RELATIONSHIP BETWEEN GENOTYPE AND COGNITIVE
PHENOTYPE IN FEMALES WITH FRAGILE X
B.A., California State University, Fullerton, 1993
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
Christie A. Hartman
This thesis for the Master of Arts
Christie A. Hartman
has been approved
23 / 2-oo /
Hartman, Christie A. (M.A., Psychology)
Relationship between Genotype and Cognitive Phenotype in Females with Fragile X
Thesis directed by Professor Mitchell Handelsman
This study examined general intelligence (IQ) and executive function ability in
females with fragile X, a genetic disorder that causes cognitive impairment. Two
groups were used, 100 with the premutation and 43 with the full mutation. The study
had three goals: to confirm that the full mutation group would perform significantly
lower than the premutation group on all tasks, to look for specific cognitive strengths
and weaknesses for each group, and to investigate whether biochemical factors (CGG
repeat number, activation ratio, and FMRP) were significantly related to cognitive
measures. Results confirmed that the premutation group performed significantly
better than the full mutation group on all IQ tasks, but not executive function (EF).
This is likely due to the fact that subjects with an IQ below 74 did not take the EF
measure, thus skewing the sample. No significant strengths or weaknesses emerged
for the full mutation group, though a trend for a weakness on the arithmetic subtest
was reported. The premutation group reported small but significant results, including
strength on the matrix reasoning subtest and significant weaknesses on arithmetic and
information subtests. Both groups were scored significantly below the population
mean on the executive function measure, but only the premutation group performed
below what would be expected from mean IQ scores. CGG repeat number did not
correlate to any cognitive measure. Though activation ratio produced more robust
correlations, every significant result and trend for both activation ratio and FMRP
resulted from tasks requiring visual-spatial or short-term memory ability, confirming
prior research. Though considered to be cognitively unaffected, the premutation
groups below-average performance on the executive function task generates
questions that need investigation. In general, the role of executive function in fragile
X and in other causes of cognitive impairment needs more investigation.
This abstract accurately represents the content of the candidates thesis. I recommend
I would like to thank Randi J. Hagerman for her generous permission to use the
Genotype-Phenotype data. I would also like to thank Mitchell Handelsman and Ann
Reynolds for their support and guidance during the several months I worked on my
General Cognitive Ability.......................12
Executive Function Ability......................14
General Cognitive Ability.......................19
Executive Function Ability......................22
4. DISCUSSION 27
Activation Ratio and FMRP..............................30
Strengths and Weaknesses...............................31
2.1 N Size and Mean IQ Data for Both Groups................................11
2.2 Hierarchical Structure of the Wechsler Tests............................12
2.3 Descriptions of Wechsler Indexes and Individual Subtests...............13
2.4 Wisconsin Card Sort Test Description...................................15
3.1 Comparing Premutation to Full Mutation Groups on Cognitive Measures.....18
3.2 Single Sample T-test Results for Wechsler Subtest Strengths and Weaknesses 21
3.3 Executive Function: Deviations from Standard Mean of 100...............22
3.4 Correlation Results for Full Mutation Group: Wechsler Scores and FMRP/AR 25
Fragile X is the leading inherited cause of mental retardation (Hagerman, 1999). It
derives its name from the fragile site or break on the long arm of the X
chromosome when cells are incubated in folic acid-deficient culture media. In 1991,
researchers identified a mutation in the FMR1 gene on the X chromosome (Verkerk et
al., 1991; Yu et al., 1991; Vincent et al., 1991; Bell et al., 1991) involving the
expansion of trinucleotide sequence CGG (cytosine, guanine, guanine). CGG repeats
number between 6 and 54 in normal populations (Fu et ah, 1991), 50-200 in the
premutation, and over 200 in the full mutation. Over 200 repeats is typically
associated with hypermethylation of the gene, which shuts down the production of
FMR1 protein (FMRP) and leads to the cognitive and behavioral phenotype
associated with fragile X.
The fragile X phenotype often includes physical features such as elongated face,
prominent ears, and hyper-extensible joints (Hagerman, 1996). Common behavioral
features include attentional problems, hyperactivity, shyness, tantrums, and
perseverative speech and behaviors. Autistic features are also associated with fragile
X, such as hand flapping, poor eye contact, and sensitivity to sensory stimulation.
Another primary feature of the fragile X phenotype is cognitive impairment, the focus
of this paper.
The degree of cognitive impairment seen in those with fragile X varies
considerably for a number of reasons. Those carrying the fragile X premutation are
considered cognitively unaffected (DeVries et al., 1996; Riddle et al., 1998; Sobesky
et al., 1994), as hypermethylation has not occurred and therefore FMRP function is
basically intact. Among those with the full mutation, cognitive impairment (and the
fragile X phenotype in general) is more severe in males than in females. Females have
two X chromosomes, and the unaffected X can still produce the FMR1 protein
thereby ameliorating some of the effects of the mutation. Males have only one X
chromosome; if the gene is mutated, males will typically display cognitive
impairment in the mentally retarded range. According to Merenstein et al. (1996),
there is some variation in IQ for males a higher IQ is associated with mosaicism
(when genetic tests show both full mutation and premutation cells) as well as with
incomplete methylation (where some FMRP is produced).
This paper focuses on females with fragile X. Among females with the full
mutation, DeVries et al. (1996) found that approximately 71% fall into the borderline
range or below on cognitive testing. However, the range of IQ scores is broad scores
for subjects with the full mutation in this study ranged from the 50s to above
average. This variability relates to the activation ratio: the percentage of cells with the
unaffected X chromosome as the active X chromosome (Hagerman, 1999). Since
females have two X chromosomes, only one of them is active or producing protein
in each cell. This inactivation of one X occurs randomly. In fragile X, if the active X
does not have the mutation, then FMRP is produced. Of course, each cell differs as to
whether the mutated or non-mutated X is active, so the activation ratio estimates to
what degree that females cells display the unaffected X as the active X. This ratio has
an impact on phenotype, and several studies have shown that activation ratio (AR) is
positively correlated with IQ and other cognitive measures (DeVries et al.,1996;
Sobesky et al., 1996; Tassone et al., 1999).
This study sought to examine the relationship between genotype and cognitive
phenotype in females with fragile X. Three questions were asked. First, are there
significant differences in cognitive test results between premutation and full mutation
groups? Second, does a pattern of cognitive strengths and weaknesses exist in females
with fragile X, as shown in some studies? And third, is there a significant relationship
between biochemical variables (such as activation ratio) and cognitive measures? In
respective order, I will elaborate on each of these questions.
I. The first question is whether there are significant group differences in cognitive
ability between full mutation and premutation groups. To assess cognitive ability, this
study included two types of measures: a general cognitive (IQ) measure and an
executive function measure. Executive function (EF) refers to the ability to undertake
goal-directed and planning behavior, and involves working memory, cognitive
flexibility, and inhibition of behavior (Bennetto & Pennington, 1996). Studies have
shown that females with the full mutation score significantly lower than premutation
and control groups on tests measuring IQ (DeVries et al., 1996; Sobesky et al., 1996)
as well as executive function (Sobesky et al., 1994 and 1996). In both Sobesky
studies, results were significant even after controlling for IQ, SES, and age.
This study sought to confirm these results, predicting that the premutation group
would perform significantly better than the full mutation group on all cognitive tasks.
In addition, most of the studies on cognitive ability in fragile X report IQ results from
the Wechsler Adult Intelligence Scale Revised (1981) and the Wechsler Intelligence
Scale for Children Revised (1974). This study utilized the revised and updated
WAIS-III (1997) and WISC-III (1991).
II. The second question pertains to cognitive strengths and weaknesses. Are there
particular tasks on which the fragile X population performs particularly well or
poorly? Because females with fragile X function at a higher cognitive level than the
males (who typically score low on all tasks), they provide a better opportunity to
examine cognitive strengths and weaknesses. Because certain abilities are associated
with specific brain regions, attempting to find a pattern of strength/deficit may help us
eventually understand how the fragile X gene causes cognitive deficit.
Evidence of a pattern of strengths and weaknesses has emerged for females with
the full mutation. As a group, studies report relative deficits in visuo-spatial, short-
term auditory memory, and math tasks, and strengths in verbal comprehension
(Bennetto & Pennington, 1996; Mazzocco & Reiss, 1999), as measured by Wechsler
subtests block design, arithmetic, digit span, and vocabulary (Wechsler, 1974, 1981).
Consistent with this, Miezejeski et al. (1986) and Kemper et al. (1986) studied
females with fragile X and reported weakness on block design, arithmetic, and digit
span, and Brainard et al. (1991) reported strength on the vocabulaiy subtest. However,
these studies were completed prior to the 1991 sequencing of the FMR1 gene. At that
time, diagnosis of fragile X status was based on cytogenetic testing, where cells were
examined for the fragile break on the X chromosome. This method lacked accuracy,
particularly with females, and false negative results were not uncommon (Adesman,
Research also suggests that executive function ability may be a weakness in
females with fragile X. Mazzocco et al. (1993) found that females with the full
mutation performed in the impaired range on executive function tasks when compared
to premutation and control groups. The differences were significant even after
controlling for IQ, a key point because its important to rule out that EF deficit is not
simply due to low IQ (Bennetto & Pennington, 1996).
The present study looked for patterns of relative strengths and weaknesses in
females with fragile X using both IQ and executive function measures, focusing
primarily on ipsative comparisons rather than normative ones. It attempted this task
by using the updated DNA testing as well as the most recent versions of the WAIS
and WISC. Based on results from previous research, I predicted that the full mutation
group would display overall relative weakness on block design, arithmetic, and digit
span, as well as significant weakness on the EF measure, the Wisconsin Card Sort
Test (Heaton et al., 1993). I also predicted that the premutation group would display
relative strength in vocabulary and object assembly, and weakness in arithmetic and
WCST. Despite numerous reports that the premutation does not affect cognitive
ability, observation and other preliminary data from the Fragile X Project in Denver,
CO suggest otherwise.
III. The third question addresses the relationship between biochemical factors and
cognitive measures. Specifically, does the amount of FMR1 protein or the CGG
repeat number in the mutation relate to how one scores on cognitive tests? Studies
have documented significant relationships between activation ratio (the ratio of cells
with the unaffected X as the active X one method of estimating FMRP) and
performance on both IQ and executive function testing. For example, positive
relationships were found between activation ratio (AR) and both full scale IQ (FSIQ)
and performance IQ (PIQ), but not verbal IQ (VIQ) (DeVries et al., 1996; Riddle et
al., 1998, Tassone et al., 1999). However, Cornish et al. (1998) did not get these
findings. Sobesky et al. (1996) found that AR significantly relates to executive
function (EF). Reiss et al. (1995) compared girls with fragile X to controls and found
that mean parental IQ predicted 50% of the variance in IQ in controls versus only
26% in the fragile X group, whereas AR predicted 33% of the IQ variance in the
fragile X group. This study highlighted the predictive value of the activation ratio in
In addition to activation ratio, researchers have examined two other biochemical
variables and their relationship to cognitive phenotype in fragile X: the number of
CGG repeats (CGG) in the mutation, and the percentage of cells producing FMR1
protein (FMRP). Many studies have examined the relationship between CGG repeat
number and cognitive ability, but with mixed results. Some found significant
relationships between CGG and FSIQ (Sobesky et al., 1994; Abrams et al., 1994),
executive function (Sobesky et al., 1994), PIQ and picture completion (Abrams et al.,
1994), and object assembly (Cornish et al., 1998). However, others did not find such
relationships (Sobesky et al., 1996; Reiss et al., 1995). Once the CGG repeats number
200 or more the gene becomes hypermethylated, which shuts down FMRP production
- because of this, it is questionable to approach CGG as a continuous variable. The
current study ran analyses with CGG data in an attempt to clarify this issue.
More recent studies have addressed the relationship between cognitive ability and
FMRP in females with fragile X. Tassone et al. (1999) found that FMRP, like
activation ratio, is significantly related to both FSIQ and PIQ, but not VIQ. Kaufinann
et al. (1999) found that both AR and FMRP correlated with FSIQ. The consistent lack
of results with Verbal IQ may be due to the fact that VIQ reflects school learning to a
much greater degree than PIQ perhaps those with fragile X have learned to
compensate for some deficits through learning.
If relationships have been established between biochemical variables and IQ,
which aspects of the IQ score are responsible for these effects? The specific
relationship between biochemical variables and each IQ subtest is still unclear, and
needs examination for the same reason specific strengths and weaknesses need study.
Do these biochemical variables somehow relate to specific skills, such as spatial
ability? Few studies have looked at the relationship between the individual Wechsler
subtests and activation ratio. Abrams et al. (1994) found relationships between AR
and arithmetic, picture arrangement, block design, and coding subtests. Reiss et al.
(1995) found that AR correlated with block design but not object assembly, and did
not examine other subtests. Cornish et al. (1998) found no significant relationship
between AR and block design or object assembly, but also did not look at other
subtests. No studies have examined the relationship between Wechsler subtest
performance and FMRP.
This study examined the relationship between biochemical variables (CGG, AR,
and FMRP) and all cognitive measures (including Wechsler subtests). For the full
mutation group only, I predicted a significant relationship between biochemical
variables (AR, FMRP) and FSIQ and PIQ, but not VIQ, as established in previous
studies. At the subtest level, I predicted that for both groups those previously
established areas of weakness (block design, arithmetic, and digit span) would have a
significant positive relationship with both AR and FMRP, but no relationship to
CGG. This is based on Reiss et al. (1995), who suggested that AR correlated with
those tasks where this population shows the greatest impairment. Also, activation
ratio and FMRP have shown a more promising relationship to cognitive phenotype in
the fragile X population than has CGG.
To summarize all three hypotheses:
1) The premutation group will perform significantly better overall than the full
mutation group on all cognitive measures.
2) The full mutation group will display relative weakness on block design,
arithmetic, digit span, and weakness on the executive function measure. The
premutation group will display relative strength on vocabulary and object
assembly, and relative weakness on arithmetic and executive function.
3) For the full mutation group only, a significant relationship between biochemical
variables (AR, FMRP) and FSIQ and PIQ (but not VIQ) is expected. Also, for
both groups, those previously established areas of weakness (block design,
arithmetic, and digit span) will significantly relate to both AR and FMRP, but not
Data for this study come from a Genotype-Phenotype research project, part of the
Fragile X Project at The Childrens Hospital in Denver, CO and funded by the
National Institutes for Heath. The author would like to thank Dr. Randi Hagerman for
permission to use the data. Both this study and the genotype-phenotype study
received approval by the appropriate human subjects committees. All subjects signed
consent forms and agreed to have their blood taken to determine genetic status and to
take a battery of cognitive and emotional measures. This study included 144 females
with the fragile X gene: 101 with the premutation (mean age = 41.0, range = 7-68)
and 43 with the full mutation (mean age = 22.5, range = 6-51). General IQ data for
both groups is summarized in Table 2.1:
N Size and Mean IQ Data for Both Groups
FSIQ VIQ PIQ
Full Mutation Total n = 40 WISC-III n = 17 WAIS-HI n = 23 74.18 (SD = 16.11) range 40-118 75.80 (SD= 17.72) range 40-119 75.55 (SD= 15.09) range 46-117
Premutation Total n = 100 WISC-III n = 5 WAIS-III n = 95 101.71 (SD = 13.96) range 67-133 99.79 (SD = 13.30) range 66-128 103.73 (SD = 14.55) 64-138
General Cognitive Ability
The Wechsler Adult Intelligence Scale Third Edition (WAIS-III) and the Wechsler
Intelligence Scale for Children Third Edition (WISC-III) were administered by
qualified individuals to all subjects. Both tests have a mean of 100 and standard
deviation of 15.
To aid interpretation of results, the structure of the Wechsler tests and descriptions
of the indexes and individual subtests are provided in Tables 2.2 and 2.3, and are
based on Kaufman et al. (1999). Table 2.3 provides a brief description of each index
and subtest; in parentheses are the abilities tapped by that index or subtest.
Hierarchical Structure of the Wechsler Tests
Full Scale IQ (FSIQ)
Verbal IQ (VIQ) Performance IQ (PIQ)
-comprehension -picture arrangement
Verbal Comprehension Index Freedom from Distractibility Index Perceptual Organization Index Processing Speed Index
-vocabulary -similarities -information -digit span -arithmetic -letter-number sequencing* -block design -matrix reasoning* -picture completion -coding -symbol search
* WAIS-III only
Descriptions of Wechsler Indexes and Individual Subtests
Verbal Comprehension Index Verbal conceptualization, knowledge, and expression
Freedom from Distractibility Index Number ability, sequential processing, attention span
Perceptual Organization Index Non-verbal thinking, visual-spatial and visual motor skills
Processing Speed Index Response speed to non-verbal problems
Picture Completion Finding whats missing in various pictures (visual organization and recognition)
Vocabulary Giving word definitions (word knowledge, verbal expression)
Coding Copying symbols (visual processing speed)
Similarities Explaining how two items or concepts are alike (abstract/categorical thinking)
Block Design Using blocks to make designs like they appear in pictures (spatial visualization, analysis of whole into parts)
Arithmetic Doing arithmetic problems without paper/pencil (working memory, arithmetic ability)
Matrix Reasoning Choosing the design that best completes the overall design pattern (visual-spatial skill, analogic reasoning) WAIS-III only
Digit Span Repeating back increasing numbers of digits (short term auditory memory, rote recall)
Information Answering fact-based questions (general knowledge, long term memory)
Picture Arrangement Rearranging cards so they tell a logical story (temporal sequencing, visual organization)
Comprehension Includes questions about social institutions and expected social behavior (social judgment, knowledge of social conventions)
Symbol Search Matching target symbols to a group of symbols (visual processing speed)
Table 2.3 (Cont.)
Letter-number Sequencing Repeating back groups of numbers and letters in a specified order (working memory, sequential processing ability) WAIS-III only
Object Assembly Constructing an object with puzzle pieces (ability to understand relationships between parts and synthesize into a whole)
Executive Function Ability
EF refers to the ability to undertake goal-directed and planning behavior, involving
working memory, cognitive flexibility, and inhibition of behavior (Bennetto and
Pennington, 1996). The Wisconsin Card Sort Test (WCST) involves executive
function ability and has been shown to be useful in detecting dysfunction in those
parts of the brain responsible for planning behaviors (Heaton et al., 1993). The test
measures the ability to develop and maintain a problem solving strategy and to shift
that strategy in response to changing variables. Subjects view cards with varying
colors and shapes, and must match the cards based on specific rules. They are told
only if they are right or wrong. The rules periodically change and so the test focuses
on not only the ability to adhere to each rule but also to exhibit flexibility when the
rules change. To aid interpretation, the scoring components of the WCST are
explained below in Table 2.4. The first six measures utilize standard scores (mean =
100, SD = 15). The last four measures are coded by percentile rank, with 5 = >16,4 =
11-16, 3 = 6-10, 2 = 2-5, and 1 = 1 or <1
Wisconsin Card Sort Test Description
Total errors total errors made during administration
% total errors total errors/total number of trials
% perseverative responses Percent of total trials where subject persists on responding according to the same rule (can be correct or incorrect)
% perseverative errors Percent of total trials where subject persists in making the same type of error
% non-perseverative errors Percent of trials where errors are not perseverative
% conceptual level responses Consecutive correct responses of three of more, showing understanding of the current rule
# categories completed six possible, and need 10 consecutive correct answers to complete a category
# trials to first category how many trials until subject picks up on rules
Fail to maintain set five or more correct answers followed by an error before completion of that category (which requires 10)
Learning to learn Increased efficiency at task (learning how task works)
Subjects underwent genetic testing to determine CGG repeat number (which
determines whether subject has a premutation or a full mutation), activation ratio, and
FMRP. CGG repeat that numbers 50-200 is considered to be premutation, and 200 or
more is considered to be a full mutation (Fu et al., 1991). CGG repeat number is
obtained through polymerase chain reaction (PCR) and southern blotting, the details
of which are described, respectively, in Pergolizzi (1992) and Taylor (1994). Both the
shortest and longest number of repeats found for each subject are obtained during this
procedure; the average of these two numbers was used for this study.
The activation ratio is the percentage of cells that have the unaffected (non-
mutated) X chromosome as the active X chromosome (Hagerman, 1999). In females
with fragile X, it should correlate with how much FMRP activity there is. Activation
ratio is obtained by southern blot analysis, as described by Rousseau (1991).
The FMRP measure refers to the percentage of lymphocytes producing FMRP,
another way of estimating FMRP production. This measure, determined through a
rapid antibody test procedure, was developed by Willemsen (1995) and is described
in Tassone et al. (1999).
All analyses were conducted using SPSS. Because this study is exploratory in
nature and so many analyses were done, the chances of finding a significant result are
increased considerably. A stricter p-value of 0.01 was adopted and anything between
0.01 and 0.05 is reported as a trend. Significant results are noted with **, and trends
are noted with *. However, even with the stricter p-value there may be spurious
findings and therefore all results need to be interpreted with great caution.
Of the 23 adult females with the full mutation, 14 (61%) scored in the borderline
range or below on the WAIS-III. Of the 17 girls with the full mutation, 14 (82%)
scored in the borderline range or below on the WISC-III. Combined, 70% of the full
mutation group fall into the borderline range or below, consistent with previous
research by DeVries et al. (1996).
The results of each question this study addresses are presented in order.
As predicted, independent samples t-tests showed that the full mutation group
scored significantly lower than the premutation group (p < .0001) on all general
cognitive measures. See Table 3.1 for data. However, when comparing these two
groups on the executive function measure, some trends emerged but no significant
differences were seen. It is possible that the limited sample size for full mutation
females (n = 23 compared with n = 93 for premutation) made detection of differences
too difficult. More importantly, this analysis does not include those subjects with the
full mutation who had IQ scores of 73 or below because they were unable to either
undergo or finish the test. Therefore, it is likely that the sample of WCST scores for
the FM group is skewed.
Comparing Premutation (PM) to Full mutation (FM) Groups on Cognitive Measures
ISlilSilSIi PM (n, mean) FM (n, mean) T P
FSIQ 100, 101.71 40, 74.18 10.079 0.000**
VIQ 100, 99.79 40, 75.80 8.734 0.000**
PIQ 100, 103.73 40, 75.55 10.244 0.000**
VCI 100,100.05 37, 78.22 7.872 0.000**
POI 100,105.02 37, 76.65 9.787 0.000**
FDI 100, 99.23 37, 73.68 7.423 0.000**
PSI 100,102.02 36, 77.00 8.742 0.000**
Picture Completion 100,10.53 37, 6.73 6.165 0.000**
Vocabulary 100,10.50 37, 5.76 7.884 0.000**
Coding 100,10.35 37, 6.32 6.991 0.000**
Similarities 100, 9.96 37, 5.49 6.245 0.000**
Block Design 100, 10.76 37, 5.68 9.164 0.000**
Arithmetic 100, 9.08 37, 4.49 8.066 0.000**
Matrix Reasoning 95, 11.24 21,6.29 6.238 0.000**
Digit Span 100, 9.86 37, 6.38 6.498 0.000**
Information 100, 9.73 37, 6.73 5.588 0.000**
Picture Arrangement 100, 10.09 37, 5.86 7.047 0.000**
Table 3.1 (Cont.)
Comprehension 100,10.67 37, 5.62 8.689 0.000**
Symbol Search 100, 10.44 36,5.11 9.131 0.000**
Letter-number Seq. 95, 10.48 21,6.81 3.641 0.001**
Object Assembly 93,10.24 36, 5.94 7.588 0.000**
Total errors 93, 92.67 23, 87.00 1.551 0.124
% total errors 93,91.77 23, 83.35 2.108 0.037*
% persev resp. 93, 94.76 23, 83.65 2.441 0.016*
% persev errors 93, 93.42 23, 83.57 2.183 0.031*
% non-persev err. 93,95.15 23,91.87 0.885 0.378
% cone level resp 93, 92.37 23, 84.17 1.967 0.052
# categ completed 93, 4.99 23, 3.65 0.942 0.348
# trials to 1st categ 93, 4.30 23, 3.57 2.097 0.045*
Fail to maint set 93,4.58 23, 4.22 1.168 0.253
Learning to learn 85, 4.29 19, 3.68 1.885 0.062
General Cognitive Ability
To examine overall patterns of relative strength and weakness for both premutation
and full mutation groups, I used a procedure undertaken by Freund and Reiss (1991).
First, a mean subtest score was calculated for each subject. Each subtest score was
then converted to a figure that represents the deviation of that score from that
individuals overall mean. Then for each subtest, an overall mean was calculated from
the individual deviation scores. For each subtest, a single-sample t-test was used to
determine if the deviation mean differed significantly from zero. This method was
chosen because it determines strength or weakness with regard to ones individual
ability and presents it in aggregate form. Using normative data for cognitively
impaired populations only tells us that they are below average this study sought to
look at strengths and weaknesses regardless of IQ.
The full mutation group did not exhibit the predicted strengths or weaknesses,
except that the arithmetic subtest produced a trend for a potential weakness. The mean
scores for arithmetic were more than 2 scaled points below the mean, which signifies
performance below the 25"1 percentile. This is consistent with the report that girls and
boys with fragile X tend to struggle with math more than any other subject. However,
the premutation group displayed some small but significant strengths and weaknesses,
as summarized in Table 3.2. This group did not display strength in the predicted areas
of vocabulary and object assembly, but like the full mutation group displayed
weakness in arithmetic. Yet even the 2 most robust results, arithmetic (weakness) and
matrix reasoning (strength), only deviated from the mean by roughly one scaled point.
Large sample size is likely responsible for such a result.
Single Sample T-test Results for Wechsler Subtest Strengths and Weaknesses
Full mutation n M diff. t P S/W
Picture Completion 37 -1.56E-03 -0.001 0.999
Vocabulary 37 0.946 0.920 0.363
Coding 37 0.393 0.370 0.713
Similarities 37 1.208 1.130 0.266
Block Design 37 1.025 0.991 0.328
Arithmetic 37 2.183 2.245 0.031* Weakness
Matrix Reasoning 19 0.814 0.762 0.456
Digit Span 37 0.341 0.329 0.744
Information 37 -1.56E-03 -0.001 0.999
Picture Arrangement 37 0.840 0.826 0.414
Comprehension 37 1.077 1.039 0.306
Symbol Search 36 1.625 1.582 0.122
Letter-number Seq. 19 0.814 0.753 0.461
Object Assembly 37 0.917 0.872 0.389
Premutation n M diff. t P S/W
Picture Completion 100 -0.245 -1.164 0.247
Vocabulary 100 -0.215 -1.088 0.279
Coding 100 -6.46E-02 -0.276 0.783
Similarities 100 0.325 1.790 0.076
Block Design 100 -0.475 -2.420 0.017* Strength
Arithmetic 100 1.205 6.438 0.000** Weakness
Matrix Reasoning 95 -0.988 -4.254 0.000** Strength
Digit Span 100 0.425 2.024 0.046* Weakness
Information 100 0.555 2.911 0.004** Weakness
Picture Arrangement 100 0.195 0.861 0.392
Comprehension 100 -0.385 -2.020 0.046* Strength
Symbol Search 100 -0.155 -0.628 0.532
Letter-number Seq. 95 -0.230 -1.051 0.296
Object Assembly 93 -1.24E-02 -0.056 0.955
Executive Function Ability
To determine if either premutation or full mutation groups displayed strength or
weakness on the WCST, single sample t-tests were used to compare mean scores to a
standard normative mean score of 100. For the last four categorical variables, mean
scores were compared to 5, which represents a percentile rank of >16. Results suggest
that, consistent with previous research, executive function is a potential weakness for
females with fragile X. As predicted, the full mutation group performed significantly
below average on most of the scoring dimensions when compared to those in their age
group and with their education level (see Table 3.3). As predicted, but not consistent
with the literature, the premutation group performed significantly below the mean
when compared to others in their age group and with their education level.
Executive Function: Deviations from Standard Mean of 100
Full mutation n Mean t P
Total errors 23 87.00 -3.600 0.002**
% total errors 23 83.35 -4.631 0.000**
% perseverative resp. 23 83.65 -4.467 0.000**
% perseverative errors 23 83.57 -3.951 0.001**
% non-persev. Errors 23 91.87 -2.158 0.042*
% concept, level resp. 23 84.17 -4.368 0.000**
# categories completed 23 3.65 -4.236 0.000**
# trials to 1 category 23 3.57 -4.326 0.000**
Fail to maintain set 23 4.22 -2.719 0.013*
Learning to learn 19 3.68 -4.763 0.000**
Table 3.3 (Cont.)
Premutation N Mean T P
Total errors 93 92.67 -4.631 0.000**
% total errors 93 91.77 -4.628 0.000**
% perseverative resp. 93 94.76 -2.526 0.013*
% perseverative errors 93 93.42 -3.298 0.001**
% non-persev. Errors 92 95.15 -3.031 0.003**
% concept.level resp. 92 92.37 -4.069 0.000**
# categories completed 92 4.99 -0.015 0.988
# trials to 1 category 92 4.30 -5.823 0.000**
Fail to maintain set 92 4.58 -3.949 0.000**
Learning to learn 85 4.29 -5.047 0.000**
An important question is how did each group perform on the executive function
task with respect to IQ? Because those with an IQ of 73 or below did not take the
WCST, there are only 20 subjects in the full mutation group who completed both IQ
and EF tasks. For this group, the average EF score is 83.88 compared to a mean IQ of
84.45. So although the full mutation group performed significantly below the
population mean on executive function, their mean EF scores were close to their mean
IQ scores. By contrast, the premutation group performed in the average range on the
EF measure, but significantly below others in their age/education group (M = 93.36),
and below what might be expected from their mean IQ (M = 101.7).
Pearsons product-moment correlation was utilized to examine relationships
between biochemical variables and scores on all cognitive measures. Within the
premutation group, no significant relationships were found between any of the
biochemical and cognitive measures, confirming the general finding that the
premutation does not affect cognitively ability. The predicted relationships with block
design, arithmetic, and digit span were not seen. For the full mutation group, no
significant relationships were found between biochemical measures and the WCST,
or between CGG and any cognitive measures. However, both AR and FMRP
correlated significantly with some cognitive measures. For FMRP, a significant
positive relationship emerged with the digit span subtest (predicted), and trends with
arithmetic (predicted), picture arrangement, and object assembly subtests as well as
the perceptual organization index (POI) and the freedom from distractibility index
(FDI). The predicted relationship with block design was not observed.
For activation ratio, significant relationships emerged with block design
(predicted), digit span (predicted), picture arrangement, and object assembly, and with
the POI, FDI, and PIQ (predicted). Trends also exist with FSIQ (predicted), picture
completion, and arithmetic (predicted). Results are summarized in Table 3.4.
Correlation Results for Full Mutation Group: Wechsler Scores and FMRP/AR
FMRP (n, r, p) Activation Ratio (n, r, p)
FSIQ 34, 0.334, p = 0.054 35, 0.402, p = 0.017*
VIQ 34, 0.260, p = 0.138 35, 0.279, p = 0.104
PIQ 34, 0.333, p = 0.055 35, 0.429, p = 0.010**
VCI 31, 0.303, p = 0.097 32, 0.319, p = 0.075
POI 31, 0.362, p = 0.046* 32, 0.484, p = 0.005**
FDI 31, 0.370, p = 0.041* 32, 0.473, p = 0.006**
PSI 30, 0.255, p = 0.173 31, 0.275, p = 0.134
Picture Completion 31, 0.264, p = 0.152 32, 0.409, p = 0.020*
Vocabulary 31, 0.307, p = 0.093 32, 0.342, p = 0.055
Coding 31, 0.188, p = 0.310 32, 0.187, p = 0.305
Similarities 31, 0.278, p = 0.131 32, 0.206, p = 0.258
Block Design 31, 0.353, p = 0.052 32, 0.471, p = 0.007**
Arithmetic 31, 0.390, p = 0.030* 32, 0.411, p = 0.019*
Matrix Reasoning 18, 0.046, p = 0.855 18, 0.179, p = 0.477
Digit Span 31, 0.495, p = 0.005** 32, 0.606, p = 0.000**
Information 31, 0.210, p = 0.256 32, 0.285, p = 0.114
Picture Arrangement 31, 0.413, p = 0.021* 32, 0.499, p = 0.004**
Comprehension 31, 0.088, p = 0.640 32, 0.122, p = 0.507
Symbol Search 30, 0.236, p = 0.210 31, 0.277, p = 0.131
Letter-number sequencing 18, -0.108, p = 0.670 18, 0.377, p = 0.123
Object Assembly 31, 0.372, p = 0.039* 31, 0.521, p = 0.003**
Though the results for activation ratio were more robust, a similar pattern emerged
for both AR and FMRP. Both measures displayed significance or trends with many of
the same cognitive measures, including object assembly, picture arrangement, digit
span, arithmetic, the perceptual organization index, and the freedom from
distractibility index. Overall, every significant result and every trend resulted with
tasks requiring visual-spatial or short-term memory ability, consistent with the report
that these are areas of weakness for females with fragile X (Bennetto & Pennington,
This study confirmed that females with the fragile X full mutation score
significantly lower on cognitive tasks than those with the premutation. The
premutation group performed in the normal range on the IQ measure, but scored
significantly below the mean on the executive function measure. Although these
scores are in the normal range, they are lower than others in their age group and with
their education level, and lower than their mean IQ scores. Research generally
supports the idea that those with the fragile X premutation are cognitively unaffected.
This seems questionable when one considers that females with the premutation
display significant differences from control populations with medical and emotional
functioning, including higher rate of premature menopause (Partington et al., 1996)
and higher prevalence of anxiety disorders (Franke et al., 1996).
Interestingly, there is some preliminary evidence that some men with the
premutation display signs of neurological impact when they get older. Tremors have
been reported in this sub-group, as have deficits in performance IQ and executive
function. This suggests the possibility of a cognitive phenotype in premutation
individuals, something worth studying further.
Some studies report that females with fragile X dont perform as well on tasks
requiring visuo-spatial skill. Miezejeski et al. (1986) and Kemper et al. (1986) found
block design to be a weakness in this group, and Abrams et al. (1994) found that
block design and object assembly are positively correlated with activation ratio. Yet
Cornish et al. (1998) did not get such results. Their study examined spatial ability in
girls with fragile X, dividing the spatial tasks into 4 categories: visuo-spatial, visou-
construction, visuo-perception, and visuo-motor. The subjects displayed weakness on
three of the four visuo-construction tasks only, one of which was block design. The
results for object assembly were not significant. The authors suggested that this
population finds abstract designs (block design) more difficult than meaningful ones
(object assembly) this assumption makes sense, for this study and others report
weakness with abstract visual-spatial and short-term memory tasks but not with
verbal tasks. Verbal tasks utilize crystallized intelligence, which is developed through
school learning and can be improved with time.
Matrix reasoning, new to the WAIS-III, has not been examined in the literature it
is a task that requires visual processing and is abstract in nature. In accordance with
the literature, one might expect subjects to perform weakly on matrix reasoning, but
mean scores on this subtest were among the higher subtest scores. This was
particularly true for the premutation group, whose mean score on this task was not
only the highest of all their subtest means, but slightly higher than the population
mean. One interesting point is that the matrix reasoning task is not timed, unlike
block design, and it is possible that the extra time provided made up for any potential
weakness. Bennetto and Pennington (1996) suggest that the weakness in spatial
ability seen in this population may actually be due to a deficit in abstract reasoning, a
component of fluid intelligence. Fluid intelligence is more sensitive to neurological
damage, and performance on such tasks tends to decrease as IQ decreases. With such
mixed results on spatial ability in this population, it may be interesting to examine the
other abilities tapped by these visuo-spatial tasks, including abstract reasoning.
Is EF a weakness for women with fragile X? Bennetto and Pennington (1996)
suggested that a deficit in this area may underlie the deficits seen on more specific
tasks. Studies have documented EF as a potential weakness for females with the full
mutation, and have found that EF tends to increase as activation ratio increases
(Sobesky et al., 1996). This study found that subjects performed significantly below
the population mean on the WCST. Standard scores for the full mutation group
averaged in the 80s, putting them in the mildly impaired to below average range.
This did not include those subjects with IQs below 74. The premutation group scores
ranged from the low to mid 90s, placing them mostly in the average range, yet still
lower than would be expected from their mean IQ. Perhaps it is females with the
fragile X premutation, rather than the full mutation, that exhibit executive function
Activation Ratio and FMRP
Between these two biochemical measures, the activation ratio was the more robust
in its ability to correlate with the cognitive test scores, suggesting that it is a better
predictor of cognitive phenotype. In this study the AR correlated with those measures
that are consistent with the literature: PIQ (not VIQ), POI and FDI, block design, digit
span, picture arrangement, and object assembly. FMRP had similar results, but these
results were not quite statistically significant. So despite the mixed results in trying to
find profile strengths and weaknesses, there is some evidence supporting that as AR
(and FMRP) increases, scores on the aforementioned tasks increase.
This study found no relationships between CGG and any of the cognitive
measures. CGG is useful in diagnosing fragile X, placing patients into full or
premutation categories, and therefore predicting how an individual with the gene will
function. But CGG appears to have little predictive value within premutation and full
mutation groups, and probably cannot be viewed as a continuous variable. According
to Hagerman (1999), more useful predictive variables include activation ratio for
females and percent methylation for males.
Strengths and Weaknesses
Attempting to identify a cognitive profile among the fragile X population has
produced some results, but they arent entirely consistent. Put together, studies most
consistently reported weaknesses in block design, arithmetic, and digit span. This
study only confirmed the weakness in arithmetic, for both full mutation and
Ideally, this study would have a control group comprised of family members
whom tested negative for fragile X; this would allow for comparison of fragile X and
control groups without the confound of the genetic variability created by unrelated
controls. Such a group is hard to come by, as unaffected family members dont often
get evaluated if data are collected in a clinical setting.
One cannot study IQ without taking socioeconomic status into consideration, as
there exists a relationship between the two (Bouchard & Segal, 1985). This study
could not address this issue because of insufficient data. There are data for education
level, but not for every subject. Also, education is only an estimate of SES; even then,
it would only work for adults, and this study included a significant number of
children. Ideally, the influence of SES would be covaried out during analysis.
Another issue in this study is that the mean ages differed significantly between full
mutation and premutation groups. All cognitive measures used in this study utilize
age norms, making scores comparable. But some studies suggest that activation ratio
changes with age (Rousseau et al, 1991), which could possibly influence the results of
As mentioned previously, because so many analyses were done in this study it is
important to interpret these results with caution. Thus, the primary value of this
exploratory study can be to provide new directions to pursue.
Since much is still unknown regarding how the fragile X mutation impacts
cognitive phenotype, future research may want to employ additional
neuropsychological measures. Executive function studies with fragile X subjects have
raised some questions, and more work in this area would be intriguing, especially
considering the possibility than even premutation females may show some executive
function deficit. Designing studies in this area may not only help to understand fragile
X cognitive phenotype better, but may help us better understand executive function
ability and how it applies to other forms of cognitive disability.
Abrams, MT, Reiss, AL, Freund, LS, Baumgardner, TL, Chase, GA, Denckla, MB.
1994. Molecular-neurobehavioral associations in females with the fragile X full
mutation. American Journal of Medical Genetics 51: 317-327.
Adesman, A. 1995. Fragile X Syndrome. In A. J. Capute and P. J. Accardo (eds.),
Developmental Disabilities in Infancy and Childhood, 2nd ed. Baltimore: P.H.
Brookes Publishing Co.
Bell, MV, Hirst, MC, Nakahori, Y, MacKinnon, RN, Roche, H, Flint, TJ, Jacobs, PA,
Tommerup, N, Tranebjaerg, L, Froster-Iskenius, U, Kerr, B, Turner, G,
Lindenbaum, RH, Winter, R, Pembrey, M, Thibodeau, S, Davies, KE. 1991.
Physical mapping across the fragile X: Hypermethylation and clinical expression of
the fragile X syndrome. Cell 64: 861-866.
Bennetto, L and Pennington, BF. 1996. The neuropsychology of fragile x syndrome.
In R. J. Hagerman and A. Cronister-Silverman (eds.), Fragile XSyndrome:
Diagnosis, Treatment, and Research, 2nd ed. Baltimore: Johns Hopkins University
Bouchard Jr., TJ and Segal, NL. 1985. Environment and IQ. In B.B. Wolman (Ed.),
Handbook of Intelligence: Theories, Measurements, and Applications. New York:
Brainard, SS, Schreiner, RA, Hagerman, RJ. 1991. Cognitive profiles of the carrier
fragile X woman. American Journal of Medical Genetics 38: 505-508.
Cornish, KM, Munir F, Cross, G. 1998. The nature of the spatial deficit in young
females with fragile X syndrome: a neuropsychological and molecular perspective.
Neuropsychologia 36(11): 1239-1246.
DeVries, BB, Wiegers, AM, Smits, APT, Mohkamsing, S, Duivenvoorden, HJ, Fryns,
JP, Curfs, LMG, Halley, DJJ, Oostra, BA, van den Ouweland, AMW, Niermeijer,
MF. 1996. Mental status of females with an FMR1 gene full mutation. American
Journal of Human Genetics 58: 1025-1032.
Franke, P, Maier, W, Hautzinger, M, Weiffenbach, O, Gansicke, M, Iwers, B,
Proustka, F, Schwab, SG, Froster, U. 1996. Fragile X carrier females: Evidence for
a distinct psychopathological phenotype? American Journal of Medical Genetics
64 (2): 334-339.
Fu, Y, Kuhl, DP A, Pizzuti, A, Pieretti, M, Sutcliffe, JS, Richards, S, Verkerk, JMH,
Holden, JJA, Fenwick Jr., RG, Warren, ST, Oostra, BA, Nelson, DL, Caskey, CT.
1991. Variation of the CGG repeat at the fragile X site results in genetic instability:
resolution of the Sherman paradox. Cell 67: 1047-1058.
Hagerman, RJ. 1996. Physical and Behavioral Phenotype. In R. J. Hagerman and A.
Cronister-Silverman (eds.), Fragile X Syndrome: Diagnosis, Treatment, and
Research, 2nd ed. Baltimore: Johns Hopkins University Press.
Hagerman, RJ. 1999. Fragile X Syndrome. In Neurodevelopmental Disorders:
Diagnosis and Treatment. New York: Oxford University Press.
Heaton, RK, Chelune, GJ, Talley, JL, Kay, GG, Curtiss, GC. 1993. The Wisconsin
Card Sort Test manual Revised. Odessa, FL: Psychological Assessment
Kaufman, AS and Lichtenberger, EO. 1999. Essentials of WAIS-IIIAssessment. New
York: John Wiley & Sons, Inc.
Kaufmann, WE, Abrams, MT, Chen, W, Reiss, AL. 1999. Genotype, molecular
phenotype, and cognitive phenotype: correlations in fragile X syndrome. American
Journal of Medical Genetics, 83, 286-295.
Kemper, MB, Hagerman, RJ, Ahamad, RS. 1986. Cognitive profiles and the spectrum
of clinical manifestation in heterozygous fragile X females. American Journal of
Medical Genetics 23: 139-156.
Mazzocco, MM, Pennington, BF, Hagerman, RJ. 1993. The neurocognitive
Phenotype of female carriers of fragile X: Additional evidence for specificity.
Developmental and Behavioral Pediatrics 14: 328-334.
Mazzocco, MM and Reiss, AL. 1999. A behavioral neurogenetics approach to
understanding the fragile X syndrome. In Tager-Flusberg, H. (ed.),
Neurodevelopmental Disorders. Cambridge, MA: The MIT Press.
Meizejeski, CM, Jenkins, EC, Hill, AL. 1986. A profile of cognitive deficit in females
From fragile X families. Neuropsychologia 24: 405-409.
Merenstein, SA, Sobesky, WE, Taylor, AK, Riddle, JE, Tran, HX, Hagerman, RJ.
1996. Molecular-clinical correlations in males with an expanded FMR1 mutation.
American Journal of Medical Genetics 64 (2): 388-394.
Partington, MW, Moore, DY, Turner, GM. 1996. Confirmation of early menopause
in fragile X carriers. American Journal of Medical Genetics 64 (2): 370-372.
Pergolizzi, RG, Erster, SH, Goonewardena, P, Brown, WT. 1992. Detection of full
Fragile X mutation. Lancet 339: 271-271.
Reiss, AL, Freund, LS, Baumgardner, TL, Abrams, MT, Denckla, MB. 1995.
Contribution of the FMR1 gene mutation to human intellectual dysfunction. Nature
Genetics 11: 331-334.
Riddle, JE, Cheema, A, Sobesky, WE, Gardner, SC, Taylor, AK, Pennington, BF,
Hagerman, RJ. 1998. Phenotypic involvement in females with the FMR1 gene
mutation. American Journal on Mental Retardation 102 (6): 590-601.
Rousseau, F, Heitz, D, Tarleton, J, MacPherson, J, Malmgren, H, Dahl, N, Bamicoat,
A, Mathew, C, Momet, E, Tejada, I, Maddalena, A, Spiegel, R, Schinzel, A,
Marcos, JAG, Schorderet, DF, Schaap, T, Maccioni, L, Russo, S, Jacobs, PA,
Schwartz, C, Mandel, JL. 1994. A multicenter study on genotype-phenotype
correlations in the fragile X syndrome, using direct diagnosis with probe StB 12.3:
The first 2,253 cases. American Journal of Human Genetics 55: 225-237.
Rousseau, F, Heitz, D, Oberle, I, Mandel, JL. 1991. Selection in blood cells from
female carriers responsible for variable phenotypic expression of the fragile X
syndrome: Inverse correlation between age and proportion of active X carrying the
fragile X mutation. Journal of Medical Genetics 28: 830-836.
Sobesky, WE, Taylor, AK, Pennington, BF, Bennetto, L, Porter, D, Riddle, J,
Hagerman, R. 1996. Molecular/Clinical correlations in females with fragile X.
American Journal of Medical Genetics 64: 340-345.
Sobesky, WE, Pennington, BF, Porter, D, Hull, CE, Hagerman, RJ. 1994. Emotional
and neurocognitive deficits in fragile X. American Journal of Medical Genetics 51:
Tassone, F, Hagerman, RJ, Ikle, DN, Dyer, PN, Lampe, M, Willemsen, R, Oostra,
BA, Taylor, AK. 1999. FMRP expression as a potential prognostic indicator in
fragile X syndrome. American Journal of Medical Genetics 84: 250-261.
Taylor, AK, Safanda, JF, Fall, MZ, Quince, C, Lang, KA, Hull, CE, Carpenter, I,
Staley, LW, Hagerman, RJ. 1994. Molecular predictors of involvement in fragile X
females. Journal of the American Medical Association 271: 507-514.
Verkerk, AJ, Peiretti, M, Sutcliffe, JS, Fu, YH, Kuhl, DP, Pizzuti, A, Reiner, O,
Richards, S, Victoria, MF, Zhang, F, Eussen, BE, van Ommen, GJ, Blonden, LJ,
Riggins, GJ, Chastain, JL, Kunst, CB, Galjaard, H, Caskey, CT, Nelson, DL,
Oostra, BA, Warren, ST. 1991. Identification of a gene (FMR1) containing a CGG
repeat coincident with a breakpoint cluster region exhibiting length variation in
fragile X syndrome. Cell 65: 905-914.
Vincent, S, Heitz, D, Petit, C, Kretz, C, Oberle, I, Mandel, JL. 1991. Abnormal
pattern detected in fragile X patients by pulsed-field gel electrophoresis. Nature
Wechsler, D. 1974. Wechsler Intelligence Scale for Children Revised. San Antonio,
TX: The Psychological Corporation.
Wechsler, D. 1981. Wechsler Adult Intelligence Scale Revised. San Antonio, TX:
The Psychological Corporation.
Wechsler, D. 1991. Wechsler Intelligence Scale for Children Third Edition. San
Antonio, TX: The Psychological Corporation.
Wechsler, D. 1997. Wechsler Adult Intelligence Scale Third Edition. San Antonio,
TX: The Psychological Corporation.
Willemsen, R, Mohkamsing, S, DeVries, B, Devys, D, van den Ouweland, A,
Mandel, JL, Galjaard, H, Oostra, B. 1995. Rapid antibody test for fragile X
Syndrome. Lancet 345: 1147-1148.
Yu, S, Pritchard, M, Kenner, E, Lynch, M, Nancarrow, J, Baker, E, Holman, K,
Mulley, JC, Warren, ST, Schlessinger, D, Sutherland, GR, Richards, RI. 1991.
Fragile X genotype characterized by an unstable region of DNA. Science 252: 1179-