, complex language, and reasoning domains, and by intact or superior
performance in the attention, simple memory, simple language, and visual–spatial domains. This profile is not
consistent with mental retardation or with a general deficit syndrome, but rather with a selective impairment in
complex information processing that does not involve visual–spatial processing. This profile is not consistent with a
single primary deficit, but with a multiple primary deficit model in which the deficit pattern within and across
domains is reflective of the complexity of the information processing demands. This neuropsychologic profile is
furthermore consistent with the neurophysiologic characterization of autism as a late information processing
disorder with sparing of early information processing. (JINS, 1997, 3, 303–316)
Keywords: Autism, Neuropsychologic function, Cognitive profiles, Information processing
INTRODUCTION
Although autism is now widely accepted as being of neuro-
logic origin, the neurobehavioral basis of the clinical syn-
drome remains highly controversial, with widely disparate
views regarding the core deficit or deficits that underlie the
abnormal behavior typical of this syndrome.Aneurobehav-
ioral basis for autism first gained acceptance in the 1960s
(Rimland, 1964), and a number of neurobehavioral models
were proposed and became widely known during the 1960s
and 1970s. Some of these theories, such as those postulat-
ing core deficits in sensory perception or relating autism to
an amnesic disorder, persisted into the 1980s, but others,
such as the left hemisphere language theory, were later set
aside as a result of incompatibility with evolving research
findings demonstrating symmetric bihemispheric abnormal-
ities (Bauman & Kemper, 1985; Rumsey & Hamburger,
1988; Minshew, 1992). In the last 15 years, neurobehav-
ioral theories have existed that hypothesized primary defi-
cits in nearly every aspect of neuropsychologic functioning
as the cause of this clinical syndrome. During the early 1980s,
these models generally postulated a single primary deficit
in an aspect of information acquisition. Such models hy-
pothesized a core deficit in a basic aspect of sensory per-
ception, attention, or memory function. Although these
models may now be viewed as simplistic in light of present
knowledge about brain organization and about autism, the
clinical integrity of information acquisition in autism re-
mains a critical issue for all current neurobehavioral models
of autism, and is largely undocumented by neuropsycho-
logic data.
An additional, but little known, theory from the early
1980s hypothesized a selective impairment in auditory in-
formation processing based on the initial observations of
Reprint requests to: Nancy J. Minshew, Western Psychiatric Institute
and Clinic, 3811 O’Hara Street, Bellefield Towers, Room 430, Pittsburgh,
PA 15213.
Journal of the International Neuropsychological Society (1997), 3, 303–316.
Copyright © 1997 INS. Published by Cambridge University Press. Printed in the USA.
303
dramatic abnormalities in auditory cognitive potentials with
minimal abnormalities in visual cognitive potentials (Nov-
ick et al., 1980). This theory was not pursued beyond the
initial report, but the disparity between auditory and visual
cognitive potential abnormalities was subsequently repli-
cated by several investigators, and issues related to visual
information processing and to the involvement of the pos-
terior regions of the cerebral hemispheres persist as impor-
tant unanswered questions in autism.
Neurobehavioral theories proposed in the late 1980s and
the 1990s have hypothesized primary deficits in various
aspects of information processing. These theories have pro-
posed core deficits in higher order memory abilities, con-
ceptual reasoning, executive function, complex attentional
abilities, or higher order abilities in general. Three of these
five theories, the higher order memory, complex attention,
and executive function theories, appear to specify a very
similar neuropsychologic impairment involving executive
control over information processing, but differ largely as a
result of the anatomic localization proposed for this func-
tion.All of these theories, with the exception of the last theory
proposing multiple primary deficits in higher order abili-
ties, are single primary deficit models based on the premise
that the behavioral syndrome of autism will be unified at
the clinical level by a neuropsychologic deficit in a single
domain or sensory modality.
The primary deficits hypothesized in most of these neuro-
behavioral models were proposed either on the basis of
neurophysiologic abnormalities without clinical evidence
of a neuropsychologic deficit, or neuropsychologic data
obtained prior to the early 1980s before the importance
of excluding autistic subjects with coexisting infectious,
metabolic, and genetic disorders from research studies was
appreciated (Damasio et al., 1980; Rumsey et al., 1984; Crea-
sey et al., 1986; Minshew & Goldstein, 1993). Addition-
ally, many of these models were developed on the basis of
studies focusing exclusively on function in a single do-
main, precluding identification of potential deficits in other
domains and consideration of the significance of coexisting
deficits for the neurobehavioral formulation.
Although numerous neuropsychologic studies were com-
pleted in autism in the 1970s and 1980s (Rumsey, 1992),
these studies typically focused on a single cognitive do-
main. One of the first studies to investigate the profile of
neuropsychologic functioning across domains using a com-
prehensive test battery and strict diagnostic criteria, exclud-
ing autistic individuals with other coexisting causes of
neurologic abnormalities, involved 10 autistic men with
WAIS Verbal and Performance IQ scores above 80 and
respective mean scores of 103 and 104 (Rumsey & Ham-
burger, 1988). This study reported a profile of neuropsy-
chologic functioning in autism that was characterized by
dramatic impairments in problem solving abilities, rela-
tively intact language, memory and motor abilities, and in-
tact sensory perception and visual–spatial abilities. This
general profile was replicated in several studies (Prior &
Hoffman, 1990; Ozonoff et al., 1991), and ultimately be-
came the basis for two neurobehavioral models hypothesiz-
ing the predominance of conceptual reasoning (Rumsey &
Hamburger, 1988) and executive function (Ozonoff et al.,
1991, 1994) deficits.Although the neuropsychologic test bat-
tery used by Rumsey and Hamburger (1988) was compre-
hensive in many respects, the reliance in the motor, memory,
and language domains on a few tests predominately of ba-
sic abilities and the evidence, albeit less dramatic, of im-
pairments in domains other than problem solving suggested
the need for further characterization of the profile of neuro-
psychologic functioning in autism.
The present study was designed to provide further am-
plification of the profile of intact and deficient abilities in
autism relevant to the investigation of a number of neuro-
behavioral models in acceptance since 1980. These models
have hypothesized core neuropsychologic deficits in (1) sen-
sory input or perception (Ornitz & Ritvo, 1968; Ornitz,
1983); (2) attention to extrapersonal space (Ornitz, 1985;
Ornitz et al., 1985; Dawson & Lewy, 1989); (3) antero-
grade memory resulting in a Korsakoff’s type amnesia (De-
Long, 1978; Boucher, 1981; Bachevalier, 1991), or in higher
order memory abilities (DeLong, 1992; Bachevalier, 1994;
Bauman & Kemper, 1994); (4) auditory information process-
ing (Novick et al., 1980); (5) conceptual reasoning (Rum-
sey & Hamburger, 1988); (6) executive function (Ozonoff
et al., 1991, 1994); (7) multiple control mechanisms of at-
tention involving selective attention (Courchesne et al., 1984,
1987), attention to extrapersonal space (Courchesne et al.,
1993a; Townsend & Courchesne, 1994), and shifting atten-
tion (Courchesne et al., 1993b); and (8) higher order abili-
ties across domains with equal involvement of both the
auditory and visual modalities (Minshew et al., 1992; Min-
shew & Goldstein, 1993; Minshew et al., 1994a; Smith &
Bryson, 1994). All of the deficits proposed in the most re-
cent versions of these theories were in evidence by the time
the present study was designed and initiated, with the ex-
ception of the impairment in cross-modal shifting of atten-
tion in response to a complex contingency (Courchesne
et al., 1993b).
To test the original primacy arguments hypothesized in
these eight theories, and to provide a more detailed charac-
terization of the profile of intact and deficient neuropsycho-
logic abilities in autism, a more comprehensive test battery
than those previously published was designed so as to in-
clude tests of (1) multiple aspects of attention; (2) simple
sensory and higher cortical sensory perception; (3) elemen-
tary motor and skilled motor abilities; (4) multiple aspects
of auditory and visual memory; (5) oral and written lan-
guage functions ranging from phonetics and simple word
knowledge to semantic–pragmatic language and text com-
prehension; (6) problem solving; and (7) the rule-learning,
concept formation, and flexibility aspects of abstraction. With
regard to attention, tests were included from all four factors
in the Mirsky model (Mirsky et al., 1991) as well as tests of
other relevant aspects of attention so as to provide an as-
sessment of encoding, sustained attention, selective atten-
tion, attention to extrapersonal space, focused attention, and
304 N.J. Minshew et al.
intramodal shifting of attention. Thus, all of the attentional
impairments proposed in neurobehavioral models of autism
were considered except for the deficit in cross-modal shift-
ing of attention. In the sensory domain, tests of tactile sensory
perception were selected to address hypotheses proposing
primary impairments in sensory input or perception. In the
motor domain, consideration in test selection was given to
both elementary motor and skilled motor abilities based on
our hypothesis of generalized involvement of higher order
abilities related to neocortical function (Minshew, 1992;
Minshew et al., 1992; Minshew & Goldstein, 1993). Rele-
vant to the selective auditory information processing model,
tests presenting information in both the auditory and visual
modalities were included in all domains in which this issue
was relevant, namely, language, memory, and reasoning.
Within the language and memory domains, test selection also
reflected the need to address hypotheses implicating either
basic or higher order abilities. The assignment of language
and memory tests to simple or complex domains was based
on objective operational definitions and previous research
findings, as reviewed in Minshew & Goldstein (1993), Min-
shew et al. (1994b), and Minshew et al. (1995). In the case
of language tests, assignment to the simple or complex do-
main was based on whether the test assessed procedural or
mechanical language skills, as exemplified by phonetics,
word fluency, and spelling, or interpretative skills, as ex-
emplified by text and metaphor comprehension. In the case
of memory, classification of tests as simple or complex was
based on whether tasks were dependent on simple associa-
tive processes or required a mediational strategy to promote
remembering.Thus, associative memory tasks, such as short-
term recall of simple information, were classified as tests of
simple memory abilities, while tasks that involved consis-
tent long-term retrieval and delayed recall of complex in-
formation requiring a self-initiated organizational strategy,
were classified as involving complex memory processes
(Minshew & Goldstein, 1993). Test selection for the rea-
soning domain included problem solving tests, consistent
with the study of Rumsey and Hamburger (1988), and tests
of the rule-learning and self-initiated concept formation as-
pects of abstraction, consistent with our previous observa-
tions (Minshew et al., 1992). A visual–spatial domain was
added to the eight domains above to complete the profile of
abilities in autism. Visual–spatial abilities have long been
recognized as an area of intact function or strength in au-
tism and thus have not been a consideration in neurobehav-
ioral models, but are important in defining the profile of
neuropsychologic functioning.
METHODS
Research Participants
The participants for this study consisted of 33 individually
matched pairs of high functioning autistic and normal con-
trol adolescents and young adults between the ages of 12
and 40 years. All participants had Full Scale and Verbal IQ
scores above 80, and demonstrated sufficient cooperation
to complete testing. Demographic data for the two groups
are provided in Table 1. The absence of a significant dis-
parity between the mean Verbal and Performance IQ scores
in our autistic subjects is comparable to that obtained by
Rumsey and Hamburger (1988). The lack of a significant
Verbal–Performance IQ difference is typical of school age
and adult autistic individuals with Full Scale IQs over 80
(Mesibov, 1996; Siegel et al., 1996).
Potential autistic subjects were excluded if found to have
an associated neurologic, genetic, infectious, or metabolic
disorder, such as tuberous sclerosis, fragile-X syndrome, or
fetal cytomegalovirus infection. The diagnostic instru-
ments elicited early developmental history, and individuals
with impairments in social interaction and restricted pat-
terns of behavior, but with no clinically significant delay in
language, cognition, and adaptive behavior, were consid-
ered to have Asperger’s disorder and were excluded.
The diagnosis of autism was established through expert
clinical evaluation in accordance with accepted clinical de-
scriptions of high functioning autism (Rutter & Schopler,
1987; Minshew & Payton, 1988; Rapin, 1991; Minshew,
1996a) and two structured diagnostic instruments, the Au-
tism Diagnostic Interview (ADI: Lecouteur et al., 1989; Lord
et al., 1994) and theAutism Diagnostic Observation Sched-
ule (ADOS: Lord et al., 1989). Technician reliability in ad-
ministration and scoring of these instruments was established
through training with Dr. Catherine Lord, one of the devel-
opers of these instruments. Ongoing reliability of adminis-
tration and diagnostic accuracy were documented through
review and rescoring of the audiotaped ADI and videotaped
ADOS for all 33 autistic subjects by Dr. Lord or a member
of her research group. Eligibility for the study was depen-
dent on consistency of diagnosis across all assessments. The
autistic sample in this study was functionally and behavior-
ally comparable to the samples of Rumsey and Hamburger
(1988), Lord’shighest functioning group for theADOS (Lord
et al., 1989), and Ozonoff’s highest functioning group (Ozo-
noff et al., 1994).
Table 1. Demographic data for autistic and control subjects
Autistic Group Control Group
Variable MSDMSD
Age 20.91 9.69 21.21 9.99
Education (years) 10.54 2.90 11.24 3.07
SES
a
3.75 1.43 3.30 1.26
Verbal IQ 102.48 16.35 101.30 12.42
Performance IQ 97.45 11.19 99.09 10.59
Full Scale IQ 100.09 12.96 100.48 11.65
Male:Female 29:4 29:4
a
These values reflect middle-class status (e.g., administrative personnel,
small business owners).
Neuropsychological profile in autism 305
Neuropsychiatrically normal, medically healthy, control
participants were recruited from community volunteers as
individual age, sex, race, and IQ matches for the autistic
participants. Potential control participants were excluded if
they had a history or evidence of birth or developmental
abnormalities; acquired brain injury; poor school atten-
dance; a learning or language disability; a current or past
history of psychiatric or neurologic disorder; a medical dis-
order with implications for the central nervous system or
requiring regular medication usage; or a family history of
autism, developmental cognitive disorder,learning disability,
mood disorder, anxiety disorder, alcoholism, or other neuro-
psychiatric disorders thought to have a genetic component.
To qualify as a match for an autistic participant, the age
difference for an autism–control pair could be no more than
6 months for participants 17 years old or younger and no
more than 12 months for subjects older than 17 years. Wech-
sler Full Scale IQ scores for a pair could differ by no more
than 5 points. Socioeconomic status (SES) of the family of
origin was determined with a modification of the Hollings-
head method (Hollingshead, 1957) and matched at a group
level.
Neuropsychologic Test Battery
The tests and variables used in the present study are listed
in Table 2. To maintain a reasonable participants-to-variable
ratio, no more than five variables were considered for each
domain. The domain assignments of tests reflected the gen-
erally accepted classification systems (Lezak, 1983) and fol-
lowed previous methods (Rumsey & Hamburger, 1988;
Minshew et al., 1992). In this study, the verbal fluency test
was assigned to the simple language domain as a test of lan-
guage production ability. This assignment was consistent
with its use in the Rumsey and Hamburger (1988) study.
The Trail Making Tests A and B (Reitan & Wolfson, 1993)
were assigned to the motor domain and reasoning domain,
respectively, at the time of battery design, since perfor-
mance on Trails A more strongly reflects the psychomotor
demands of the task (Reitan & Wolfson, 1993) whereas
Trails B more strongly reflects the executive function de-
mands of the task. Delayed recall of the Rey–Osterrieth Fig-
ure (Osterrieth, 1944) was assigned to the memory domain
as a test of memory for complex visual information, con-
sistent with Lezak (1983); the means and standard devia-
tions for the copy score are provided under the visual–
spatial domain in Table 2. The group mean score of 32 for
our autistic participants on the copy score was within the
established range of normal function for individuals of this
age, IQ, and educational level (Lezak, 1995), thus docu-
menting that any impairments found in delayed recall would
be related to memory for complex visual material and not to
a visual–spatial deficit. Similarly, the Developmental Test
of Visual-Motor Integration (DVMI: Beery & Buktenica,
1989) was selected a priori for its demands on skilled mo-
tor abilities and assigned to the motor domain, although it is
also viewed as a test of visual–spatial perception.
The test battery was administered by a trained neuropsy-
chology technician working under the supervision of a clin-
ical neuropsychologist.Test sessions were adjusted in length
to the individual participant’s capabilities.
Data Analysis
The major statistical method used to demonstrate the profile
of deficits and intact abilities was Wilks’s stepwise discrim-
inant analysis. The stepwise procedure generates the dis-
crimination by sequential entering of variables. The most
discriminating variable, typically the one that produces the
highest F ratio, is entered first, followed by other variables
that combine with the first variable in a way that increases
discriminatory accuracy. Variables are entered or removed
until a preestablished tolerance test is failed, indicating that
additional entry of available variables would make no fur-
ther contribution to discriminative accuracy. The statistical
significance of the classification matrices generated by this
method was evaluated with kappa, a coefficientof agreement
for nominal scales. Relatively high kappas indicate that the
variables passing the tolerance test discriminated well be-
tween autistic subjects and controls, while low kappas in-
dicate the reverse.According to the Landis and Koch criteria
(1977), a kappa of .40 to .75 indicates fair to good agree-
ment beyond chance while a kappa less than .40 indicates
poor agreement beyond chance. Examination of variables
passing the tolerance test compared to those failing the tol-
erance test provided some additional information regarding
the nature of the abilities that best distinguished or failed to
distinguish autisticfrom control participants. In domains with
a kappa in the good agreement range, failure of a test to
pass the tolerance test did not necessarily mean that the in-
dividual test did not discriminate well between autistic par-
ticipants and controls but that, for various reasons including
high intercorrelations and multicollinearity, the test did not
add further to the accuracy of the prediction beyond that
achieved with the combination of entered variables. In such
cases, evaluation of individual test performance is neces-
sary to determine if tests failing the tolerance test did or did
not demonstrate between-group differences. For this reason,
p values based on t tests are provided in Table 2 for each of
the test variables and variables with significant p values are
marked with asterisks. Correspondingly, in domains with
poor agreement kappas, even though some tests passed the
tolerance test, they did not have sufficient discriminatory
accuracy to achieve satisfactory agreement levels.
Since stepwise analyses are exploratory in nature and can
capitalize on chance, we followed Shutty’s (1991) recom-
mendation that preliminary direct method analyses should
be performed prior to using stepwise procedures.The direct
method, involving entering all variables simultaneously, was
found to yield the same classification rates as the stepwise
method. The stepwise method was employed in the final
analyses and reported here, because of the additional infor-
mation it provides concerning the discriminatory power of
the combinations of individual tests.
306 N.J. Minshew et al.
Table 2. Psychometric data used for discriminant analysis
Autistic group Control group
Tests entered into prediction equations MSDMSDp
Attention domain
WAIS–R Digit Span 9.88 3.81 10.52 2.46 .424
Serial Digit Learning (correct responses) 16.52 8.17 17.42 7.91 .648
Continuous Performance Test (mean reaction time correct responses) 0.34 0.62 0.23 0.66 .487
Letter Cancellation (omissions) 1.09 1.63 0.45 1.00 .061
Number Cancellation (omissions) 3.27 4.03 4.39 5.38 .342
Sensory perception domain
Luria-Nebraska Tactile Scale:
Simple Touch errors 0.29 0.55 0.17 0.48 .407
Stereognosis errors 0.46 0.59 0.21 0.42 .096
Sharp–Dull Discrimination errors 0.88 0.80 0.58 0.72 .189
Position Sense errors 0.00 0.00 0.08 0.41 .328
Finger Position errors 0.67 1.27 0.46 1.02 .535
*Halstead-Reitan: Fingertip Number Writing (errors) 5.38 4.30 2.79 2.84 .019*
Motor domain
Finger Tapping—dominant hand 44.27 13.78 45.19 16.24 .805
Developmental Test of Visual–Motor Integration (total points) 15.42 32.43 22.18 31.69 .465
*Grooved Pegboard—dominant hand (time in seconds) 86.73 18.30 70.67 16.03 .000*
*Trail Making A (time in seconds) 31.52 15.81 20.45 7.99 .001*
Simple language domain
WAIS–R Vocabulary 9.45 3.02 9.70 2.26 .713
K-TEA Reading Decoding 97.48 13.60 102.79 10.19 .078
Controlled Oral Word Association (FAS) (number of words) 36.00 13.31 34.00 16.18 .586
K-TEA Spelling 102.58 16.93 100.91 11.50 .642
Woodcock Reading Mastery–Word Attack 107.24 11.55 103.52 15.53 .273
Complex language domain
*Woodcock Reading Mastery—Passage Comprehension 92.27 15.04 104.27 14.34 .002*
*K-TEA Reading Comprehension 91.36 14.43 103.06 12.45 .001*
*Test of Language Competence—Metaphoric Expression (scaled score) 6.85 3.25 9.42 3.70 .004*
*Binet Verbal Absurdities (raw score) 9.30 3.64 12.48 3.97 .001*
Token Test (number correct) 18.03 2.19 18.42 5.19 .690
Simple memory domain
Paired-Associate Learning (number correct) 42.55 23.13 48.76 24.21 .290
3 Word Short Term Memory (number of correct sequences) 3.24 3.04 2.91 3.15 .663
Maze Recall (correct/incorrect) 0.42 0.61 0.52 0.57 .534
CVLT A List—Trial 1 (number correct) 4.50 3.90 6.30 3.90 .072
Complex memory domain
Paired-Associates—Delayed Recall 16.00 7.46 17.45 6.13 .390
CVLT A List—Long Delay 7.00 5.49 9.00 5.55 .146
*Nonverbal Selective Reminding—Consistent Long-Term Retrieval 19.94 15.09 37.39 16.09 .000*
*WMS–R Logical Memory—Delayed Recall (elements) 5.58 5.79 8.45 6.02 .052*
*Rey-Osterrieth Figure—Delayed Recall (number of elements) 16.83 8.58 21.94 7.49 .012*
Reasoning domain
Trail Making B (time in seconds) 65.48 37.19 52.42 23.31 .093
Halstead Category Test (errors) 46.24 28.71 40.73 22.46 .388
Wisconsin Card Sorting Test (perseverative errors) 16.45 15.48 13.27 11.13 .342
*Binet Picture Absurdities (raw score) 20.00 11.46 27.52 6.12 .002*
*20 Questions (% constraint seeking) 35.49 23.82 56.08 14.02 .000*
Visual–spatial domain
WAIS–R Picture Completion 8.76 2.22 9.21 2.27 .415
WAIS–R Object Assembly 9.88 3.63 9.73 2.88 .852
WAIS–R Block Design 10.79 3.25 9.70 2.14 .113
Rey-Osterrieth Copy 31.30 4.80 33.09 3.75 .096
* An asterisk denotes p values that represent statistically significant differences between groups.
Neuropsychological profile in autism 307
RESULTS
The results of the Wilks’s stepwise discriminant analyses
are provided in Table 3, including test variables failing the
tolerance test, test variables passing the tolerance test in or-
der of entry, percentage of correct classifications based on
test variables passing the tolerance test, and kappa coeffi-
cients. Discriminant analyses revealed kappas in the fair to
good agreement range (.40–.75) for five domains. Autistic
participants thus were found to perform significantly more
poorly than controls in the four domains of motor function,
complex language, complex memory, and reasoning. Autis-
tic subjects were found to perform significantly better than
matched controls in the simple language domain. Kappas
were in the poor range for the attention, sensory perception,
simple memory, and visual–spatial domains,indicating poor
discrimination between the performance of autistic and con-
trol participants.
Attention domain
The kappa for this domain was in the poor range, indicating
that autistic participants were not distinguishable from con-
trols on the basis of performance in the attention domain.
Only the cancellation tests (Mesulam, 1985) passed the tol-
erance test, and performance on these tests revealed a very
low rate of omissions by both groups, with no specific pre-
dilection for errors in any of the visual quadrants. Notably,
the only tests in the attention domain to pass the tolerance
test were those with a psychomotor speed component,
whereas tests of attentional processes alone did not. Thus, a
motor impairment rather than an impairment in attention may
have been responsible for the entry of the cancellation tests
into the classification equation.
The fourth factor in the Mirsky model relates to shifting
attention, and refers to complex attentional abilities gener-
ally agreed to be of frontal origin. In the Mirsky model, this
ability is assessed by the Wisconsin Card Sorting Test
(WCST: Grant & Berg, 1948), although the WCST is more
commonly classified in the reasoning domain, as was done
in this study. Regardless of its domain classification, the per-
formance of our autistic participants on the WCST was in-
distinguishable from that of controls, and thus changes in
the domain assignment of the WCST would not have al-
tered the findings of this study with regard to attention.
Sensory perception domain
The kappa for the sensory perception domain was in the
poor range, indicating that an impairment in sensory per-
ception was not a core feature of neuropsychologic func-
tioning in autism. Two tests passed the tolerance test, both
involving higher cortical sensory perception rather than el-
ementary sensory abilities. Examination of individual test
performance revealed nearly error-free performance on all
tests of sensory perception except for a small number of
errors in both the autistic and control groups on Fingertip
Number Writing (Reitan & Wolfson, 1993). Thus, neither
the kappa for this domain nor performance on individual
Table 3. Discriminant analysis results by domain and by order of entry
Domain Tests failing tolerance test Tests passing tolerance test
%
Correct
%
Jackknife kappa
Attention Serial Digit Learning; Digit Span;
Continuous Performance
Letter Cancellation; Number Cancellation 66.7 66.7 .33
Sensory
perception
Luria-Nebraska Tactile Scale:
Touch, Position, Finger Position
and Stereognosis items
Finger Tip Writing; Luria-Nebraska
Sharp/Dull Tactile Scale item
64.6 62.5 .29
Motor Finger Tapping; Developmental
Test of Visual Motor Integration
Grooved Pegboard; Trail Making A 75.8 75.8 .52
Simple
language
WAIS–R Vocabulary K-TEA Reading Decoding;
K-TEA Spelling; WRMT–R Word
Attack; Controlled Oral Word
Association
71.2 66.7 .42
Complex
language
WRMT–R Passage Comprehension;
TLC—Metaphoric Expression
K-TEA Reading Comprehension;
Verbal Absurdities; Token Test
72.7 65.2 .45
Simple
memory
Paired Associates; 3 Word Short
Term Memory; Maze Recall
CVLT Trial 1 65.2 65.2 .30
Complex
memory
Paired Associates—Delayed; CVLT
Long Delay
NVSRT-Consistent Long Term Retrieval;
WMS–R Logical Memory—Delayed
Recall; Rey Figure—Delayed Recall
77.3 75.8 .55
Reasoning Category Test; Wisconsin Card
Sort Test
20 Questions; Picture Absurdities;
Trail Making B
75.8 72.7 .52
Visual–spatial WAIS–R Picture Completion,
Object Assembly
WAIS–R Block Design 56.1 56.1 .12
308 N.J. Minshew et al.
tests supported the presence of an impairment in sensory
input or perception as the cause of the clinical syndrome of
autism.
Motor domain
The kappa for the motor domain was in the fair to good
range of agreement, indicating that an impairment in skilled
motor abilities was a significant feature of the neuropsy-
chologic profile in autism. The discrimination was achieved
with two tests: the Grooved Pegboard (Matthews & Klove,
1964); and Trails A, both of which involve skilled motor
behavior. In contrast, there was no difference between au-
tistics and controls on the Finger Tapping Test (Reitan &
Wolfson, 1993), the only test of elementary motor skills.
Thus, the kappa for the motor domain provides support for
an impairment in skilled motor movements as a significant
feature of the neuropsychologic profile in autism.
Simple language domain
The kappa for the simple language domain was in the fair
to good range of agreement, and was exceptional in that it
reflected a superior performance by the autistic participants
relative to individually matched controls. Tests in this do-
main assessed basic language abilities, such as fluency, pho-
netics, spelling, and vocabulary. Reassignment of the verbal
fluency test to the reasoning domain as a test of executive
function would not therefore have altered the findings of
the present study in terms of support for the executive func-
tion model.
Complex language domain
The kappa for this domain was in the fair to good range of
agreement, indicating that impaired function in the Complex
Language Domain was a significant feature of the profile of
neuropsychologic functioning in these autistic individuals.
The tests entered into the classification equation included
the Reading Comprehension subtest from the Kaufman-
Test of Educational Achievement (K-TEA: Kaufman &
Kaufman, 1985), the Verbal Absurdities subtest from the
Stanford-Binet Intelligence Scale (Terman & Merrill, 1973),
and the Token Test (Boller & Vignolo, 1966). These tests
assess multiple higher order features of language, specifi-
cally text comprehension, verbal problem solving, and the
comprehension of complex grammatical constructions, re-
spectively. Of further note, tests entered into the classifica-
tion equation involved information presented in both the
auditory and visual modalities.
Two tests in this domain were not selected for the clas-
sification equation: Passage Comprehension from the Wood-
cock Reading Mastery Test–Revised (Woodcock, 1987) and
Metaphoric Expression from the Test of Language Com-
petence (Wiig & Secord, 1989). However, failure of these
variables to pass the tolerance test was likely related to multi-
collinearity, as examination of p values for these variables
(Table 2) and previous studies revealed that the autistic par-
ticipants performed significantly less well than controls on
both of these tests (Minshew et al., 1995).
Simple memory domain
The kappa for the simple memory domain was in the poor
range of agreement, indicating that an impairment in basic
associative memory abilities was not a core feature of the
neuropsychologic profile in autism. Of the five variables in
the simple memory domain, only Trial 1 of the California
Verbal Learning TestAList (CVLT: Delis et al., 1987) passed
the tolerance test. This was the only task in the simple mem-
ory domain that involved free recall of a lengthy list of ma-
terial outside the immediate span of attention and without
stimulus cues, which may account for its selection. This vari-
able was not associated with a significant intergroup differ-
ence. The tests in this domain assessed simple associative
processes in both the visual and auditory modalities. The
failure of the simple memory domain to distinguish be-
tween autistic and control participants on these tests sup-
ports the integrity of basic memory processes and the absence
of an amnesic disorder in autism.
Complex memory domain
The kappa for the complex memory domain was in the fair
to good range of agreement, indicating that impaired com-
plex memory abilities are a significant feature of the profile
of neuropsychologic functioning in autism. Tests in this cat-
egory included delayed recall and consistent long-term re-
trieval of complex visual and auditory information. Three
variables were entered into the classification equation. The
long-term retrieval measure from the Nonverbal Selective
Reminding Test (NVSR: Fletcher, 1985) was the single best
group discriminator, followed by the delayed recall score
for story recall from the Logical Memory Test (Wechsler,
1987), and the delayed recall score for the Rey-Osterrieth
Figure. All three tests require sustained retention of com-
plex information in long-term memory for which perfor-
mance would benefit from the use of mediational strategies.
Notable also is the entry into the classification equation of
tasks in both the visual and auditory modalities.
Reasoning domain
The kappa for this domain was in the fair to good range of
agreement indicating that impairments in the reasoning do-
main were a significant feature of the neuropsychologic pro-
file in autism. The three tests entered into the classification
equation were the 20 Questions procedure (Olver & Hornsby,
1966), PictureAbsurdities (Thorndike et al., 1986), and the
Trail Making Test B. The first two of these tests require self-
initiated concept formation to produce a solution, whereas
Trails B assesses the cognitive flexibility aspect of abstrac-
tion. Trails B is also commonly conceptualized as an exec-
utive function test. The selection of both concept formation
and executive function tests does not support a single pri-
Neuropsychological profile in autism 309
mary deficit in the executive function aspect of abstraction,
but rather a broader impairment in conceptual reasoning abil-
ities as proposed by Rumsey and Hamburger (1988).
Visual–spatial domain
The discriminant analysis for the visual–spatial domain re-
vealed a kappa in the poor range of agreement. Review of
the performance of the two groups on these tests revealed
that the autistic group performed as well or better than con-
trols on all three subtests, indicating that even the slight in-
crease in correct classifications above chance was related to
superior function by the autistic participants and not to a
subtle deficit in visual–spatial function. These results con-
firm the integrity of visual–spatial abilities as a feature of
the profile of neuropsychologic functioning in autism.
DISCUSSION
The study of non-mentally-retarded autistic individuals
provides the maximal opportunity for defining the impair-
ments associated with autism and demonstrating their spec-
ificity to autism rather than to the confounds associated with
mental retardation. The greater capabilities of high func-
tioning autistic individuals have been essential to the ap-
plication of more sophisticated methods for investigating
cognitive function to the study of autism. These methods
have included clinical and experimental neuropsychologic
tests, cognitive psychology tests, cognitive evoked poten-
tials, and, in the future, functional imaging technologies re-
quiring activation paradigms. The validity of the subgroup
method in all of these studies was based on the premise that
the participants have the disorder under study. Non-mentally-
retarded autistic individuals have all of the clinical features
and developmental history associated with autism (Kanner
et al., 1972), the same neuropathologic findings as have been
demonstrated in mentally retarded autistic individuals (Bau-
man & Kemper, 1994), the same imaging abnormalities
(Piven et al., 1996), and the same family history character-
istics (Piven et al., 1997).
The present study of the profile of neuropsychologic func-
tioning in non-mentally-retarded autistic adolescents and
young adults provides evidence within the same subject
sample of significant impairments in the motor, complex
language, complex memory, and reasoning domains. This
profile was further defined by intact or superior perfor-
mance in the attention, sensory perception, simple memory,
simple language, and visual–spatial domains. These results
provide empiric evidence of previously undocumented
impairments in skilled motor, complex language, and com-
plex memory abilities that redefine the profile of neuropsy-
chologic functioning reported for autism in prior studies
(Rumsey & Hamburger, 1988; Prior & Hoffman,1990; Ozo-
noff et al., 1991). The deficient and intact abilities identi-
fied in the present study define a pattern of cognitive
functioning in autism that is characterized by selective im-
pairment of certain higher order abilities, sparing or en-
hancement of simpler abilities in these same domains, and
intact basic skills.
The intact abilities identified in this study are of particular
significance in demonstrating that the neuropsychologic pro-
file in autism is not that of a general deficit syndrome or of
mental retardation, as autistic participants performed as well
or better than individually matched normal controls on many
tests. This profile is also in sharp contrast to the neuropsy-
chologic profile reported for schizophrenia, in which sub-
jects perform more poorly than controls on most tests, and
hence, are often considered to have a general deficit syn-
drome (Chapman & Chapman, 1973). Intact neuropsycho-
logic abilities may prove to be as much of a determinant of
the behavioral expression of autism as deficient skills. For
example, the focus on details and reaction to trivial envi-
ronmental changes so characteristic of autism might not be
features of this clinical syndrome if basic attention, sensory
perception, and associative memory abilities were not suf-
ficiently intact to support the awareness of details. Thus,
neuropsychologic functioning in autism is best character-
ized by a two-part model giving equal consideration to in-
tact abilities and deficits.
The results of the present study confirm and significantly
extend the definition of neuropsychologic functioning re-
ported for autismby Rumsey andHamburger(1988) and rep-
licated by others (Prior & Hoffman, 1990; Ozonoff et al.,
1991) as characterized by prominent impairment in prob-
lem solving abilities in the absence of clinically significant
impairments in sensory perception, memory, or language
ability. The profile reported by Rumsey and Hamburger is
identical to that obtained in the present study in the reason-
ing, simple memory, simple language, and sensory percep-
tion domains. The documentation by the present study of
significant impairments in domains other than reasoning re-
flects the expansion of the test battery to include tests of
higher order abilities in other domains, the separate consid-
eration of simple and complex memory and language tests,
and the utilization of a much larger sample.
Adirect examination of subject performance in the Rum-
sey and Hamburger study (1988) reveals findings essen-
tially identical to ours, including indications of the presence
of impairments in domains other than reasoning. In the Rum-
sey and Hamburgerstudy, the motor domain consisted solely
of the Grooved Pegboard Test, which revealed impaired per-
formance bilaterally by the autistic participants (p 5 .05),
but below the significance level established for the study.
This was interpreted as lack of evidence for unilateral brain
dysfunction as the cause of autism. Review of their data for
the Trail Making Test also reveals poorer performance on
Trails A (scores twice that of controls) than on Trails B
(scores 1.5 times that of controls), mirroring the findings of
the present study and our prior study (Minshew et al., 1992),
and providing further evidence of a clinically significant im-
pairment in skilled motor function as a feature of the neuro-
psychologic profile in autism. The language domain in the
Rumsey and Hamburger study was confined to tests of for-
mal language and, thus, did not find the deficits in complex
310 N.J. Minshew et al.
language reported in the present study. The memory do-
main was composed of one test of simple associative pro-
cesses, list learning on the Verbal Selective Reminding Test
(Buschke & Fuld, 1974), and a second test of complex mem-
ory abilities (recall of paragraphs and designs from the Wech-
sler Memory Scale–Revised). As in the present study, the
performance of the autistic subjects in the Rumsey and Ham-
burger study on these two tests was divergent, with im-
paired performance on the test of complex memory abilities
(p 5 .05) and intact performance on the simple memory
test (p 5 .92).
Since the Rumsey and Hamburger study, there have been
a number of studies focusing on function in individual neuro-
psychologic domains that have provided evidence of the im-
pairments reported in the present study in reasoning (Ozonoff
et al., 1994; Klinger & Dawson, 1995), complex memory
(Tager-Flusberg, 1991; Minshew & Goldstein, 1993; Min-
shew et al., 1996), complex language (Minshew et al., 1995),
and skilled motor function (Smith & Bryson, 1994; Hughes,
1996; Leary & Hill, 1996). Of these deficits, the impair-
ments in complex memory abilities and skilled motor move-
ments have been the most recent to be documented, and
remain incompletely defined. In addition, in a recent factor
analytic study of the neuropsychologic performance of 3-
to 7-year-old autistic children with Performance IQ scores
greater than 80, Rapin and colleagues described a four-
factor structure involving separate factors for simple and
complex language abilities, a factor for motor abilities, and
a factor for visual–spatial skills (Fein et al., 1996; Rapin,
1996). This factor structure is consistent with the results of
the present study.
Our findings of significant coexisting deficits in skilled
motor, complex language, complex memory, and reasoning
abilities, involvement of the auditory and visual modalities,
with intact or superior abilities in the attention, sensory per-
ception, simple language and simple memory domains, have
significant implications for neurobehavioral models pro-
posed for autism in the past 15 years. All but one of these
models have hypothesized the presence of a clinically ap-
parent deficit in a single domain or modality of neuropsy-
chologic functioning as the basis for the clinical syndrome
of autism. The argument for primacy in each of these mod-
els was based on evidence provided of the current presence
of the specified deficit in autistic children, adolescents, and
adults. Other mechanisms for primacy were often pro-
posed, but without the support of empiric data. These alter-
native arguments typically hypothesized temporal primacy,
with an onset of the proposed primary deficit in early life
preceding all other manifestations of autism, primacy at the
neurobiologic level in terms of the way in which the brain
accomplishes the involved function, or a combination of both
of these. However, neither the neurobehavioral models them-
selves, nor the present study, have data to support or refute
these alternative arguments. Thus, the major questions for
the single primary deficit models are which of the hypoth-
esized deficits can be demonstrated to be present, and of
these, which ones can feasibly produce the pattern of neuro-
psychologic deficits and intact abilities that has been de-
fined for autism in this and other studies.
Neurobehavioral models from the early 1980s postulat-
ing inconstancy of sensory input, generalized inattention,
or amnesia are clearly not supported by the findings of this
study in the attention, sensory perception, and simple mem-
ory domains. A revision of one of these early models (Or-
nitz, 1985) and one recently developed model (Courchesne
et al., 1993b;Townsend & Courchesne, 1994) have hypoth-
esized a generalized form of neglect or inattention to extra-
personal space as a primary deficit in autism. Neglect and
attention to extrapersonal space were specifically investi-
gated in the present study with the auditory, visual, and so-
matosensory double simultaneous stimulation tasks from the
Halstead-Reitan Sensory–Perceptual Examination (Reitan
& Wolfson, 1993) and two cancellation tasks, which re-
vealed the absence of evidence of unilateral or bilateral ne-
glect or inattention to extrapersonal space. A review of the
two theories proposing neglect as a central deficit reveals
that this deficit was inferred on purely theoretical grounds
(Ornitz, 1985) or on the basis of nonquantitative imaging
abnormalities of parietal cortex (Townsend & Courchesne,
1994). Thus, the models hypothesizing deficits in attention
to extrapersonal space or neglect also are not compatible
with the findings reported in this study or with the profile
of neuropsychologic functioning defined in other studies.
The most recently proposed neurobehavioral theories for
autism have hypothesized core deficits in various aspects of
information processing rather than in its acquisition. The
first of the information processing models proposed a se-
lective defect in auditory processing (Novick et al., 1980).
A selective impairment in auditory information processing
is not, however, supported by the findings of the present
study, which instead yielded evidence of comparable diffi-
culty in the processing of auditory and visual information
in multiple domains. In the complex memory and complex
language domains, the long-term retrieval score from the
Nonverbal Selective Reminding Test and the Reading Com-
prehension score from the K-TEA were the first variables
selected for the classification equation. Similarly, Picture
Absurdities, Verbal Absurdities, and Trails B were all se-
lected for the classification equation, thus demonstrating
difficulty in the analysis of both visual and auditory mate-
rial in the reasoning domain. Because Picture Absurdities
requires the formulation of a verbal response and Trails B
involves numbers and letters, it could be argued that per-
formance on these tests reflects a language impairment or a
verbal reasoning deficit, rather than difficulty analyzing vi-
sual information. However, the superior performance of our
autistic subjects in the Simple Language Domain would not
support an impairment in the capacity to formulate lan-
guage or read letters as the cause of impaired performance
on PictureAbsurdities and Trails B. In addition, other stud-
ies in autism have provided evidence in high functioning
autistic individuals of difficulty with problem solving tasks,
such as the Tower of Hanoi (Borys et al., 1982), which have
no language component (Ozonoff et al., 1991). Thus, the
Neuropsychological profile in autism 311
results of the present and other studies suggest that autistic
individuals have difficulty with the processing of informa-
tion regardless of the modality of presentation.
Two of the more recently proposed information process-
ing models have hypothesized primary deficits in concep-
tual reasoning (Rumsey & Hamburger, 1988) or in executive
function (Ozonoff et al., 1991, 1994) as a result of frontal
systems dysfunction. The findings of the present study con-
firm the presence of significant impairments in reasoning
and problem solving. However, the absence of difficulty on
the part of the autistic participants in this study on the WCST
and the selection of the 20 Questions procedure and Picture
Absurdities test as the first and second variables to be en-
tered into the classification equation does not support cog-
nitive inflexibility or inability to shift sets as the defining
feature of the reasoning impairment in autism; rather, these
data suggest that the deficit in reasoning abilities is broader
or more generalized. The presence of significant deficits in
multiple other domains does not support the primacy or pre-
dominance of reasoning deficits. The final consideration with
regard to assessing the viability of these two models is
whether the deficits in other domains can be accommo-
dated within the frontal systems localization proposed in both
models. Although complex memory and skilled motor def-
icits are compatible with a frontal systems localization, def-
icits in higher order language comprehension and reading
comprehension are not and would indicate more wide-
spread involvement beyond frontal systems.
A third model in the executive-function–reasoning cat-
egory hypothesizes a deficit in an executive function that
regulates the attachment of meaning to information during
memory and learning, and further proposes that this func-
tion resides in the hippocampus or limbic system ( DeLong,
1992; Bachevalier, 1994). This model, therefore, argues pri-
macy both at the clinical level in terms of the primacy of a
memory and learning impairment, and also at the neurobio-
logic level in terms of how the brain assigns meaning to
incoming information. The neuropsychologic findings cited
in this model as support for a deficit in the assignment of
meaning to information appear to be the same impairments
demonstrated in the present and other studies with tasks that
are conventionally assigned to the reasoning, complex mem-
ory, and complex language domains. From a clinical neuro-
psychologic perspective, the classification of all of the tests
demonstrating impairments in this study as tests of memory
and learning would defy accepted conventions for the clas-
sification of such tests and the underlying premise that these
measures reflect separate functions in the brain. The pri-
macy argument of this model lacks construct validity and is
therefore not supported at the clinical level.
The remaining single primary deficit model in autism hy-
pothesizes multiple deficits in the control mechanisms for
attention, namely in selective attention (Courchesne et al.,
1984, 1987), attention to extrapersonal space (Courchesne
et al., 1993a; Townsend & Courchesne, 1994), and shifting
attention (Courchesne et al., 1993b), and argues that cur-
rent deficits in these mechanisms for controlling the focus
of attention are the cause of the clinical manifestations of
autism. The empiric support for this model was based on
data derived from the testing of high functioning school age
and adult autistic individuals, who were thus similar in func-
tion to the autistic subjects in the present study. The empiric
support for the deficits in selective attention and in the fo-
cus of attention consisted of neurophysiologic abnormali-
ties in cognitive potentials in the presence of intact subject
performance on the attentional task (Courchesne et al., 1984,
1985, 1987) and on qualitative imaging abnormalities in-
volving volume loss in the parietal lobes (Townsend &
Courchesne, 1994), and not neuropsychologic data. The em-
piric support provided for the shifting attention deficit was
based on data from high functioning autistic adolescent and
young adult autistic subjects using a modality shift experi-
ment to a complex contingency paradigm (Courchesne
et al., 1993b). Thus, the attentional task used to demon-
strate a shifting attention deficit in autism involved a sub-
stantial information processing component in addition to the
demand to shift attention across modalities. Thus, we would
propose that the difficulty high functioning autistic individ-
uals have been reported to have on this task reflects the in-
formation processing or cognitive demands of the task rather
than the demand for a shift in attention at the perceptual
level. The present study has provided a comprehensive eval-
uation of attention assessing the four elements of attention
in the Mirsky model as well as attention to extrapersonal
space. The results of this study fail to provide evidence of a
clinically discernible deficit in encoding, sustained atten-
tion, the ability to selectively focus attention, to attend to
all quadrants of extrapersonal space, and to make cognitive
shifts in attention, in individuals who nonetheless have all
the signs and symptoms of autism as well as deficits in higher
order abilities across multiple domains. The findings of this
study fail to support the primacy arguments of this model
of clinically apparent deficits in the control of attention
as the cause of the signs and symptoms of autism, or of at-
tentional deficits as the cause of impairments in skilled mo-
tor, complex language, complex memory, and reasoning
abilities.
The findings of the present study are most consistent with
a multiple primary deficit model, as has been proposed pre-
viously by Rutter (1988) and Goodman (1989). The pres-
ence of coexisting impairments in skilled motor, complex
language, complex memory, and reasoning abilities in au-
tism suggest a central problem with the capacity for pro-
cessing the complex features of information within these
domains. Complexity as defined by the data in this study is
linked to domain, and thus to the manner in which the brain
accomplishes the processing of the most complex informa-
tion within each domain. Domains vary in and of them-
selves in terms of complexity, and the relative complexity
of the information processing demands of the various neuro-
psychologic domains may explain the preponderance of
symptomatology in autism in domains with the highest com-
plex information processing demands.Although the test bat-
tery for the present study did not include tests of theory of
312 N.J. Minshew et al.
mind or nonverbal communicative abilities (Baron-Cohen,
1995), such deficits are consistent with a complex informa-
tion processing model and have been interpreted as such by
others. Klinger and Dawson, for example, have said that the
“social impairments displayed by individuals with autism
result from an inability to process social information because
of its novel unpredictable nature” (Klinger & Dawson, 1995,
p. 120). To this, we would add that social information is ex-
ceedinglycomplex,asaresultofmultiplecompetingsources
ofinformation, and a rapid pace ofinformationpresentation.
Notable for its absence among the domains demonstrat-
ing impairments in this study was the visual–spatial do-
main. Visual–spatial ability has long been accepted to be an
area of strength in autism, as evidenced by the preservation
of function on the Performance IQ scale with declining Full
Scale IQ and by the often remarkable facility with tasks such
as puzzle assembly. Although visual–spatial tasks are com-
plex, they are spared by the faulty information processing
mechanism in autism. Thus, it appears that the brain mech-
anisms mediating the processing of visual–spatial informa-
tion are likely to be different at a neurobiologic level from
the mechanisms used for processing complex information
in other domains (Rutter, 1983). One additional possible
explanation for the concurrent involvement of social, lan-
guage, and reasoning domains and sparing of the visual–
spatial domain in autism is that the neural systems subserving
social, language, and reasoning abilities must be intercon-
nected in order for these cognitive functions to occur and
be appropriate. In contrast, the visual–spatial system can
function independent of these other neural systems without
affecting clinical competence in this domain.
The profile of neuropsychologic functioning documented
in the present study is consistent with the profile for infor-
mation acquisition and information processing defined in
autism with evoked potentials. Neurophysiologic studies in
autism in the last decade have demonstrated a profile that is
characterized by intact brainstem potentials, intact mid-
latency potentials, and abnormalities in late, endogenous
potentials (reviewed in Minshew, 1991, 1996b). This neuro-
physiologic profile provides independent corroborating ev-
idence of the integrity of information acquisition and of early
events in information processing in autism, and of the pri-
macy of impairments in late events in information process-
ing.The selective involvement of late information processing
potentials in autism and the sparing of early information pro-
cessing potentials supports the unusual constellation de-
fined in this study of deficits in the most complex abilities
within domains with preservation of simpler abilities within
these same domains.
The neuroanatomic basis of autism and of the deficit pro-
file defined in the present study is not yet known and is con-
troversial.The prevailing neurobehavioral models for autism
demonstrate wide debate with regard to how the brain ac-
complishes the identification of information as important
and arrives at a determination of the appropriate response.
Thus, a major limitation on neurologic localization in au-
tism is that it is not known how the normal brain subserves
these complex abilities, with the exception of the general
consensus that this often involves neural systems rather than
regional brain structures. Current neurobehavioral models
for autism are consistent in hypothesizing that the brain ab-
normality underlying autism is most likely to be at the neu-
ral systems level. Equally compelling evidence exists at
present implicating cerebral cortex, limbic structures, and
the cerebellum in the affected neural systems in autism, as
well as the abnormal development of neural connections be-
tween these regions (Horwitz et al., 1988; Bauman &
Kemper, 1994; Courchesne et al., 1994; Zilbovicius et al.,
1995; Minshew, 1996b; Piven et al., 1996).
In conclusion, the use of a comprehensive neuropsycho-
logic test battery has provided evidence within a single co-
hort of autistic individuals of the coexistence of deficits in
skilled motor, complex language, complex memory, and rea-
soning abilities, and involvement of both visual and audi-
tory information processing. This test battery has furthermore
provided evidence of intact or superior simpler abilities in
these same functional areas, and the integrity of information
acquisition and visual–spatial abilities. The neuropsycho-
logic profile defined in this study is not readily explainable
in terms of a single primary deficit, but is most compatible
with a multiple primary deficit syndrome resulting from a
disorder of complex information processing that spares
visual–spatial processing. This neuropsychologic pattern
is consistent with the neurophysiologic characterization of
autism as a late information processing disorder.
ACKNOWLEDGMENTS
This work was supported by National Institute of Neurologic Dis-
orders Grant NS33355 to Nancy J. Minshew and the Department
of Veterans Affairs.
REFERENCES
American Psychiatric Association. (1987). Diagnostic and statis-
tical manual of mental disorders (Rev. 3rd ed.). Washington,
DC: Author.
American Psychiatric Association. (1994). Diagnostic and statis-
tical manual of mental disorders (4th ed.). Washington, DC:
Author.
Bachevalier, J. (1991). An animal model for childhood autism:
Memory loss and socioemotional disturbances following neo-
natal damage to the limbic system in monkeys. In C.A. Tam-
minga & S.C. Schulz (Eds.), Advances in neuropsychiatry and
psychopharmacology: Vol. 1. Schizophreniaresearch (pp. 129–
140). New York: Raven Press.
Bachevalier, J. (1994). Medial temporal lobe structures and au-
tism: A review of clinical and experimental findings. Neuro-
psychologia, 32, 627–648.
Baron-Cohen, S. (1995). Mindblindness: An essay on autism and
theory of mind. Cambridge Massachusetts: MIT Press.
Bauman, M.L. & Kemper, T.L. (1994). Neuroanatomic observa-
tions of the brain in autism. In M.L. Bauman & T.L. Kemper
(Eds.), The neurobiology of autism (pp. 119–145). Baltimore:
Johns Hopkins Press.
Neuropsychological profile in autism 313
Beery, K.E. & Buktenica, N.A. (1989). Developmental Test of
Visual–Motor Integration. Odessa, FL: Psychological Assess-
ment Resources.
Benton,A.L. & Hamsher, K.deS. (1989). MultilingualAphasia Ex-
amination. Iowa City, IA: AJAAssociates.
Boller, F. & Vignolo, L.A. (1966). Latent sensory aphasia in
hemisphere-damaged patients:An experimental study with the
Token Test. Brain, 89, 815–830.
Borys, S.V., Spitz, H.W., & Dorans, B.A. (1982). Tower of Hanoi
performance of retarded young adults and nonretarded chil-
dren as a function of solution length and goal state. Journal of
Experimental Child Psychology, 33, 87–110.
Boucher, J. (1981). Immediate free recall in early childhood au-
tism. Another point of behavioral similarity with the amnesic
syndrome. British Journal of Psychology, 72, 211–215.
Buschke, H. & Fuld, P.A. (1974). Evaluating storage, retention,
and retrieval in disordered memory and learning. Neurology,
11, 1019–1025.
Chapman, L.J. & Chapman, J.P. (1973). Disordered thought in
schizophrenia. New York: Meredith Corporation.
Courchesne, E., Courchesne, R.Y., Hicks, G., & Lincoln, A.J.
(1985). Functioning of the brainstem auditorypathway innon-
retarded autistic individuals.Electroencephalography and Clin-
ical Neurophysiology, 61, 491–501.
Courchesne, E., Elmasian, R., & Yeung-Courchesne, R. (1987).
Electrophysiological correlates of cognitive processing: P3b and
Nc, basic, clinical, and developmental research. In A.M. Hal-
liday, S.R. Butler, & R. Paul(Eds.), Atextbook of clinical neuro-
physiology (pp. 645–676). New York: Wiley.
Courchesne, E., Kilman, B.A., Galambos, R., & Lincoln, A.J.
(1984). Autism: Processing of novel auditory information as-
sessed by event-related brain potentials. Electroencephalogra-
phy and Clinical Neurophysiology, 59, 238–248.
Courchesne, E., Press, G.A., &Yeung-Courchesne, R. (1993a). Pa-
rietal lobe abnormalities detected on magnetic resonance im-
ages of patients with infantile autism. American Journal of
Roentgenology, 160, 387–393.
Courchesne, E., Saitoh, O., Yeung-Courchesne, R., Press., G.A.,
Lincoln,A.J., Haas, R.H.,& Schreibman, L. (1994). Abnormal-
ity of cerebellar vermian lobules VI and VII in patients with
infantile autism: Identification of hypoplastic and hyperplastic
subgroups with MR imaging. American Journal of Roentgen-
ology, 162, 123–130.
Courchesne, E., Townsend, J.P., Akshoomoff, N.A., Yeung-Cour-
chesne, R., Press, G.A., Murakami, J.W., Lincoln, A.J., James,
H.E., Saitoh, O., Egaas, B., Haas, R.H., & Schreibman, L.
(1993b).Anew finding: Impairment in shifting attention in au-
tistic and cerebellar patients. In S.H. Broman & J. Grafman
(Eds.), Atypical deficits in developmental disorders: Implica-
tions for brain function (pp. 101–137). Hillsdale, NJ: Law-
rence Erlbaum Associates.
Creasey, J., Rumsey, J.M., Schwartz, M., Duara, R., Rapoport, J.L.,
& Rapoport, S.I. (1986). Brain morphometry in autistic men as
measured by volumetric computed tomography. Archives of Neu-
rology, 32, 669–672.
Damasio, H., Maurer, R.G., Damasio, A.R., & Chui, H.C. (1980).
Computerized tomographic scan findings in patients with au-
tistic behavior. Archives of Neurology, 37, 504–510.
Dawson, G. & Lewy, A. (1989). Arousal, attention, and the socio-
emotional impairments of individuals with autism. In G. Daw-
son (Ed.), Autism: Nature, diagnosis,and treatment (pp. 49–74).
New York: Guilford.
Delis, D.C., Kramer, J.H., Kaplan E., & Ober, B.A. (1987). Cali-
fornia Verbal Learning Test: Adult Version. San Antonio, TX:
The Psychological Corporation.
DeLong, G.R. (1978). A neuropsychological interpretation of in-
fantile autism. In M. Rutter & E. Schopler (Eds.), Autism
(pp. 207–218). New York: Plenum.
DeLong, G.R. (1992). Autism, amnesia, hippocampus, and learn-
ing. Neuroscience & Biobehavioral Reviews, 16, 63–70.
Fein, D., Dunn, M.,Allen, D.A.,Aram, D.M., Hall, N., Morris, R.,
& Wilson, B.C. (1996). Language and neuropsychological find-
ings. In I. Rapin (Ed.), Preschool children with inadequatecom-
munication: Developmental language disorder, autism, low IQ
(pp. 123–154). London: Mac Keith Press.
Fletcher, J.M. (1985). Memory for verbal and nonverbal stimuli in
learning disability subgroups: Analysis by selective remind-
ing. Journal of Experimental Child Psychology, 40, 244–259.
Goodman, R. (1989). Infantile autism:Asyndrome of multiple pri-
mary deficits? Journal of Autism and Developmental Disor-
ders, 3, 409–424.
Grant, D.A. & Berg, E.A. (1948).A behavioral analysis of the de-
gree of reinforcement and ease of shifting to new responses in
a Weigl-type card sorting problem. Journal of Experimental
Psychology, 38, 404–411.
Hollingshead, A.B. (1957). Two factor index of social position.
New Haven, CT: Author.
Horwitz, B., Rumsey, J.M., Grady, C.L., & Rapoport, S.I. (1988).
The cerebral metabolic landscape in autism: Intercorrelations
of regional glucose utilization. Archives of Neurology, 45,
749–755.
Hughes, C. (1996). Brief report: Planning problems in autism at
the level of motor control. Journal of Autism and Developmen-
tal Disorders, 26, 99–107.
Kanner, L., Rodriguez, A., & Ashenden, B. (1972). How far can
autistic children go in matters of social adaptation? Journal of
Autism and Childhood Schizophrenia, 2, 9–33.
Kaufman, A.S. & Kaufman, N.L. (1985). Kaufman Test of Educa-
tional Achievement. Circle Pines, MN: American Guidance
Service.
Klinger, L.G. & Dawson, G. (1995).A fresh look at categorization
abilities in persons with autism. In E. Schopler & G. B. Mesi-
bov (Eds.), Learning and cognition in autism (pp. 119–136).
New York: Plenum.
Landis, J.R. & Koch, G.G. (1977). The measurement of observer
agreement for categorical data. Biometrics, 33, 671–679.
Leary, M.R. & Hill, D.D. (1996). Moving on: Autism and move-
ment disturbance. Mental Retardation, 34, 39–53.
Le Couteur,A., Rutter, M., Lord, C., Rios, P., Robertson, S., Hold-
grafer, M., & McLennan, J. (1989). Autism diagnostic inter-
view:Astandardized investigator-based instrument. Journal of
Autism and Developmental Disorders, 19, 363–387.
Lezak, M.D. (1983). Neuropsychologicalassessment (2nd ed.). New
York: Oxford University Press.
Lezak, M.D. (1995). Neuropsychological assessment (3rded.). New
York: Oxford University Press.
Lord, C., Rutter, M., Goode, S., Heemsbergen,J., Jordan, H., Maw-
hood, L., & Schopler, E. (1989). Autism diagnostic observa-
tion schedule: A standardized observation of communicative
and social behavior. Journal of Autism and Developmental Dis-
orders, 19, 185–212.
Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism Diagnostic
Interview–Revised:Arevised version of a diagnostic interview
for caregivers of individuals with possible pervasive develop-
314 N.J. Minshew et al.
mental disorders. Journal of Autism and Developmental Dis-
orders, 24, 659–685.
Matthews, C.G. & Klove, H. (1964). Instruction manual for the
Adult Neuropsychology Test Battery. Madison, WI: University
of Wisconsin Medical School.
Mesibov, G. (1996). Commentary on Siegel, Minshew, and Gold-
stein [Editorial preface].Journal ofAutism and Developmental
Disorders, 26, 387–388.
Mesulam, M.N. (1985). Principles of behavioral neurology. Phil-
adelphia: F.A. Davis.
Minshew,N.J. (1991). Indicesofneuralfunctionin autism:Clinical
and biologic implications. Pediatrics, 87 (Suppl. 2), 774–780.
Minshew, N.J. (1992). Neurological localization in autism. In E.
Schopler & G. Mesibov (Eds.), High functioning individuals
with autism (pp. 65–89). New York: Plenum Press.
Minshew, N.J. (1996a).Autism. In R. D.Adams & M. Victor (Eds.),
Principles of child neurology (pp. 1713–1729). New York:
McGraw-Hill.
Minshew, N. J. (1996b). Brief report: Brain mechanisms in au-
tism: Functional and structural abnormalities. Journal of Au-
tism and Developmental Disorders, 26, 205–209.
Minshew, N.J. & Goldstein, G. (1993). Is autism an amnesic dis-
order? Evidence from the California Verbal Learning Test.
Neuropsychology, 7, 209–216.
Minshew, N.J., Goldstein, G., Muenz, L.R., & Payton, J.B. (1992).
Neuropsychological functioning in non-mentally retarded au-
tistic individuals. Journal of Clinical and Experimental Neuro-
psychology, 14, 740–761.
Minshew, N.J., Goldstein, G., & Siegel, D.J. (1995). Speech and
language in high functioning autistic individuals. Neuropsy-
chology, 9, 255–261.
Minshew, N.J., Goldstein, G., & Siegel, D.J. (1996).Complex mem-
ory impairments in autism. Journal of the International Neuro-
psychological Society, 2, 190.
Minshew, N.J., Goldstein,G., Siegel, D.J., & Nicholson, M. (1994a).
Selective pattern of cognitive functioning in autism: A func-
tion of information processing load. Biological Psychiatry, 35,
632.
Minshew, N.J., Goldstein, G., Taylor, H.G., & Siegel, D.J. (1994b).
Academic achievement in high functioning autistic individu-
als. Journal of Clinical and Experimental Neuropsychology, 16,
261–270.
Minshew, N.J. & Payton, J.B. (1988). New perspectives in autism,
Part I: The clinical spectrum of autism. Current Problems in
Pediatrics, 18, 561–610.
Mirsky,A.F.,Anthony, B.J., Duncan, C.C.,Ahearn, M.B., & Kellam,
S.G. (1991). Analysis of the elements of attention: A neuro-
psychological approach. Neuropsychology Review, 2, 109–
145.
Novick, B., Kurtzberg, D., Vaughn, H.G., Jr., & Simpson, R. (1980).
An electrophysiologic indication of auditory processing de-
fects in autism. Psychiatry Research, 3, 107–114.
Olver, R.R. & Hornsby, J.R. (1966). On equivalence. In J.S. Bruner,
R.R. Olver, P.M. Greenfield, J.R. Hornsby, H.J. Kennedy, M.
Maccoby, N. Modiano, S.A. Mosher, D.R. Olson, M.C. Potter,
L.C. Reich, & A.M. Sonstroem (Eds.), Studies in Cognitive
Growth (pp. 68–85). New York: Wiley & Sons.
Ornitz, E.M. (1983). The functional neuroanatomy of infantile au-
tism. International Journal of Neuroscience, 19, 85–124.
Ornitz, E.M. (1985). Neurophysiology of infantile autism. Jour-
nal of the American Academy of Child and Adolescent Psychi-
atry, 24, 251–262.
Ornitz, E.M.,Atwell, C.W., Kaplan,A.R., & Westlake, J.R. (1985).
Brain-stem dysfunction in autism. Archives of General Psychi-
atry, 42, 1018–1025.
Ornitz, E.M. & Ritvo, E.R. (1968). Perceptual inconstancy in early
infantile autism. Archives of General Psychiatry, 18, 76–98.
Osterrieth, P.A. (1944). Le test de copie d’une figure complexe
[The Complex Figure Copy Test]. Archives de Psychologie, 30,
206–356.
Ozonoff, S., Pennington, B.F., & Rogers, S.J. (1991). Executive
function deficits in high-functioning autistic individuals: Re-
lationship to theory of mind. Journal of Child Psychology and
Psychiatry, 32, 1081–1105.
Ozonoff, S., Strayer, D.L., McMahon, W.M., & Filloux, F. (1994).
Executive function abilities in autism and tourette syndrome:
An information processing approach. Journalof ChildPsychol-
ogy and Psychiatry, 35, 1015–1032.
Piven, J., Arndt, S., Bailey, J., & Andreasen, N. (1996). Regional
brain enlargement in autism: A magnetic resonance imaging
study. Journal of the American Academy of Child and Adoles-
cent Psychiatry, 35, 1–7.
Piven, J., Palmer, P., Jacobi, D., Childress, D., &Arndt, S. (1997).
The broader autism phenotype: Evidence from a family history
study of multiple-incidence autism families. American Journal
of Psychiatry, 154, 185–190.
Prior, M. & Hoffmann, W. (1990). Brief report: Neuropsycholog-
ical testing of autistic children through an exploration withfron-
tal lobe tests. Journal of Autism and Developmental Disorders,
20, 581–590.
Rapin, I. (1991). Autistic children: Diagnosis and clinical fea-
tures. Pediatrics, 87 (Suppl. 2), 751–760.
Rapin, I. (1996). Preschool children with inadequate communica-
tion: Developmental language disorder, autism, low IQ. Lon-
don: Mac Keith Press.
Reitan, R.M. & Wolfson, D. (1993). Halstead-Reitan Neuropsy-
chological Test Battery: Theory and clinical interpretation.Tuc-
son, AZ: Neuropsychology Press.
Rimland, B. (1964). Infantile autism. NewYork: Appleton-Century-
Crofts.
Rumsey, J.M. (1992). Neuropsychological studies of high-level
autism. In E. Schopler & G.B. Mesibov (Eds.), High function-
ing individuals with autism (pp. 41–64). New York: Plenum
Press.
Rumsey, J.M., Grimes, A.M., Pikus, A.M., Duara, R., & Ismonds,
D.R. (1984).Auditory brainstem responses in pervasive devel-
opmental disorders. Biological Psychiatry, 19, 1403–1418.
Rumsey, J.M. & Hamburger, S.D. (1988). Neuropsychological find-
ings in high-functioning men with infantile autism, residual state.
Journal of Clinical and Experimental Neuropsychology, 10,
201–221.
Rutter, M. (1983). Cognitive deficits in the pathogenesis of
autism. Journal of Child Psychology and Psychiatry, 4, 513–
531.
Rutter, M. (1988). Biological basis of autism: Implications for in-
tervention. In F.J. Menolascino & J.A. Stark (Eds.), Preventive
and curative intervention in mental retardation (pp. 265–294).
Baltimore: Brookes.
Rutter, M. & Schopler, E. (1987). Autism and pervasive develop-
mental disorders: Concepts and diagnostic issues. Journal of
Autism and Developmental Disorders, 2, 159–186.
Shutty, M., Jr. (1991). Guidelines for presenting multivariate sta-
tistical analysis in rehabilitation psychology. Rehabilitation Psy-
chology, 36, 133–136.
Neuropsychological profile in autism 315
Siegel, D.J., Minshew, N.J., & Goldstein, G. (1996). Wechsler IQ
profiles in diagnosis of high-functioning autism. Journal of Au-
tism and Developmental Disorders, 26, 389–406.
Smith, I.M. & Bryson, S.E. (1994). Imitation and action in autism:
A critical review. Psychological Bulletin, 116, 259–273.
Tager-Flusberg, H. (1991). Semantic processing in the free recall
of autistic children: Further evidence of a cognitive deficit. In
G. Dawson (Ed.), Autism: Nature, diagnosis, and treatment
(pp. 92–109). New York: Guilford.
Terman, L.M. & Merrill, M.A. (1973). Stanford-Binet Intelligence
Scale. Manual for the Third Revision, Form L-M. Boston:
Houghton-Mifflin.
Thorndike, R.L., Hagen, E.P., & Sattler, J. M. (1986). Stanford-
BinetIntelligenceScale(4thed.).Chicago:RiversidePublishing.
Townsend, J. & Courchesne, E. (1994). Parietal damage and nar-
row “spotlight” spatial attention. Journal of Cognitive Neuro-
science, 6, 218–230.
Wechsler, D. (1987). Wechsler Memory Scale–Revised. San Anto-
nio, TX: The Psychological Corporation.
Wiig, E. H. & Secord, W. (1989). Test of Language Competence–
Expanded Edition. New York: The Psychological Corporation.
Woodcock, R.W. (1987). Woodcock Reading Mastery Tests–
Revised. Circle Pines, MN: American Guidance Service.
Zilbovicius, M., Garreau, B., Samson, Y., Remy, P., Barthelemy,
C., Syrota,A., & Lelord, G. (1995). Delayedmaturation of fron-
tal cortex in childhood autism. American Journal of Psychia-
try, 152, 248–252.
316 N.J. Minshew et al.
