Acta Pñdiatr Suppl 445: 60±64.

2004

Memory development and intellectual disabilities

S Vicari
Ospedale Pediatrico Bambino Gesù, Scientific Institute (IRCCS), Rome, Italy

Vicari S. Memory development and intellectual disabilities. Acta Pædiatr 2004; Suppl 445: 60–64.
Stockholm. ISSN 0803-5326
Neuropsychological research has permitted different cognitive profiles among subjects with
intellectual disabilities (ID) of different etiology to be defined. For example, numerous authors
have stressed that the typical language profile for people with Down’s syndrome (DS) consists of
poor production with greater compromise of morphosyntax than of lexical abilities, but relatively
preserved comprehension. Children with Williams’ syndrome (WS) often show marked impairment
in certain visuospatial abilities (especially praxic–constructive) and relative preservation of both
productive and receptive language, at least concerning the phonological elements. These obser-
vations seem to support a theoretical approach that considers ID not as a mere slowing of normal
cognitive development, but as distinct, individual profiles that can be qualitatively specified. The
importance of this approach was shown in several recent studies of memory, especially implicit
memory in subjects with ID. Neuropsychological studies suggest insufficient development of the
mnemic function in ID at different levels of articulation. Long-term memory has been extensively
investigated in people with ID both in the explicit and in the implicit component. According to
recent studies, people with ID should show a diffuse impairment of declarative mnesic abilities
and a relative preservation of implicit memory. The focus of this study is on the characteristics of
long- and short-term memory in children with ID and, particularly, with DS and WS. The results
are relevant to knowledge on the qualitative aspects of the anomalous cognitive development in
mentally retarded people and the neurobiological substrate underlying this development.
Key words: Down’s syndrome, implicit memory, learning, Williams’ syndrome
S Vicari, Servizio di Neurologia e Riabilitazione, Ospedale Pediatrico Bambino Gesù, IRCCS,
Lungomare Guglielmo Marconi 36, Santa Marinella, I-00058 Rome, Italy (Tel. ‡39 0766
5244258, fax. ‡39 0766 5244244, e-mail. vicari@opbg.net)

Neuropsychological research has permitted different
cognitive profiles among subjects with intellectual
disabilities (ID) of different etiology to be defined.
For example, numerous authors have stressed that the
typical language profile for people with Down’s
syndrome consists of poor production with greater
compromise of morphosyntax than of lexical abilities,
but relatively preserved comprehension (1). Williams’
syndrome (WS) is another genetic condition, less
frequent but equally characterized by ID and typified
by a number of severe medical anomalies, such as facial
dysmorphology and abnormalities of the cardiovascular
system (2). Children with WS often show marked
impairment in certain visuospatial abilities (especially
praxic–constructive) and relative preservation of both
productive and receptive language, at least concerning
the phonological elements (3). Different cognitive
profiles were described in subjects with comparable
intellectual deficits or even with the same etiopatho-
logical picture (3). All of these observations seem to
support a theoretical approach that considers ID not as a
mere slowing of normal cognitive development, but as
distinct, individual profiles, that can be qualitatively
specified. In line with this theoretical point of view
(which also suggests the need for strongly individua-
lized rehabilitation treatment protocols), many recent
studies emphasized the need to define more clearly not
only the impaired cognitive abilities in each subject, but
just as importantly, the respective strengths, or rela-
tively preserved abilities in children with ID. The
importance of this approach was shown in several
recent studies of memory, especially implicit memory
in subjects with ID.
The neuropsychological studies reported in literature
suggest insufficient development of the mnesic function
in ID at different levels of articulation. With some
exceptions (e.g. children with WS) (4), multiple deficits
have been identified in short-term memory (STM)
functioning. The peripheral system of articulatory re-
iteration and the central system that direct information
processing seem to be deficient in these subjects.
Long-term memory (LTM) has been also extensively
investigated in people with ID both in the explicit and in
the implicit component. Explicit memory concerns
intentional recall or recognition of experiences or
information. Implicit memory is manifested as a
facilitation (i.e. an improvement in performance) in
perceptual, cognitive and motor tasks, without any

 2004 Taylor & Francis. ISSN 0803-5326 DOI 10.0180/08035320310021291

ACTA PÆDIATR SUPPL 445 (2004)
conscious reference to previous experiences. Explicit
memory deficits in people with ID have also been
extensively documented. According to recent studies, as
a result of this diffuse impairment of mnesic abilities,
people with ID should show a relative preservation of
implicit memory.
A recent study (5), described LTM abilities in people
with DS and in others with ID of unknown etiology,
comparing them with normal subjects of similar mental
age. The performance of the normal subjects in explicit
memory tests was significantly better than in children
with ID of unknown etiology, and the latter were better
than those with DS. However, the performances of the
three groups did not differ in an implicit memory test
(repetition priming); both ID groups performed as well
as controls matched for mental age. These results seem
to confirm a dissociation between explicit and implicit
memory in subjects with ID. However, there are many
limits in the studies reported so far on this issue: results
are often contradictory and methodological limits
include the use of populations with ID of often
undefined etiology. The selection criteria used for the
control group (chronological age, mental age) and the
limited number of tests used for evaluating especially
implicit memory (almost always visual priming tests)
are other significant methodological limits. This last
point has important implications for both theoretical
and applied issues. Specifically, if the presumed facility
demonstrated by people with ID in repetition priming
tests was confirmed, for example, in procedural learning
tests, this would suggest substantial preservation of
implicit memory functions, and thus would support the
theoretical distinction between implicit and explicit
memory. From a more applied prospective, these find-
ings would suggest the possibility of using techniques
based on automatic learning in the rehabilitation of
these subjects.
This paper presents the results of two recent studies
carried out by the author’s group in the unit in Santa
Marinella. They concern the different aspects of
implicit and explicit memory in two groups of people
with ID (DS and WS) compared with normal subjects
matched for mental age, i.e. with a comparable global
cognitive level of functioning.
The aims were, first, to verify the hypothesis that
people with ID would have impaired explicit memory
abilities compared with controls, but that the groups do
not differ significantly in implicit memory abilities; and
secondly, determine whether this profile is characteris-
tic of all ID people or, alternatively, whether distinct
profiles may be described in different etiological groups
of ID.
Material and methods
Participants
The performances of three groups of people were
Memory and intellectual disabilities 61
examined. The first consisted of 14 individuals with
free trisomy 21 DS (chronological age, mean ± SD:
21 ± 2.42 y; mental age: 6.5 ± 0.76 y). The second
consisted of 12 people with WS (chronological age:
14.7 ± 2.8 y; mental age: 6.5 ± 0.8 y). The deletion on
the chromosome 7 was confirmed in all subjects by
fluorescence in situ hybridization (FISH). The control
group comprised 32 children with normal cognitive
abilities, of comparable mental age, evaluated with the
L-M form of the Stanford-Binet Intelligence Scale. DS
and WS groups did not differ in mental age but were
significantly different in their chronological age. For
this reason, distinct control groups were identified for
each experimental group (DS and WS).
Neuropsychological tests
Consistent with the hypothesis, the neuropsychological
battery included tests for evaluating implicit memory
[(Tower of London (TOL) test, Fragmented Pictures
Test, Serial Reaction Time (SRT) test, Word Comple-
tion], episodic explicit memory tests for verbal material
(free recall of a list of unrelated words) and episodic
explicit memory tests for visual–perceptual material
(explicit recognition of material studied in the Frag-
mented Pictures Test). All tests used in the study have
been described previously (6–8).
The subjects were tested individually; administration
of the entire protocol required two sessions of approxi-
mately 1 h each, on 2 successive days.
Results
Results obtained by DS subjects and their controls in the
implicit memory tasks are reported in Table 1. DS and
normal controls did not differ on any of the tasks
considered. In particular, a similar pattern within the
two groups was observed in both the SRT and TOL
tests. Concerning the WS group, the results were quite
different (Table 2). Indeed, although WS were similar to
normal controls in the priming repetition tasks (for
verbal as well as for visual material), they failed to show
the similar pattern of normal controls in both the SRT
and TOL.
On explicit memory, DS were always poorer than
Table 1. Results obtained by Down’s syndrome subjects (DS) and
mental age normal controls (MA) on the implicit memory tasks.
Implicit memory task DS MA p-Value
Serial Reaction Time (V–IV trials) ‡102.7 ‡162.5 ns
Tower of London (II–I testing 2.8 ± 4.5 3.2 ± 1.9 ns
session score)
Fragmented Pictures Test 3.64 ± 5.1 6.3 ± 3.7 ns
Word Stem Completion 4.9 ± 2.9 5.65 ± 2.4 ns
Data are shown as mean ± SD.
ns: not significant.
62 S Vicari ACTA PÆDIATR SUPPL 445 (2004)

Table 2. Results obtained by Williams’ syndrome subjects (WS) and global cognitive impairment affecting ID people but,
mental age normal controls (MA) on the implicit memory tasks. rather, it is a peculiarity of the WS group. It presumably
Implicit memory task WS MA p-Value
results from some specific characteristics of their
anomalous brain development. Concerning this study,
Serial Reaction Time (V–IV trials) ‡62 ‡219 =0.01 any attempt to identify which neuroanatomical structure
Tower of London (II–I testing 1.2 ± 2.6 3.2 ± 1.5 <0.05 is specifically involved in the implicit memory impair-
session score)
Fragmented Pictures Test 3.9 ± 2.9 4.8 ± 3.2 ns ment displayed by WS subjects must necessarily be
Word Stem Completion 5.3 ± 3.7 5.9 ± 2.5 ns based on qualitative analogies of their deficit with that
displayed by adult neurological patients, i.e. Hunting-
Data are shown as mean ± SD. ton’s disease, with a degenerative loss of neurons at the
ns: not significant.
level of basal ganglia, and patients with cerebellar
damage (9).
Brain development in WS children is characterized
Table 3. Results obtained by Down’s syndrome subjects (DS) and by both a remarkable atrophy of basal ganglia (10) and a
mental age normal controls (MA) on the explicit memory tasks. neurochemical alteration (reduction in the neurotrans-
Explicit memory task DS MA p-Value
mitter N-acetylaspartate) in the cerebellum (11), thus
suggesting a neurobiological substrate for the impaired
Free recall 35 ± 8.2 45.3 ± 6.4 <0.001 maturation of procedural learning.
Word recognition 25.6 ± 5.2 29 ± 1.1 <0.05 There may be two reasons for attributing a prevalent
Picture recognition 24.8 ± 5.4 28 ± 0 <0.01
role to the volumetric reduction in basal ganglia in these
Data are shown as mean ± SD. syndromes. First, the performance profile exhibited by
WS children resembles Huntington’s disease patients
more than patients with cerebellar damage. Secondly,
DS subjects, despite severe atrophy of cerebellum, show
normal controls (Table 3), whereas the performance of normal procedural learning of both visuomotor and
WS did not differ from that of normal controls (Table cognitive tasks, thus undermining the role of cerebellar
4). circuit maturation in the development of skill learning.
Further studies, directly evaluating the possible correla-
tion between morphovolumetric and spectroscopic
indices of brain functioning and the ability of WS
Discussion subjects to learn visuomotor and cognitive procedures,
The main result of this study was the documented are needed to understand more clearly the relative
distinct memory patterns in people with DS and with contribution of basal ganglia and abnormal cerebellar
WS. With regard to explicit memory abilities, WS development in the impaired maturation of procedural
subjects showed a similar performance profile to the memory in these subjects.
normal MA matches. In contrast, subjects with DS
obtained lower performance scores than the other two Acknowledgements.—I gratefully acknowledge GA Carlesimo and S
groups. In the implicit memory domain, although Bellucci’s contributions to this study. Many thanks are due to the
children who participated in the study, as well as their parents.
comparable results were observed between the two
experimental groups in repetition priming tasks, WS
subjects were impaired in the ability to learn new
procedures. References
The discrepant performance profiles shown by 1. Vicari S, Caselli MC, Tonucci F. Early language development in
children with DS and WS suggest that the procedural Italian children with Down syndrome: asynchrony of lexical and
learning deficit exhibited by WS (as well as the deficit in morphosyntactic abilities. Neuropsychologia 2000; 38: 634–
explicit memory of DS) is not the expression of the 44
2. Frangiskakis JM, Ewart AK, Morris CA. LIM-kinase-1 hemi-
zygosity implicated in impaired visuospatial constructive cogni-
tion. Cell 1996; 86: 59–69
3. Pezzini G, Vicari S, Volterra V, Milani L, Ossella MT. Children
Table 4. Results obtained by Williams’ syndrome subjects (WS) and with Williams’ syndrome: is there a single neuropsychological
mental age normal controls (MA) on the explicit memory tasks. profile? Dev Neuropsychol 1999; 15: 141–55
4. Vicari S, Carlesimo GA, Brizzolara D, Pezzini G. Short-term
Explicit memory task WS MA p-Value memory in children with Williams’ syndrome: a reduced con-
tribution of lexical–semantic knowledge to word span. Neuro-
Free recall 42.9 ± 6.8 47.1 ± 6.6 ns psychologia 1996; 34: 919–25
Word recognition 28.7 ± 1.05 29.2 ± 1.03 ns 5. Carlesimo GA, Marotta L, Vicari S. Long-term memory in
Picture recognition 27.3 ± 1.02 28 ± 0 ns mental retardation: evidence for a specific impairment in
subjects with Down’s syndrome. Neuropsychologia 1997; 35:
Data are shown as mean ± SD. 71–9
ns: not significant. 6. Vicari S, Bellucci S, Carlesimo GA. Implicit and explicit
ACTA PÆDIATR SUPPL 445 (2004)
memory: a functional dissociation in persons with Down syn-
drome. Neuropsychologia 2000; 38: 240–251
7. Vicari S, Bellucci S, Carlesimo GA. Procedural learning deficit
in children with Williams’ syndrome. Neuropsychologia 2001;
39: 665–77
8. Vicari S, Carlesimo GA. Children with intellectual disabilities.
In: Baddeley A, Wilson B, Kopelman M, editors. Handbook of
memory disorders. New York: John Wiley; 2002: 501–20
9. Molinari M, Leggio MG, Solida A. Cerebellum and procedural
learning: evidence from focal cerebellar lesions. Brain 1997;
120: 1753–62
10. Jernigan TJ, Bellugi U. Anomalous brain morphology on
magnetic resonance images in Williams’ syndrome. Arch Neurol
1990; 47: 529–33
11. Rae C, Karmiloff-Smith A, Lee MA. Brain biochemistry in
Williams’ syndrome. Evidence for a role of the cerebellum in
cognition. Neurology 1998; 51: 33–40
Discussion 6
Volterra: The points that I would like to focus on for
this discussion could also be applied to the interesting
presentation on AD/HD children. I think that children
with WS and DS give us a unique opportunity to study
links between brain genes and behavior, whereas
children with AD/HD offer us a great opportunity to
see the importance of comparison across cultures and
languages. If I can go back to the problem of children
with AD/HD, we have a different approach here in Italy,
because, as Dr Jacobson has underlined, in the USA
these children are treated with medication, while here in
Italy, together with medication we often offer psycho-
logical support or neurological rehabilitation. Now, we
should also ask ourselves whether a different family
context and especially a different educational context,
could in some cases explain the differences found in this
particular disorder. Moving to another aspect, it has
been written that children with WS are the living proof
that cognition could be separated from language,
because, despite being mentally retarded, they seem to
have a very sophisticated language. However, results
produced so far in WS have only been obtained with
English-speaking children affected by this condition,
and English in some way could be considered a quite
simple language, especially from a grammatical point of
view. When similar research started on French- or
Italian-speaking children with WS, we suddenly rea-
lized that, contrary to expectations, the language of
these children is far from intact. So, I think that research
across cultures and languages is very important and, by
comparing data obtained from such research, could
significantly change our approach to designing rehabi-
litation programs, for example. Therefore, I am inter-
ested in opinions from both of you and also to hear
suggestions from the audience.
Vicari: I also believe that it is very interesting to study
the relationship between cognition and language, but I
also think that the question about memory and language
is very important, in the sense that language is highly
Memory and intellectual disabilities 63
related to cognition. What we have found, for example
in the memory domain, is that children with DS have
very poor STM, especially verbal STM. They also have
very poor phonological ability and we know that
phonological ability is quite useful. On the other hand,
children affected by WS have good phonological
abilities and indeed they are good performers on STM
tests; however, the difficulties that these children have,
e.g. in semantics or grammar, are reflected in LTM, in
which they are very poor. Language and memory fit
well into these two types of evaluation. Indeed, it is also
important to look at transcultural work. For this
purpose, we are involved in international programs, as
we are working with some groups in San Diego and also
with a group in London.
Jacobson: Just a few comments on the cultural
differences that might affect diagnosis and incidence
rates of AD/HD. It is clear that children in a large
classroom, perhaps with inexperienced teachers, will
probably have more problems than children who are
learning one to one, taught in their home or in very
small groups. Across cultures, there seems to be some
support for the use of stimulating medication; in South
America caffeine is used instead of fetamine, with the
same results.
Cappa: Dr Vicari, what you have described is a
classical example of “behavior phenotype”. I think it
is important not only in DS and WS, but also in Prader–
Willi syndrome; however, I think this should be added
as an extra tool to the classical ones, such as molecular
biologists and technicians, to make a complete diag-
nosis for many of the dysmorphological syndromes.
Bottazzo: What is the actual relationship between
physiology and pathology? Once you have finally
identified the differences in terms of anatomical and
functional relationship, is there any practical possibility
of helping these children to solve their problems?
Vicari: I think we need to know the neurophysiological
basis of mental retardation. Emerging data indicate very
different profiles in the brains of children with DS and
WS. Is this knowledge useful for children with mental
retardation? I do not know whether there is a direct
implication that might improve their lives; knowing
their brains better, we could probably help them better,
avoiding a lot of unnecessary therapeutic strategies, e.g.
psychological approaches. If we knew better the
behavior of children with these syndromes, I am sure
we could implement better strategies to improve their
learning, and this is exactly what I tried to convince you
about, when I showed the data related to implicit and
explicit memory. So, the strategies to the approach
become quite different and this depends on which type
of mental retardation one faces. For example, children
with DS can work very easily at a low level of
responsibility and in repetitive work; we are collecting
experiences in this sense. I agree with Marco Cappa; we
also need the phenotypes of all these different syn-
dromes.
64 S Vicari
Cilio: I was quite impressed to see that from the overall
population of Rochester, 5–7% had a definite diagnosis
and the others had a probable diagnosis of AD/HD. This
is probably due to a low level of tolerance of scholastic
failure in the USA. It should be stated that in Italy
psychostimulants, such as methylphenidate, which is
the most important drug for these disorders, are still not
available because they are not allowed. So, there is
probably a small proportion of children with AD/HD
even in Italy, who could definitely benefit from this
drug, but they cannot get it.
Jacobson: I was not aware of this. I think in America we
continue to teach that the treatment of children with
AD/HD is much more than the medication itself, which
stresses the idea of structured school and home
environments. In particular, work done at home could
be important, in which distractions have been removed,
the TV is off, the brothers and sisters are away, and
short periods of study are combined with periods of
relaxation. Frequently, one-to-one teaching can be done
instead of group lessons, to reduce distractions. None-
theless, the medication really does seem to help. It
appears that stimulant medications work. The medica-
tions are difficult to use; some children seem to respond
very well to 5 or 15 mg, while some children seem to
need more. You cannot make the decision based on
bodyweight, as there does not appear to be tolerance to
the medication. Frequently, children begin on a dose of
10 mg and remain on that dose for the next 15 y of life,
ACTA PÆDIATR SUPPL 445 (2004)
while other children seem to need higher or lower doses,
and again remain on that steady dose. Some children do
not respond at all to methylphenidate and might switch
to dextroamphetamine, with wonderful results. It is
surprising how little we know about the pharmacology
of these medications. Often, children who become
intolerant to the side effects of the stimulant medica-
tions, e.g. insomnia or weight loss, do very well.
Di Ciommo: I have two questions: first, going back to
the incidence data you have presented, I do not think it
was a “per year incidence”, but a “cumulative inci-
dence”: is that right? Secondly, do you have informa-
tion about risk factors? Having such a good database for
a well-defined population, for example, is there a
relationship with birthweight?
Jacobson: The answer to your first question is yes, the
reference was to a strict cumulative incidence for
children, who enter the cohort at the age of 5 y without
the diagnosis and by the age of 19 y they have the
diagnosis. You are correct, it was not a per year
incidence. In answer to the second question: Dr
Barbaresi and his colleagues are only beginning to
analyze the information contained in the database of this
cohort and are very eager to come back, perhaps in a
year or two, and report to you findings regarding
birthweight, gestational age, breastfeeding versus bottle
feeding, cigarette smoking, etc. We have only begun to
explore the surface of the data.