Handbook of Learning Disabilities

228 CAUSES AND BEHAVIORAL MANIFESTATIONS

in phonological awareness, Journal of Experi- between phonological processing abilities and
mental child Psychology, 63, 1–21. word-level reading as children develop from be-
Tyler, A., & Nagy, W. (1990). Use of derivational ginning to skilled readers: A 5-year longitudinal
morphology during reading. (1990). Cognition, study. Developmental Psychology, 33, 468–479.
36, 17–34. Watt, W. C. (1989). Getting writing right. Semioti-
Tzeng, O. J. L., Hung, D. L., & Wang, W. (1977). ca, 75, 279–315.
Speech recoding in reading Chinese characters. Wolf, M. (1984). Naming, reading and the dyslexi-
Journal of Experimental Psychology: Human as: A longitudinal overview. Annals of Dyslexia,
Learning and Memory, 3, 621–630. 34, 87–115.
van Orden, G. C. (1987). A rows is a rose: Spelling, Wolf, M., & Bowers, P. G. (1999). The double-
sound, and reading. Memory and Cognition, 15, deficit hypothesis for the developmental dyslexia.
181–198. Journal of Educational Psychology, 91, 415–438.
Walley, A. C. (1993). The role of vocabulary devel- Wolf, M., Bowers, P. G., & Biddle, K. (2000). Nam-
opment in children’s spoken word recognition ing-speed processes, timing, and reading: A con-
and segmentation ability. Developmental Review, ceptual review. Journal of Learning Disabilities,
13, 286–350. 33, 387–407.
Wagner, R. K., & Torgesen, J. K. (1987). The nature Wolf, M., & Goodglass, H. (1986). Dyslexia, dys-
of phonological processing and its causal role in nomia and lexical retrieval: A longitudinal inves-
the acquisition of reading skills. Psychological tigation. Brain and Language, 28, 159–168.
Bulletin, 101, 192–212. Yopp, H. K. (1988). The validity and reliability of
Wagner, R. K., Torgesen, J. K., Rashotte, C. A., phonemic awareness tests. Reading Research
Hecht, S. A., Barker, T. A., Burgess, S. R., Don- Quarterly, 23, 159–177.
ahue, J., & Garon, T. (1997). Changing relations

14

Self-Concept and Students
with Learning Disabilities

Batya Elbaum
Sharon Vaughn

A man cannot be comfortable without his own approval.
—TWAIN (1929, p. 17)

In the educational, psychological, and popu- which may lower self-esteem. These acade-
lar literature, self-concept has long been mic difficulties are often coupled with diffi-
considered important both in and of itself culties in the social domain. The effect of
and as a variable that mediates other signifi- these academic and social challenges on the
cant outcomes, such as academic achieve- self-concept of students with LD can range
ment (Carlock, 1999; Chapman, 1988; in severity from minimal (or none) to quite
Cronin, 1994; Haager & Vaughn, 1997; pronounced. Clinical experience with stu-
Marsh & Yeung, 1997a; Purkey & Novak, dents with LD indicates that “children with
1996). Individuals who have a positive sense learning disorders appear to suffer psycho-
of self-worth tend to be happier than others logically more than their peers who do not
(Swann, 1996) and to grapple more success- have learning disorders. Their psychological
fully with failure experiences and other ad- suffering cannot be measured through man-
verse circumstances (Carlock, 1999). Once ifest symptoms alone, as many do not dis-
an individual develops negative self-percep- play such symptoms” (Palombo, 2001, p.
tions, these perceptions can be extremely re- 3). Though few students with LD may re-
sistant to change (Swann, 1996), even when quire clinical treatment for problems relat-
the individual achieves success (Achenbach ing to self-concept, many are affected to
& Zigler, 1963). some degree by the negative perceptions
they hold of themselves as readers, as stu-
Self-concept has particular relevance to dents, or as members of their social group.
students with learning disabilities (LD).
Learning disabilities have been consistently In this chapter, we first provide a brief
linked to poor self-concept (De Francesco & overview of self-concept as it relates to stu-
Taylor, 1985; Kloomok & Cosden, 1994; dents with LD. In doing so, we emphasize
Vaughn & Elbaum, 1999). Children’s expe- the dimensions of self-concept that are espe-
riences in school, particularly in the early cially relevant to students with LD: academ-
grades, can have a powerful influence on ic self-concept, social self-concept, and
their self-perceptions. Difficulties in reading, global self-worth. We then describe the find-
writing, and spelling make students with LD ings and implications of our own research
more vulnerable to failure experiences, in the area of self-concept. This research has

229

230 CAUSES AND BEHAVIORAL MANIFESTATIONS

focused on the impact of specific school fac- The development of multidimensional
tors—identification, educational placement, models has enabled researchers investigat-
and school-based interventions—on the self- ing students with LD to ask important ques-
concept of students with LD. tions beyond whether students with LD
have lower self-concept than do students
Overview without LD. These questions include the
following: In which domains do students
Current conceptualizations of self-concept with LD have self-perceptions that differ
place it in the area of social cognition, from those of their peers without disabili-
which deals with the mental representations ties? How do self-perceptions in specific do-
and processes that underlie self-awareness, mains relate to perceptions of general self-
perspective taking, and the understanding worth? In the next sections, we briefly
of social relations. Social cognition has been review what is known about academic, so-
an area of serious psychological investiga- cial, and global self-concept for students
tion since the 1960s (Lefrancois, 1990), with and without LD.
though the importance of a positive evalua-
tion of oneself has been noted by psycholo- Academic Self-Concept
gists and philosophers for much longer than
that (for a history of the study of self-con- A study by Caslyn and Kenny (1977) that
cept, see Harter, 1996). Within the social included over 500 adolescents revealed that
cognitive perspective, self-concept was de- students’ academic achievement was signifi-
fined early as “the person known to himself, cantly associated with their self-concept. On
particularly the stable, important, and typi- the one hand, students with lower academic
cal aspects of himself as he perceives them” achievement subsequently exhibited lower
(Gordon & Combs, 1958, p. 433). self-concepts than did students with higher
academic achievement, suggesting that low
Though the construction of self involves achievement may be one of the causes of
descriptive and narrative aspects (Palombo, low self-evaluations of competence. At the
2001), it is usually the evaluative aspects same time, low self-esteem itself may lead to
that are at issue in considerations of psycho- lowered expectations for future success
logical adjustment. The term “self-esteem” (Chapman & Boersma, 1979) and dimin-
captures one’s overall sense of self-worth; it ished motivation for academic tasks (McIn-
is, in the words of Carlock (1999), “the way erney, Roche, McInerney, & Marsh, 1997).
you feel about yourself” (p. 3). For the pur- For high school students, self-concept has
pose of this discussion, we consider the term also been demonstrated to be related to stu-
“self-concept” to be synonymous with dents’ subsequent choices of coursework
“self-perception” and “self-esteem.” (Marsh & Yeung, 1997b) and their career
interests (McInerney et al., 1997).
Self-concept is now established as a multi-
dimensional construct, based on the evi- Because self-evaluations are based on
dence that individuals view themselves dif- comparisons we make between our own
ferently across different domains of competencies and those of people around
functioning (Harter, 1985; Marsh, 1989; us, an individual’s self-concept may also de-
Shavelson, Hubner, & Stanton, 1976). That pend on the reference group used as a com-
is, individuals may perceive that they are parison group. (A complementary perspec-
poor performers in some domains, are aver- tive is that individuals compare their own
age in others, and excel in yet others. Harter competencies across different domains
(1985), for example, identified eight do- [Marsh, 1990a].) For example, we perceive
mains of self-perception: general cognitive ourselves more favorably when we compare
competence, peer likeability, behavioral ourselves to others whose performance or
conduct, physical appearance, romantic ap- physical features are less positive than ours.
peal, close friendship, athletic competence, When students with LD compare them-
and job competence. Marsh (1988) further selves to other students with LD, their self-
divided the general domain of academic concept may be different than if they com-
competence into reading, mathematics, and pare themselves to students without LD.
general school competence.

Self-Concept and Students with LD 231

This point is supported by studies that re- suggests that having a strong academic self-
veal that students with LD demonstrated concept may be related to the interest of stu-
lower self-concepts when average- or high- dents with LD in pursuing challenging
achieving peers were used as reference coursework and participating more fully in
groups but not when low-achieving peers the general education curriculum.
were used as the reference group (Haager &
Vaughn, 1995; Vaughn, Haager, Hogan, & Social Self-Concept
Kouzekanani, 1992). Also, because students
with LD often receive most of their educa- One view of the origins and functioning of
tion in the general education classroom, self-concept is based on the idea that human
they are highly aware of how their academic beings have a strong drive to maintain sig-
performance compares with that of their nificant interpersonal relationships (Leary,
classmates (e.g., Cooley & Ayers, 1988; 1999). According to this view, self-concept
Hiebert, Wong, & Hunter, 1982; Kistner, evolved as a mechanism that enables indi-
Haskett, White, & Robbins, 1987; Mont- viduals to monitor the degree to which they
gomery, 1994). When students consistently are valued and accepted by others. Peer ac-
compare themselves to others unfavorably, ceptance is an important index of social ac-
their self-concept is negatively affected. ceptance, and low perceived peer accep-
tance is often associated with low
Renick and Harter (1989) found that per- self-esteem. A positive social self-concept is
ceptions of academic competence for stu- associated with peer acceptance, self-confi-
dents with LD who were placed in regular dence, effective coping, and psychosocial
classrooms were more highly correlated well-being (Bednar, Wells, & Peterson,
with general self-worth than they were with 1989; Harter, 1993; Parker & Asher, 1987).
perceptions of either social acceptance or Students with positive social self-concepts
athletic competence. According to Byrne perceive that others like to be around them
(1996), these results suggest that for chil- and want to have them as friends. In addi-
dren with LD at least, perceptions of how tion, they are able to make and maintain
well they perform academically may have friendships without significant difficulties.
an overriding effect on the extent to which
they like themselves as persons in general. Students who lack a positive social self-
concept are vulnerable to a host of emotion-
In a meta-analysis of studies comparing al, social, and learning problems (Brendtro,
the self-reports of children and adolescents Brokenleg, & Van Bockern, 1990), includ-
with LD to those of their peers without LD ing long-term unhappiness (Bednar et al.,
(Prout, Marcal, & Marcal, 1992), the acad- 1989; Harter, 1993) and low peer accep-
emic self-concept of students with LD was tance (Li, 1985; Vaughn, McIntosh, &
.71 of a standard deviation below that of Spencer-Rowe, 1991). One consistent find-
their peers without disabilities. According ing is the correlation between low self-con-
to Hagborg (1996), 70% of students with cept or self-perception of acceptance and
LD demonstrated significantly lower acade- depression (Leahy, 1985; Wiest, Wong, &
mic self-concepts when compared with Kreil, 1998). Research conducted by Leahy
peers without LD. In a meta-analysis by (1985) suggests that students with low self-
Chapman (1988), the average difference be- concept are more likely to experience de-
tween students with and without LD, with pression.
regard to academic self-concept, was .81 of
a standard deviation, considered quite a As early as kindergarten, same-grade
large effect. classmates perceive students later identified
as having LD as low on social acceptance
The lower self-concept of some students (Vaughn, Hogan, Kouzekanani, & Shapiro,
with LD in the academic area is of signifi- 1990), and this low peer acceptance remains
cant concern in that academic self-concept relatively stable over time (Vaughn &
may mediate students’ accomplishments re- Haager, 1994). Heath and Wiener (1996)
lated to important educational goals. Marsh reported that students with LD who scored
and Yeung (1997b) reported that students high on ratings of depression also rated
with higher self-concepts in particular acad- themselves poorly on self-perceptions of so-
emic areas were more likely to pursue sub-
sequent study in these areas. This finding

232 CAUSES AND BEHAVIORAL MANIFESTATIONS

cial acceptance and that students with LD causes other than low academic achieve-
demonstrated higher levels of depression ment—for example, poor social skills or be-
than did students without LD. The long- havioral difficulties, which themselves may
term impact of peer rejection (Alexander & be a response to the learning disability
Entwisle, 1988; Parker & Asher, 1987) and (Palombo, 2001).
the contribution of peer rejection to depres-
sion (Heath & Wiener, 1996) may operate Global Self-Concept
in part through links to poor social self-con-
cept. Overall, students with LD display lower
perceptions of self-worth than do average
There is diverging evidence regarding the achieving students without disabilities
social self-concept of students with LD. (Vaughn & Elbaum, 1999). In a meta-
Some studies have reported no difference in analysis of studies comparing the self-
the social self-concept of students with and reports of children and adolescents with LD
without LD (Berndt & Burgy, 1996; Clever, to those of their peers without LD (Prout et
Bear, & Juvonen, 1992; Durrant, Cunning- al., 1992), students with LD demonstrated a
ham, & Voelker, 1990; Hagborg, 1999), general self-concept that was .43 of a stan-
whereas others indicate that students with dard deviation below that of their peers.
LD demonstrated lower social self-concept However, this difference is unlikely to be
(Hosley, Hopper, & Gruber, 1998). A study due solely to the academic difficulties of stu-
of students with LD in the fourth and fifth dents with LD. In fact, there is evidence that
grades indicated that participants demon- students’ global self-worth is more influ-
strated lower self-perceptions of social ac- enced by nonacademic factors such as per-
ceptance and global self-concept when com- ceived physical appearance and social ac-
pared to average-achieving students (La ceptance than it is by academic achievement
Greca & Stone, 1990). In contrast, second-, (Cosden, Elliott, Noble, & Kelemen, 1999).
third-, and fourth-grade students with LD Thus, whereas students with LD may be
studied by Bursuck (1989) did not experi- aware of their low academic performance
ence lowered self-concept. In a study of (Grolnick & Ryan, 1990), a poor self-evalu-
more than 100 students with LD, 70 stu- ation of academic performance may not by
dents with behavior disorders (BD), and itself lead to a diminished sense of self-
200 average achievers in grades 9–12 (Har- worth.
ter, Whitesell, & Junkin, 1998), the students
with LD and BD both differed from the av- We now turn to a description of research
erage achievers with respect to their social that we have conducted, over the past
self-perceptions, but only the students with decade, on self-concept and students with
BD had significantly lower self-perceptions LD. The main goal of our research program
of conduct. Both groups with disabilities ex- has been to investigate the extent to which
hibited lower scores on global self-worth school factors, such as identification with a
than did average achievers without disabili- learning disability and educational place-
ties. In a study by Hagborg (1998), a small- ment, affect the self-concept of students
er sample of high school students with LD with LD. Another important goal has been
did not differ from their peers without LD to investigate the extent to which school-
on this dimension. based interventions can ameliorate the self-
concept of students with LD.
The link between poor academic perfor-
mance and low social self-concept is not Identification
clear. There is some evidence for a relation
between students’ achievement and their so- Some researchers and advocates have ar-
cial status (Wentzel & Erdley, 1993). How- gued that being identified as having a dis-
ever, in the study by La Greca and Stone ability is itself detrimental to a student’s
(1990), low-achieving students who were self-concept (Brophy & Good, 1970; Good,
not identified as having LD perceived their 1982). Vaughn and colleagues (1992) fol-
social acceptance more positively than did lowed students from kindergarten through
students with LD. Thus, the low social ac-
ceptance and low social self-concept of
some students with LD may be related to

Self-Concept and Students with LD 233

fourth grade, assessing self-concept each gories. These categories, and the number of
year. Students identified as students with LD comparisons coded into each, were regular
at the end of second grade did not differ class versus resource room (16); regular
from students not so identified with regard class versus self-contained classroom (18);
to self-concept. Other research that com- resource room versus self-contained class-
pared adults who had or had not received room (26); self-contained classroom versus
special education services found no differ- special school (3); and regular class versus
ence in self-concept (Lewandowki & Arcan- special school (2).
gelo, 1994)
The results of the meta-analysis indicated
Whereas the popular perception is that that there was no reliable association be-
being labeled as having a learning disability tween self-concept and educational place-
may lead to feelings of shame or humilia- ment for any of the major comparison cate-
tion, and hence to low self-concept, there is gories. The one exception, represented by
as yet no solid empirical evidence that iden- three samples of students from a single
tification with a learning disability results in study (Butler & Marinov-Glassman, 1994),
a diminished sense of self-worth. Indeed, for was that students with LD who received in-
some children, the knowledge that their struction in self-contained classrooms in
reading or other academic difficulties are re- regular schools exhibited lower self-concept
lated to a disabling condition, and not to compared to students attending special
low intelligence or lack of effort, may help schools. The findings of the meta-analysis
sustain positive perceptions of general intel- suggest that educational placement is not
lectual competence and self-worth. the overriding determinant of self-concept
students with LD. Rather, other factors op-
Placement erating within each context—for example,
individual teachers’ understanding and ac-
Over the past 25 years, there has been con- ceptance of students with disabilities—may
siderable debate surrounding the placement have greater influence on the way students
of students with LD in general education with LD feel about themselves in specific
classrooms for most, or all, of their instruc- classroom settings. For example, Chapman
tion. One of the arguments for general edu- (1988) reported a relation between chil-
cation placement for all of a student’s in- dren’s academic self-concepts and teachers’
struction has been that students with LD feedback and level of academic achieve-
placed in regular classrooms fare better, in ment.
terms of social acceptance, friendship rela-
tions, and self-concept, than students with Social support also influences the social
LD educated in more segregated settings self-concept of students with LD. Marsh
(Vaughn, Elbaum, & Boardman, 2001). (1990b) notes that self-perceptions “are
Predictions based on social comparison the- formed through experience with and inter-
ory would suggest that students with LD pretations of one’s environment. They are
have higher self-concept in special educa- especially influenced by evaluations by sig-
tion settings (resource rooms and self-con- nificant others, reinforcement, and attribu-
tained classrooms), where all students expe- tions for one’s own behavior” (p. 27). For-
rience similar academic challenges. man (1988) reported that students with LD
However, empirical studies have shown who had higher levels of perceived social
mixed results. support—particularly support from class-
mates—demonstrated higher self-concept
To better understand the contrasting em- regardless of school placement (resource
pirical findings, Elbaum (2002) conducted a room or self-contained special education
meta-analysis of studies that compared the setting).
self-concept of groups of students with LD
in more and less restrictive settings. A total School-Based Interventions
of 38 studies published between 1975 and
1999 were located, yielding 65 placement In light of the research indicating the many
comparisons. Each comparison was coded adverse concomitants and consequences of
into one of five placement comparison cate- negative self-concept, considerable work

234 CAUSES AND BEHAVIORAL MANIFESTATIONS

has been focused on developing interven- other data were provided, we assumed an
tions to improve the self-concept of students effect size of 0.
with LD. Elbaum and Vaughn (2001) used
meta-analysis to determine the overall effec- Findings
tiveness of interventions on the self-concept
of students with LD. Sixty-four studies were Aggregated across all categories of interven-
located that met the following criteria: (1) tion, the mean intervention effect was quite
the majority of the participants were stu- modest, d = 0.19 (where d symbolizes the
dents with LD, (2) a measure of self-concept mean weighted effected size, interpreted in
was used as one of the outcome measures, standard deviation units [Cooper, 1998]).
(3) the intervention took place in a school When outcomes of interventions across all
setting, (4) the study included both a treat- grade levels were aggregated, intervention
ment and a comparison group, (5) the study type (e.g., counseling, academic, physical)
was published or available between 1975 was not reliably associated with interven-
and 1997, and (6) sufficient data were pro- tion effect sizes. The mean weighted effect
vided to calculate an effect size. sizes for different categories of intervention
ranged from d = 0.12 to d = 0.31.
Types of Interventions
However, differences in outcomes were
Most, but not all, of the interventions in- found to be reliably associated with stu-
cluded in the meta-analysis were designed dents’ grade level. The mean weighted effect
for the primary purpose of enhancing the size for adolescents (d = 0.42) was signifi-
self-concept of students with LD. Others cantly higher than that for elementary (d =
were designed primarily to accomplish an- 0.12) and high school students (d = 0.17).
other goal, such as improved academic skills Within grade groupings, different types of
or physical abilities; however, the re- interventions were found to be most effec-
searchers hypothesized that students partici- tive. For elementary students, only academ-
pating in the intervention would also evi- ic interventions yielded effect sizes that were
dence higher self-concept than control reliably different from 0 (d = 0.17); for mid-
students. Thus, the interventions ranged dle and high school students, this was true
from what would be considered traditional only of counseling interventions (d = 0.61
counseling groups to cooperative learning and d = 0.32, respectively)
curricula, fitness programs, and so on. For
purposes of analysis, we classified the inter- Based on the earlier review of the acade-
ventions into six general categories based on mic and social self-concept of students with
the focus of the intervention. The categories LD, we were interested in the extent to
were counseling (33 studies), academic (18 which interventions had a differential im-
studies), physical (5 studies), reinforcement pact on the academic and social domains of
(5 studies), sensory/perceptual (2 studies), self-concept. Interventions had the greatest
and “other” (4 studies). The “other” cate- impact on academic self-concept (d = 0.28),
gory consisted of interventions using music, followed by social self-concept (d = 0.18)
arts and crafts, education plans provided to and general self-concept (d = 0.15).
the teacher, and a home-to-school facilita-
tor. Analyses also revealed that effect sizes
were not influenced by the self-concept
Calculation of Effect Sizes measure used. The two most frequently
used measures of self-concept, the Piers–
Effect sizes were calculated as the mean of Harris Children’s Self-Concept Scale (PHC-
the treatment group posttest score minus SCS; Piers, 1984) and the Self-Esteem Inven-
the mean of the comparison group posttest tory (SEI; Coopersmith, 1986) yielded
score divided by the pooled standard devia- almost identical effect sizes, d = 0.21 and d
tion. When means and standard deviations = 0.22, respectively.
were not available, effect sizes were estimat-
ed from t, F, or p values. When statistical In a subsequent study, Elbaum and
tests were reported as nonsignificant and no Vaughn (in press) used a subset of the inter-
vention studies from the previously de-
scribed meta-analysis to investigate whether
intervention effectiveness was associated
with the level of self-concept that students

Self-Concept and Students with LD 235

demonstrated prior to intervention. The esteem is that these interventions “change
subset of studies consisted of those that people’s perceptions of the degree to which
used the PHCSCS and provided both pre- they are socially valued individuals. Self-
and postintervention scores for the interven- esteem programs always include features
tion group. The 20 groups of students with that would be expected to increase real or
LD for whom these data were available perceived social acceptance, for example,
were divided into those whose self-concept these programs include components aimed
scores, prior to intervention, were high (at at enhancing social skills and interpersonal
or above the 75th percentile for this sam- problem solving, improving physical appear-
ple), low (at or below the 25th percentile), ance, and increasing self-control” (Leary,
and midlevel (the middle 50%). Comparing 1999, p. 35). This view accords with litera-
the average scores for these groups to the ture on the social functioning of students
mean normative score reported for the with LD which suggests that many students
PHCSCS (M = 51.84; Piers, 1984), the high with LD demonstrate overall low social skills
self-concept groups had an average score (e.g., Foss, 1991; Jarvis & Justin, 1992;
0.45 standard deviations above the norma- Kavale & Forness, 1996; Merrell, 1991).
tive mean (57.74), the middle groups had an
average score almost exactly identical to the At the same time, the findings suggest
normative mean (51.78), and the low that students with LD who have average-to-
groups had an average score approximately high levels of self-concept do not benefit
1.3 standard deviations below the mean from efforts to further enhance their self-
(34.36). When outcomes for these groups esteem. Indeed, including such students in
were compared meta-analytically, there was interventions that do not have an academic
a statistically reliable association between component, that focus exclusively on self-
self-concept level prior to intervention and concept, and that use time during which
intervention effect size. Groups of students students would otherwise be engaged in in-
with high self-concept prior to intervention structional activities, may actually be to
gained an average of 14 points (d = 1.22); their detriment in the long run.
students with midlevel self-concept gained
an average of 4 points (d = 0.29); and Summary of Findings on the Self-
groups of students with low self-concept Concept of Students with LD
prior to intervention gained an average of 3
points (d = 0.23). An analysis of residual- Students with LD often demonstrate lower
ized gain scores suggested that the observed academic self-concept than do normally
results could not be completely explained as achieving students without disabilities, and
an artifact of regression to the mean. sometimes demonstrate lower self-concepts
in the social domain. In addition, students
Taken together, the analyses of interven- with LD may demonstrate low perceptions
tion outcomes suggest that students with of general self-worth. We do not yet have a
LD who have truly low self-concept can clear understanding of why some students
benefit considerably from appropriate inter- with lower academic self-concept also have
ventions. For these students, the most effec- lower general self-perceptions, whereas oth-
tive interventions, as delineated earlier, may ers do not. It may be that a combination of
differ according to students’ age. The fact low self-evaluations across multiple do-
that the most effective interventions for mains—academic, social, and physical, for
younger students appear to be academic in- example—is a better predictor of low self-
terventions suggests that improving these worth than poor academic self-concept
students’ academic skills can have a collat- alone.
eral effect on their self-perceptions. In-
creased self-efficacy in the academic domain It is equally important to note that many
may confer a sense of empowerment that re- students with LD have self-concept scores
sults in more positive self-evaluations. that are on par with those of students with-
out disabilities (or, in some cases, even high-
For older students, the theory proposed by er; cf. Bear & Minke, 1996; Clever et al.,
Leary (1999) may be especially relevant. In 1992; Kistner et al., 1987; Kistner & Os-
this view, the explanation for the beneficial borne, 1987). Moreover, even when the self-
effects of programs that enhance self-

236 CAUSES AND BEHAVIORAL MANIFESTATIONS

concept scores of students with LD are be- Hence, considerable caution must be used
low those of students without disabilities, when comparing subscale scores of differ-
they may still be within the normal range. ent groups of students, or when comparing
Scores on measures of general self-worth the same students’ scores on different sub-
which are substantially below those of typi- scales.
cal students may indicate problems that re-
quire specialized attention. Students with Another concern has to do with the accu-
LD who experience low perceptions of self- racy of measuring change in self-concept
worth may or may not be those who are ex- over time. Again using the PHCSCS as an
periencing the greatest academic difficulties. example, test–retest reliabilities for the total
Only by assessing students’ self-perceptions scale are in the .86–.96 range for intervals
in different domains is it possible to reach a of 3–4 weeks, but in the .42–.51 range for
more complete understanding of the source intervals of 8–12 months; the median re-
of overall low self-concept and to provide ported test–retest reliability is .725. For the
appropriate interventions. studies we synthesized, the median duration
of self-concept interventions was 10 weeks;
Issues Related to the Measurement test–retest reliability of the PHCSCS over
of Self-Concept this interval is likely to be considerably
lower than the recommended level of .9.
Research on the self-concept of students This is problematic from an analytic stand-
with LD, like other self-concept research point, in that the lower the reliability of the
conducted since the 1970s, has benefited measurement instrument, the greater the
from the availability of self-report question- amount of error, and, in pre–post designs,
naires either designed from the outset (e.g., the greater the likelihood of artifacts due to
Bracken, 1992; Marsh, 1988, 1990b) or re- regression to the mean (Campbell & Kenny,
vised (e.g., Piers, 1984) to reflect evolving 1999). The problems illustrated here are
theory regarding the multidimensional na- not unique to the PHCSCS; similar con-
ture of self-concept. However, as document- cerns could be adduced with regard to
ed in considerable detail by Keith and many other commonly used self-concept in-
Bracken (1996), many widely used instru- struments.
ments lack a compelling theoretical founda-
tion and/or evidence of strong technical ad- A third issue with regard to the measure-
equacy. For example, the PHCSCS, which ment of self-concept in students with LD
was used more frequently than any other has to do with whether measures normed
measurement instrument in the intervention on samples of students without disabilities
studies we synthesized, was originally de- function similarly for students with LD.
veloped as a unidimensional measure of Only by comparing the performance of stu-
self-concept, albeit with content drawn dents with and without LD on the same in-
from different domains. According to Keith strument is it possible to verify whether stu-
and Bracken, the items were subsequently dents from these two populations respond
assigned to one or more cluster scales based similarly to item content and whether the
on factor analysis; however, the cluster subscales operate similarly for both popula-
items are not mutually exclusive (i.e., some tions. However, most instrument developers
items loaded on two or more factors), did not include students with LD in their
hence the interpretation of subscales as rep- original norming samples. The only current-
resenting discrete domains is somewhat ly available instrument that was specifically
hazardous. Moreover, whereas the internal designed to assess both students with LD
consistency of the PHCSCS total scale is ad- and students without disabilities is the Self-
equate (.88 to .93 for girls and boys in Perception Profile for Learning Disabled
grades 6 and 10), the internal consistency Students (SPPLD; Renick & Harter, 1988).
of the subscales is much lower (.73–.81). According to Keith and Bracken (1996), the
This means that the subscales are much less authors of the SPPLD developed its five aca-
reliable than the measure as a whole. demic subscales based on the finding that
students with LD responded differentially to
items that had all been part of a single acad-
emic subscale in an earlier instrument. That

Self-Concept and Students with LD 237

is, the newer instrument was developed grades received on school assignments, or
based on the evidence suggesting that com- being included or not included in the self-
pared to students without LD, students with constituting clusters that characterize the
LD had more strongly differentiated percep- social networks of students in schools
tions of their competence in different acade- (Vaughn et al., 2001).
mic domains. This fits with the typical pro-
file of students with LD as students with Third, intervention researchers should
severe deficits in reading (and sometimes gather and report much more comprehen-
also in mathematics) but not in other areas sive data on the students with LD who par-
of the curriculum insofar as these are not ticipate in intervention studies. In particular,
reading dependent. it would be extremely useful to know stu-
dents’ individual placement histories and
More recently developed measures of self- current level of academic performance (e.g.,
concept appear to address at least some of grade level in reading), as well as contextual
these concerns. For example, instruments factors such as the range of service delivery
developed by Marsh (1988, 1990b) and options used at the school(s). As well, where
Bracken (1992) provide evidence of high re- older children are concerned, it would be il-
liability as well as construct and concurrent luminating to obtain students’ own percep-
validity. Marsh’s instruments may prove to tions of the usefulness of self-concept inter-
be particularly useful in investigations in- ventions in which they have participated.
volving students with LD in that they assess
self-perceptions in three separate academic Fourth, when selecting students with LD
domains—Reading, Mathematics, and Gen- for an intervention aimed at improving their
eral School—and thus allow for a more nu- self-concept, it is essential to determine first
anced understanding of students’ academic whether they are likely to benefit from such
self-concepts. intervention. Palombo (2001) cautions that
with regard to problems of the self experi-
Implications for Future Research enced by students with LD, there is no clear
and Intervention correlation between diagnosis and treatabil-
ity. That is, there is more to the determina-
The first implication we draw for future re- tion of likely benefit than the severity of the
search on the self-concept of students with unease experienced by the child. In dis-
LD is that researchers need to be extremely cussing the advisability of individual thera-
mindful of the issues discussed earlier sur- py for self-esteem problems, Palombo
rounding the measurement of self-concept. recommends consideration of the “psycho-
In addition, given that self-concept is almost logical mindedness” of the child, that is, the
always assessed by means of written self- ability of the child to think about his or her
report instruments, consideration needs to feelings and the relation between feelings,
be given to the reading level of the instru- attitudes, and behaviors. This advice would
ment and how it is administered. Reading appear to be equally relevant to school-
the items aloud may be necessary to ensure based interventions based on group therapy
adequate comprehension by all students. models. Moreover, educators should realize
that for some students, participating in a
Second, careful longitudinal studies are group that deals explicitly with painful per-
needed to investigate the extent to which sonal issues may not be in the student’s best
students’ self-concept may change in rela- interest. In addition, Baumeister, Smart, and
tion to their academic progress and to Boden (1996) caution that when self-
changes in their educational context (e.g., appraisals are overly (unrealistically) inflat-
transitions between schools and changes in ed, the result may be increased vulnerability
placement). In some developmental stages, to ego threats and increased evaluative de-
particularly early adolescence, students’ pendency on others. Neither of these out-
self-perceptions may range from high to low comes would be desirable for any students,
in the same day, depending on the valence especially students with persistent learning
of the day’s everyday events—interactions difficulties.
with individual friends and family members,
Finally, Ellis (1998) argues that an impor-
tant means for contributing to the self-con-

238 CAUSES AND BEHAVIORAL MANIFESTATIONS

cepts of adolescents with LD is to provide Bear, G. G., Minke, K. M., Griffin, S. M., &
them opportunities to control their destinies Deemer, S. A. (1998). Achievement-related per-
and positively influence others. He suggests ceptions of children with learning disabilities and
that adolescents with LD require education- normal achievement: Group and developmental
al environments that challenge them and differences. Journal of Learning Disabilities,
provide personally meaningful work. Fur- 31(1), 91–104.
thermore, teachers can have a significant
impact on students by providing positive Bednar, R. L., Wells, M. G., & Peterson, S. R.
feedback (Bear, Minke, Griffin, & Deemer, (1989). Self-esteem: Paradoxes and innovations
1998). Appropriate, positive feedback is an in clinical theory and practice. Washington, DC:
easily administered intervention that class- American Psychological Association.
room observations reveal is used infrequent-
ly by teachers (McIntosh, Vaughn, Schumm, Berndt, T. J., & Burgy, L. (1996). Social self-con-
Haager, & Lee, 1993). cept. In B. A. Bracken (Ed.), Handbook of self-
concept (pp. 171–209). New York: Wiley.
Conclusion
Bracken, B. A. (1992). Multidimensional Self Con-
According to Palombo (2001), improve- cept Scale. Austin, TX: Pro-Ed.
ments in the feelings students with LD have
about themselves “occur when they can Brendtro, L. K., Brokenleg, M., & Van Bockern, S.
function adequately academically, have ex- (1990). Reclaiming youth at risk: Our hope for
perienced real life successes, and have devel- the future. Bloomington, IN: National Educa-
oped a good understanding of the reading tional Service.
disability. Children who reach this point can
go on to be their own advocates as they Brophy, J., & Good, T. L. (1970). Teachers’ com-
progress within the educational system” (p. munication of differential expectations for chil-
134). Educators’ main responsibility is to dren’s classroom performance: Some behavioral
ensure that students with LD accomplish data. Journal of Educational Psychology, 20,
the first of these goals, that is, that they de- 941–952.
velop the ability to function adequately in
school. The judicious use of adjunct ser- Bursuck, W. (1989). A comparison of students with
vices, including school-based interventions learning disabilities to low achieving and high
for some students, may further assist stu- achieving students on three dimensions of social
dents with LD to develop and maintain the acceptance. Journal of Learning Disabilities,
positive self-perceptions that enable them to 22(3), 188–194.
be their own best advocates.
Butler, R., & Marinov-Glassman, D. (1994). The
References effects of educational placement and grade level
on the self-perceptions of low achievers and stu-
Achenbach, T., & Zigler, E. (1963). Social compe- dents with learning disabilities. Journal of Learn-
tence and self-image disparity in psychiatric and ing Disabilities, 27, 325–334.
nonpsychiatric patients. Journal of Abnormal
and Social Psychology, 67, 197–205. Byrne, B. M. (1996). Academic self-concept: Its
structure, measurement, and relation to academ-
Alexander, K. L., & Entwisle, D. R. (1988). ic achievement. In B. A. Bracken (Ed.), Hand-
Achievement in the first 2 years of school: Pat- book of self-concept (pp. 287–316). New York:
terns and processes. Monographs of the Society Wiley.
for the Research in Child Development, 53(2),
157. Campbell, D. T., & Kenny, D. A. (1999). A primer
on regression artifacts. New York: Guilford
Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Press.
Relation of threatened egotism to violence and
aggression: The dark side of high self-esteem. Carlock, C. J. (Ed.). (1999). Enhancing self-esteem
Psychological Review, 103(1), 5–33. (3rd ed.). Philadelphia: Taylor & Francis.

Bear, G. G., & Minke, K. M. (1996). Positive bias in Caslyn, R. J., & Kenny, D. A. (1977). Self-concept
maintenance of self-worth among children with of ability and perceived evaluation of others:
LD. Learning Disability Quarterly, 19, 23–32. Cause or effect of academic achievement. Journal
of Educational Psychology, 69(2), 136–145.

Chapman, J. W. (1988). Learning disabled chil-
dren’s self-concepts. Review of Educational Re-
search, 58(3), 347–371.

Chapman, J. W., & Boersma, F. J. (1979, April).
Self-perceptions of ability, expectations and locus
of control in elementary learning disabled chil-
dren. Paper presented at the annual meeting of
the American Educational Research Association.
(ERIC Document Reproduction Service No. ED
169 738)

Clever, A., Bear, G. G., & Juvonen, J. (1992). Dis-
crepancies between competence and importance
in self-perceptions of children in integrated class-
rooms. Journal of Special Education, 26,
125–138.

Cooley, E. J., & Ayers, R. R. (1988). Self-concept

Self-Concept and Students with LD 239

and success-failure attributions of nonhandi- ities. In D. Chard, E. J. Kameenui, & J. W. Lloyd
capped students and students with learning dis- (Eds.), Issues in educating students with learning
abilities. Journal of Learning Disabilities, 21(3), disabilities (pp. 129–152). Hillsdale, NJ: Erl-
174–178. baum.
Cooper, H. (1998). Synthesizing research (3rd ed.). Hagborg, W. J. (1996). Self-concept and middle
Thousand Oaks, CA: Sage. school students with learning disabilities: A com-
Coopersmith, S. A. (1986). Self-esteem inventories. parison of scholastic competence subgroups.
Palo Alto, CA: Consulting Psychologists Press. Learning Disability Quarterly, 19, 117–126.
Cosden, M., Elliott, K., Noble, S., & Kelemen, E. Hagborg, W. J. (1998). School membership among
(1999). Self-understanding and self-esteem in students with learning disabilities and nondis-
children with learning disabilities. Learning Dis- abled students in a semirural high school. Psy-
ability Quarterly, 22(4), 279–290. chology in the Schools, 35(2), 183–188.
Cronin, E. M. (1994). Helping your dyslexic child: Hagborg, W. J. (1999). Scholastic competence sub-
A guide to improving your child’s reading, writ- groups among high school students with learning
ing, spelling, comprehension, and self-esteem. disabilities. Learning Disability Quarterly, 22(1),
Roseville, CA: Prima. 3–10.
De Francesco, J. J., & Taylor, J. (1985). Dimensions Harter, S. (1985). Manual for the Self-Perception
of self-concept in primary and middle school Profile for Children. Denver, CO: University of
learning disabled and nondisabled students. Denver.
Child Study Journal, 15, 99–105. Harter, S. (1993). Causes and consequences of low
Durant, J. E., Cunningham, C. E., & Voelker, S. self-esteem in children and adolescents. In R.
(1990). Academic, social and general self-con- Baumeister (Ed.), Self-esteem: The puzzle of
cepts of behavioral subgroups of learning dis- low self-regard (pp. 18–37). New York: Plenum
abled children. Journal of Educational Psycholo- Press.
gy, 82, 657–663. Harter, S. (1996). Historical roots of contemporary
Elbaum, B. (2002). The self-concept of students issues involving self-concept. In B. A. Bracken
with learning disabilities: A meta-analysis of (Ed.), Handbook of self-concept (pp. 1–37). New
comparisons across different placements. Learn- York: Wiley.
ing Disabilities Research & Practice, 17(4), Harter, S., Whitesell, N. R., & Junkin, L. J. (1998).
216–226. Similarities and differences in domain-specific
Elbaum, B., & Vaughn, S. (2001). School-based in- and global self-evaluations of learning-disabled,
terventions to enhance the self-concept of stu- behaviorally disordered, and normal achieving
dents with learning disabilities: A meta-analysis. adolescents. American Educational Research
Elementary School Journal, 101(3), 303–329. Journal, 35(4), 653–680.
Elbaum, B., & Vaughn, S. (in press). For which stu- Heath, N. L., & Wiener, J. (1996). Depression and
dents with learning disabilities are self-concept nonacademic self-perceptions in children with
interventions effective? Journal of Learning Dis- and without learning disabilities. Learning Dis-
abilities. ability Quarterly, 19, 34–44.
Ellis, E. S. (1998). Watering up the curriculum for Hiebert, B., Wong, B., & Hunter, M. (1982). Affec-
adolescents with learning disabilities—Part 2. Re- tive influences on learning disabled adolescents.
medial and Special Education, 19(2), 91–105. Learning Disability Quarterly, 5, 334–343.
Forman, E. A. (1988). The effects of social support Hosley, M., & Hopper, C., & Gruber, M. B. (1998).
and school placement on the self-concept of LD Self-concept and motor performance of children
students. Learning Disability Quarterly, 11, with learning disabilities. Perceptual and Motor
115–124. Skills, 87, 859–862.
Foss, J. M. (1991). Nonverbal learning disabilities Jarvis, P. A., & Justin, E. M. (1992). Social sensitiv-
and remedial interventions. Annals of Dyslexia, ity in adolescents and adults with learning dis-
41, 128–140. abilities. Adolescence, 27(108), 977–988.
Good, T. L. (1982). How teachers’ expectations af- Kavale, K., & Forness, S. R. (1996). Social skill
fect results. American Education, 18(10), 25–32. deficits and learning disabilities: A meta-analysis.
Gordon, I. J., & Combs, A. W. (1958). The learner: Journal of Learning Disabilities, 29(3), 226–237.
Self and perception. Review of Educational Re- Keith, L. K., & Bracken, B. A. (1996). Self-concept
search, 28, 433. instrumentation: A historical and evaluative re-
Grolnick, W. S., & Ryan, R. M. (1990). Self-percep- view. In B. A. Bracken (Ed.), Handbook of self-
tions, motivation, and adjustment in children concept (pp. 91–170). New York: Wiley.
with learning disabilities: A multiple group com- Kistner, J. A., Haskett, M., White, K., & Robbins,
parison. Journal of Learning Disabilities, 23, R. (1987). Perceived competence and self-worth
177–184. of LD and normally achieving students. Learning
Haager, D., & Vaughn, S. (1995). Parent, teacher, Disability Quarterly, 10, 37–44.
peer and self-reports of the social competence of Kistner, J. A., & Osborne, M. (1987). A longitudi-
students with learning disabilities. Journal of nal study of learning disabled children’s self-eval-
Learning Disabilities, 28(4), 205–215. uations. Learning Disability Quarterly, 12,
Haager, D., & Vaughn, S. (1997). Assessment of so- 133–140.
cial competence in students with learning disabil- Kloomok, S., & Cosden, M. (1994). Self-concept in

240 CAUSES AND BEHAVIORAL MANIFESTATIONS

children with learning disabilities: The relation- dren with learning disabilities: Observer-child
ship between global self-concept, academic “dis- concordance across six context-dependent do-
counting,” nonacademic self-concept, and per- mains. Journal of Learning Disabilities, 27(4),
ceived social support. Learning Disability 254–262.
Quarterly, 17, 140–153. Palombo, J. (2001). Learning disorders and disor-
La Greca, A., & Stone, W. (1990). Children with ders of the self in children and adolescents. New
learning disabilities: The role of achievement in York: Norton.
their social, personal, and behavioral function- Parker, J. G., & Asher, S. R. (1987). Peer relations
ing. In H. L. Swanson & B. Keogh (Eds.), Learn- and later personal adjustment: Are low-accepted
ing disabilities: Theoretical and research issues children at risk? Psychological Bulletin, 102,
(pp. 333–352). Hillsdale, NJ: Erlbaum. 357–389.
Leahy, R. L. (Ed.). (1985). The development of the Piers, E. V. (1984). The Piers-Harris Children’s Self-
self. Orlando, FL: Academic Press. Concept Scale. Nashville, TN: Counselor
Leary, M. (1999). Making sense of self-esteem. Cur- Recordings and Tests.
rent Directions in Psychological Science, 8, Prout, H. T., Marcal, S. D., & Marcal, D. C.
32–35. (1992). A meta-analysis of self-reported person-
Lefrancois, G. R. (1990). The lifespan (3rd ed.). ality characteristics of children and adolescents
Belmont, CA: Wadsworth. with learning disabilities. Journal of Psychoedu-
Lewandowski, L., & Arcangelo, K. (1994). The so- cational Assessment, 10(1), 59–64.
cial adjustment and self-concept of adults with Purkey, W. W., & Novak, J. M. (1996). Inviting
learning disabilities. Journal of Learning Disabil- school success: A self-concept approach to teach-
ities, 27(9), 598–605. ing, learning, and democratic practice (3rd ed.).
Li, A. K. F. (1985). Early rejected status and later Belmont, CA: Wadsworth.
social adjustments: A 3-year follow-up. Journal Renick, M. J., & Harter, S. (1988). Self-Perception
of Abnormal Child Psychology, 13, 567–577. Profile for Learning Disabled Students. Denver,
Marsh, H. W. (1988). Self-Description Question- CO: University of Denver Press.
naire, I. San Antonio, TX: Psychological Corpo- Renick, M. J., & Harter, S. (1989). Impact of social
ration. comparisons on the developing self-perceptions
Marsh, H. W. (1989). Age and sex effects in multi- of learning disabled students. Journal of Educa-
ple dimensions of self-concept: Preadolescence to tional Psychology, 81, 631–638.
adulthood. Journal of Educational Psychology, Shavelson, R. J., Hubner, J. J., & Stanton, G. C.
81, 417–430. (1976). Self-concept: Validation of construct in-
Marsh, H. W. (1990a). The influence of internal terpretations. Review of Educational Research,
and external frames of reference on the forma- 46, 407–441.
tion of math and English self-concepts. Journal of Swann, W. B. (1996). Self-traps: The elusive quest
Educational Psychology, 82, 107–116. for higher self-esteem. New York: Freeman.
Marsh, H. W. (1990b). Self-Description Question- Twain, M. (1929). What is man? and other essays.
naire, II. San Antonio, TX: Psychological Corpo- New York: Harper & Brothers.
ration. Vaughn, S., & Elbaum, B. E. (1999). The self-con-
Marsh, H. W., & Yeung, A. S. (1997a). Causal ef- cept and friendships of students with learning
fects of academic self-concept on academic disabilities: A developmental perspective. In R.
achievement: Structural equation models of lon- Gallimore, L. Bernheimer, D. L. MacMillan, D.
gitudinal data. Journal of Educational Psycholo- L. Speece, & S. Vaughn (Eds.), Developmental
gy, 89(1), 41–54. perspectives on children with high incidence dis-
Marsh, H. W., & Yeung, A. S. (1997b). Course- abilities (pp. 81–110). Mahwah, NJ: Erlbaum.
work selection: Relations to academic self-con- Vaughn, S., Elbaum, B., & Boardman, A. G.
cept and achievement. American Educational Re- (2001). The social functioning of students with
search Journal, 34(4), 691–720. learning disabilities: Implications for inclusion.
McInerney, D. M., Roche, L. A., McInerney, V., & Exceptionality, 9(1), 49–67.
Marsh, H. W. (1997). Cultural perspectives on Vaughn, S., & Haager, D. (1994). Social compe-
school motivation: The relevance and application tence as a multifaceted construct: How do stu-
of goal theory. American Educational Research dents with learning disabilities fare? Learning
Journal, 34(1), 207–236. Disability Quarterly, 17(4), 253–266.
McIntosh, R., Vaughn, S., Schumm, J., Haager, D., Vaughn, S., Haager, D., Hogan, A., &
& Lee, O. (1993). Observations of students with Kouzekanani, K. (1992). Self-concept and peer
learning disabilities in general education class- acceptance in students with learning disabilities:
rooms. Exceptional Children, 60(3), 249–261. A four- to five-year prospective study. Journal of
Merrell, K. W. (1991). Teacher ratings of social Educational Psychology, 84(1), 43–50.
competence and behavioral adjustment: Differ- Vaughn, S., Hogan, A. Kouzekanani, K., &
ences between learning-disabled, low-achieving, Shapiro, S. (1990). Peer acceptance, self-percep-
and typical students. Journal of School Psycholo- tions, and social skills of learning disabled stu-
gy, 29(3), 207–217. dents prior to identification. Journal of Educa-
Montgomery, M. S. (1994). Self-concept and chil- tional Psychology, 82(1), 101–106.

Self-Concept and Students with LD 241

Vaughn, S., McIntosh, R., & Spencer-Rowe, J. and peer acceptance in early adolescence. Devel-
(1991). Peer rejection is a stubborn thing: In- opmental Psychology, 29(5), 819–826.
creasing peer acceptance of rejected students with Wiest, D., Wong, E. H., & Kreil, D. A. (1998). Pre-
learning disabilities. Learning Disabilities Re- dictors of global self-worth and academic perfor-
search and Practice, 6(2), 65–132. mance among regular education, learning dis-
abled, and continuation high school students.
Wentzel, K. R., & Erdley, C. A. (1993). Strategies Adolescence, 33(131), 601–618.
for making friends: Relations to social behavior

15

Neurological Correlates
of Reading Disabilities

Carlin J. Miller
Juliana Sanchez
George W. Hynd

Learning disabilities are a heterogeneous learning to read (Kibby & Hynd, 2001).
group of behaviorally diagnosed disorders Hinshelwood suggested that an abnormali-
thought to have a negative impact on the ty of or damage to the angular and supra-
learning of an estimated 3–6% of all school- marginal gyri on the left side of the brain
age children (Kibby & Hynd, 2001). Diag- might manifest itself by difficulty in learn-
nosis assumes adequate intelligence, intact ing to read. Incredibly, these same areas of
sensory systems, and the absence of a handi- the brain are the focus of research on the
capping condition or environment that biological basis of reading disabilities to-
would cause a person to have significant dif- day.
ficulty learning (American Psychiatric Asso-
ciation, 2000). Although learning disabili- Reading disabilities are primarily charac-
ties exist in many domains, it is the area of terized by difficulties with reading and
severe reading disabilities, also known as de- spelling, but also include difficulties with
velopmental dyslexia, that has been the fo- phonemic segmentation (Hynd, Semrud-
cus of the majority of research and in which Clikeman, Lorys, Novey, & Eliopulos,
the neurobiological basis of the disorder is 1990), rapid and automatic recognition and
best understood. decoding of single words (Lyon, 1996), ar-
ticulation (Huettner, 1994), sensorimotor
The neurobiological basis of learning coordination (Zeffiro & Eden, 2001), and
problems has been the subject of research anomia (Temple, 1997). Although there is a
for over a century. In the late 19th century, high comorbidity with attention-deficit/
the lateralization of language to the left hyperactivity disorder (ADHD; Barkley,
hemisphere and the localization of the lan- 1997), the two appear to be unique disor-
guage areas in the brain were beginning to ders that often co-occur each with a poten-
be explored by Broca and Wernicke, among tially different etiology (Barkley, 1997;
others (Kral, Nielson, & Hynd, 1998). Hin- Hynd et al., 1990).
shelwood (1900), among others, first pro-
posed the idea that damage to or variation In the last decade, a consensus has been
in brain development of these cortical lan- reached among most researchers in the field
guage areas might lead to difficulties in that the core deficit in reading disabilities is
difficulty with phonological processing

242

Neurological Correlates of Reading Disabilities 243

(Huettner, 1994; Hynd et al., 1995; Lyon, temporale, and the perisylvian cortex, all of
1996; Shaywitz & Shaywitz, 1999; Siegel, which are thought to be involved in the pro-
1993). Although most researchers agree cessing of language (Galaburda, 1993;
that phonological processing is a core deficit Hynd, Hynd, Sullivan, & Kingsbury, 1987;
of this disorder, evidence remains that visual Hynd et al., 1990; Hynd & Semrud-Clike-
deficits also exist in some people with read- man, 1989; Leonard et al., 1993; Riccio &
ing disabilities (Eden, Van Meter, Rumsey, Hynd, 1996; Temple, 1997; Zeffiro &
& Zeffiro, 1996). There has been some de- Eden, 2001). The magnocellular pathway of
bate in the literature as to whether or not the visual system (including the lateral
orthographic processing is another core geniculate nucleus of the thalamus), the oc-
deficit of developmental dyslexia. The work cipital cortex, and the corpus callosum, all
of Eden and others has shown that individu- of which process visual information or are
als with developmental dyslexia are im- involved in the transfer of visual informa-
paired on a number of tasks that involve tion to the language centers, have also been
early sensory mechanisms, including visual implicated (Hynd et al., 1995; Zeffiro &
motion, visuomotor, and visuospatial pro- Eden, 2001). The remainder of this chapter
cessing (Zeffiro & Eden, 2001). Thus, al- provides an overview of the literature on the
though the most common and accurately neurobiological basis of reading disabilities.
identified type of reading disability is
phonological dyslexia (Stanovich, 1999), an Review of the Research Literature
orthographic or “surface” subtype might on Reading Disabilities
exist in conjunction with or independent of
phonological dyslexia. The Double-Deficit Hypothesis

Reading disabilities and associated diffi- The double-deficit theory of dyslexia is cur-
culties reflect a persistent deficit manifest rently receiving considerable attention in
persist throughout the lifespan (Lyon, 1996; the neuropsychological literature. Accord-
Shaywitz & Shaywitz, 1999). For example, ing to Wolf and Bowers (1999), the double-
of children who are reading disabled in the deficit hypothesis of dyslexia posits “pho-
third grade, 74% remain disabled in the nological deficits and processes underlying
ninth grade (Temple, 1997). This persis- naming speed represent two separable
tence throughout development is concor- sources for reading dysfunction” (p. 415),
dant with the evidence that there is a strong such that there are separate types of reading
genetic component in the development of disabilities characterized by single deficits in
these disorders, with 35% to 45% familial phonological processing or rapid naming as
risk rates consistently reported in the litera- well as a more pervasive and severe form of
ture (Hynd et al., 1995). Neurobiological dyslexia characterized by deficits in both
evidence as to the etiology of reading dis- phonological processing and rapid naming.
abilities is supported by postmortem, elec- Phonological processing is a skill that al-
trophysiological, family, genetic, and brain- lows an individual to hear and manipulate
imaging studies (Galaburda, 1993; Hynd et individual sounds in spoken language (Cut-
al., 1995; Hynd & Semrud-Clikeman, ting & Denckla, 2001). It is a part of the
1989; Kinsbourne, 1989; Lyon, 1996; Tem- larger skill units of auditory perception and
ple, 1997; Zeffiro & Eden, 2001). discrimination, but it is involved only in
sounds that correspond to speech. Rapid
Evidence from these studies suggests dis- naming is a skill that allows an individual to
ruption of the neurological system for lan- access concept names from their store of
guage in individuals with dyslexia. Most of highly automatized information using visual
the structures which have been implicated in prompts (Denckla & Rudel, 1974).
the neurobiological basis of reading disabili-
ties are known to process language and/or Many studies have examined the role
visual information; thus it is not surprising phonological processing and rapid naming
that abnormalities in these structures are ability play in the acquisition of reading
linked with the language and/or visual skills. Several studies have documented that
deficits seen in reading disabilities. These these skills contribute uniquely to the vari-
structures include, but are not limited to, ance in reading ability (e.g., Cronin &
Broca’s area, the angular gyrus, the planum

244 CAUSES AND BEHAVIORAL MANIFESTATIONS

Carver, 1998; Wagner et al., 1997). It also that an individual with a reading disability
appears that the relationship between has a specific deficit in reading, but not in
phonological processing and reading ability other domains of achievement, that is “un-
is bidirectional; thus, as phonological pro- explained” by intellectual ability
cessing develops so does reading ability (Stanovich, 1999).
which in turn allows phonological process-
ing to develop further (e.g., Wagner, Torge- These challenges in the appropriate diag-
sen, & Rashotte, 1994). In addition, early nosis of reading disability have been under
phonological processing skills predict later debate. Stanovich and Siegel have proposed
reading achievement (e.g., Torgesen, Wagn- that the aptitude–achievement discrepancy
er, & Rashotte, 1994). model be abandoned for reading disability
diagnosis (Siegel, 1988, 1999; Stanovich,
Most of the studies testing the double 1991, 1999; Stanovich & Siegel, 1994).
deficit hypothesis of dyslexia have focused They cite as basis for their proposal that
on using phonological processing and rapid fact that the neurolinguistic deficits of poor
naming skills to differentiate good readers readers are the same regardless of intellectu-
from children with reading disabilities. al ability (Fletcher, Francis, Rourke, Shay-
These studies suggest that children with im- witz, & Shaywitz, 1994; Stanovich, 1999;
paired phonological processing and rapid Stanovich & Siegel, 1994) and there is no
naming deficits are at most significant risk evidence that low-IQ and high-IQ poor
for reading disabilities and also represent readers respond differently to intervention
the most impaired readers, confirming the (Stanovich, 1999).
double-deficit hypothesis (e.g., Manis, Doi,
& Bhadha, 2000). Attempts are being made As an alternative to the aptitude–achieve-
to connect research on the double-deficit ment discrepancy model, Siegel (1999) pro-
hypothesis to neurological characteristics posed a cutoff model of reading disability
associated with dyslexia in our research and diagnosis. In this model, poor performance
by others. on a single test of psuedo-word reading is
the primary indicator of a reading disability.
The double-deficit theory of dyslexia has Performance below a standard score of 85
also been used to differentiate children with on this task, a measure of phonological pro-
dyslexia from children who are garden-vari- cessing, with intellectual ability at or above
ety poor readers. The latter group is made a standard score of 85, would indicate a
up of children who fail to exhibit a discrep- reading disability according to the Siegel cri-
ancy between reading achievement and cog- teria. Stanovich (1999) suggested that read-
nitive ability due to low-cognitive-ability ing disabilities be indicated by performance
scores. These children are often not eligible below the 15th percentile on either a mea-
for special education support services be- sure of pseudo-word reading or a test of
cause their cognitive ability is too low for word recognition, with intellectual ability
them to meet the criteria for a learning dis- playing no role in the diagnosis. But, this
ability in reading. Conversely, their cogni- model may serve to simply sample the lower
tive ability is too high to meet the criteria end of the distribution for reading achieve-
for mental retardation. The double-deficit ment, rather than assessing students who
model has been used to justify, amid signifi- experience what some would refer to as
cant controversy, providing special educa- dyslexia, a more severe form of reading dis-
tion intervention to children who display ability, always thought to have a neurobio-
weaknesses in phonological processing and logical etiology.
rapid naming, regardless of cognitive ability
(Vellutino et al., 1996). The research in this The Planum Temporale
area suggests that a redefinition of what
constitutes a reading disability may be in or- Brain-based research of dyslexia has fo-
der. Traditionally, learning disabilities have cused a great deal of attention on a struc-
been diagnosed on the basis of a discrepan- ture referred to as the planum temporale. In
cy between intellectual ability and reading a seminal study, Geschwind and Levitsky
achievement. These diagnostic criteria, (1968) reported their findings on the asym-
which have been used in research and clini- metry of the planum temporale, an area that
cal settings, are based on the assumption contains the bilateral auditory association

Neurological Correlates of Reading Disabilities 245

cortices. This area is thought to be impor- of these studies documented deviations
tant in language. Of 100 normative post- from normal patterns of plana asymmetry.
mortem brains, 65% showed a larger left
planum temporale, 11% showed a larger In an attempt to further elucidate the re-
right planum temporale, and the remaining lationship between length of the planum
24% were of equal size. These findings are temporal and reading deficits, a study the
highly significant and have led other re- following year used the same measurement
searchers to examine the role of the planum technique used by Hynd and colleagues to
temporale in language-based deficits, in- measure the plana from MRI scans. Results
cluding reading disabilities. indicated that individuals with leftward
asymmetry were significantly better on mea-
Galaburda and Kemper (1979) described sures of total reading achievement, word at-
results from a postmortem examination of a tack, reading comprehension, confronta-
young man diagnosed with dyslexia during tional naming, and rapid naming ability
his childhood. They documented symmetri- (Semrud-Clikeman, Hynd, Novey, & Eliop-
cal plana with polymicrogyri. Polymicro- ulos, 1991). The relationship held when the
gyria, a collection of many unusually small contrast groups had rightward asymmetry
gyria, are considered an anomaly of neuro- and symmetrical plana. To some degree, the
logical ontogeny. In a follow-up study, sym- literature then began to shift from measur-
metrical plana were noted in three other ing the length of the plana toward develop-
dyslexic males on postmortem examination ing typologies of the morphology of the
(Galaburda, Sherman, Rosen, Aboitiz, & perisylvian region.
Geschwind, 1985). In 1990, Humphreys,
Kaufman, and Galaburda reported on post- Gyral Morphology of the Perisylvian Region
mortem results for three female dyslexics,
all of whom had symmetrical plana. The Theory and research into the biological ba-
theoretical interpretation was that some fac- sis of dyslexia have traditionally focused on
tor had an impact on pruning the normal the left perisylvian region, which includes
process of neuronal elimination during de- the planum temporale. The left hemisphere
velopment of the right plana, thus allowing has been the focus of the majority of the re-
for inefficient processing auditory–linguistic search because it is the dominant hemi-
stimuli. Although these results provided ad- sphere for language in most individuals. In
ditional information about the nature of the recent years, some researchers have begun
dyslexic brain, a more specific protocol for to investigate the gyral and sulcal morphol-
measurement of the planum temporale was ogy of the region and its possible relation to
needed as well as a method for measuring learning disabilities.
the planum temporale in a living person.
Magnetic resonance imaging (MRI) and In a postmortem and MRI study of 120
functional magnetic resonance imaging left and right hemispheres of presumably
(fMRI) has provided the means to achieve non-learning disabled individuals, Stein-
this goal. metz, Ebeling, Huang, and Kahn (1990)
classified four main types of gyral morphol-
In 1990, a pattern of smaller left plana ogy patterns. Figure 15.1 illustrates these ty-
when compared to the right plana (reversed pologies. Figure 15.2 contains examples of
asymmetry) and symmetry was noted in the typologies from MRI scans of children.
children with dyslexia (Hynd et al., 1990). The typologies were classified on the basis
These measures were calculated from MRI of the relative size and position of the poste-
scans using clinically identified children. Us- rior ascending ramus (PAR) of the Sylvian
ing a somewhat similar methodology, fissure and the inferior postcentral sulcus
Larsen, Hoien, Lundberg, & Odegaard (POCS). The Type I, or “textbook,” pattern
(1990) found that individuals with dyslexia according to Steinmetz and colleagues, the
were more likely to have a larger right plana PAR of the Sylvian fissure ascends posterior
and this rightward asymmetry was associat- to the POCS. This morphology was the
ed with phonological processing problems. most common as it was found in 67% and
Different measurement techniques were 65% of left hemispheres and 82% and 85%
used which may have accounted for these of right hemispheres in the postmortem and
slightly different findings. Regardless, both MRI samples. Type II is characterized by the

246 CAUSES AND BEHAVIORAL MANIFESTATIONS

FIGURE 15.1. Perisylvian region subtypes (Steinmetz et al., 1990).

absence of a PAR of the Sylvian fissure and gy in the left hemisphere of dyslexics com-
was found only in 10% of the postmortem pared to controls and suggested that the in-
sample and 25% the MRI sample and only crease in Type III morphology could be re-
in left hemispheres. In the Type III morphol- lated to learning disabilities. This conjecture
ogy, which was found only in 5% of right was strengthened by the finding that the
hemispheres and only in the postmortem first-degree relatives, or individuals who
sample, a second POCS exists, creating an share 50% of their genetic material, such as
extra gyrus. Type IV gyral morphology is parents and children or full siblings, of indi-
characterized by the continuation of the viduals with dyslexia, also had increased in-
PAR into the POCS. This type was found cidence of Type III morphology in the left
in the left hemisphere in 2% and 5% of the hemisphere, which Leonard suggested could
postmortem and MRI samples, respective- be evidence for genetic transmission of gyral
ly, and 13% and 15% of the right hemi- morphology. Heimenz and Hynd found that
spheres. although the Steinmetz typologies were not
useful in the differentiated diagnosis of
Both Leonard and colleagues (1993) and dyslexia, gyral morphology did relate to
Heimenz and Hynd (2000) have attempted performance on neurolinguistic measures of
to link the Steinmetz classification system to receptive language. Specifically, left-hemi-
reading disabilities by comparing the gyral sphere Type I morphology was associated
morphology of individuals with dyslexia with better performance on the Peabody
and controls. Leonard and colleagues found Picture Vocabulary Test—Revised (PPVT-R)
an increase of Type III gyral morphology when compared to Type II and III mor-
and a smaller incidence of Type I morpholo-

Neurological Correlates of Reading Disabilities 247

FIGURE 15.2. Perisylvian region subtype examples from MRI scans of human brains.

phologies, causing the authors to conclude 1995; Geschwind & Galaburda, 1985).
that the absence of an extra gyrus in the left Galaburda goes on to suggest that the larger
hemisphere may confer an advantage in re- right planum temporale is interfering in the
ceptive language. Interestingly, Heimenz normal dominance of the left planum tem-
and Hynd found that Type II occurred only porale for language processing. As noted
in four male subjects who were diagnosed previously, the right planum temporale is
with both dyslexia and ADHD, suggesting thought to be larger due to insufficient
the possibility of a unique and rare dyslexia pruning in early ontogeny, and Galaburda,
syndrome. Rosen, and Sherman (1990) also suggest
that increased callosal connections might be
The Corpus Callosum related to insufficient pruning. In this con-
ceptualization, the interference of the right
Galaburda has theorized that development planum temporale on the left planum tem-
dyslexia is due, in part, to a larger right porale’s dominance for language may mani-
planum temporale, and evidence for this fest itself through the morphological varia-
anatomical difference in dyslexics and nor- tion in the corpus callosum.
mal readers is substantial (Galaburda,
Although there has been conflicting evi-

248 CAUSES AND BEHAVIORAL MANIFESTATIONS

dence regarding the size of the corpus callo- throughout the cortex, laminar heterotopias
sum in developmental dyslexia (Duara et are often associated with mental re-
al., 1991; Hynd et al., 1995, Preis, Stein- tardation and epilepsy (Palmini et al.,
metz, Knorr, and Jancke, 2000), most stud- 1991).
ies have found significant differences in size.
These differences are thought to be related Nodular heterotopias are clustered mass-
to reduced cortical asymmetry. Hynd and es of gray matter in irregularly formed nod-
colleagues (1995) found that the genu of the ules separated by thin myelinated fibers lo-
corpus callosum (most anterior part) was cated near lateral ventricles (Friede, 1989;
smaller in dyslexics. Other researchers Hynd & Willis, 1988). Typically, nodular
(Larsen et al., 1992), however, have found a heterotopias are asymptomatic when they
larger corpus callosum in developmental occur in small numbers (Aicardi, 1992).
dyslexia. Preis and colleagues (2000) points When there are more migrational deficits,
out that the current results in studies exam- nodular heterotopias have been associated
ining this issue are split evenly, leaving us no with agenesis of the corpus callosum and
closer to a conclusion about the potential micrencephaly (Friede, 1989). Important to
role of the corpus callosum on language and this discussion, however, nodular hetero-
language disorders. These discrepancies in topias have been found in brains of individ-
research findings are likely due methodolog- uals with dyslexia upon postmortem exam-
ical differences which remain to be resolved ination (Galaburda, et al., 1985). The
in this line of research, including criteria for distribution and severity of the dysplasias
inclusion in the clinical and control groups, may determine the degree of impact upon
the method of measuring the corpus callo- learning (Hynd & Willis, 1988). Focal dys-
sum and other relevant brain structures, and plasias are typically the result of migration
variability varieties in the imaging technolo- errors during the second gestational
gy. trimester where only a limited number of
migrational waves of neurons are affected
Cortical Abnormalities of the (Palmini et al., 1991). According to Palmini
Temporal–Parietal Region and colleagues (1993), the extent of the mi-
gration deficits and the clinical manifesta-
In addition to the structures with measur- tions may be related such that more migra-
able differences in size and morphology al- tion deficits may be associated with a
ready described, the literature suggests that greater degree of clinical impairment.
there may be differences at the level of mi-
croneuropathology in the brains of individ- As previously indicated, polymicrogyria
uals with dyslexia. Although these differ- are atypical gyri that are crowded and
ences are difficult to image and are best small; convolutional patterns are also atypi-
described from the standpoint of histology cal in polymicrogyria (Aicardi, 1992).
upon postmortem examination, it appears Polymicrogyria can cover the entire cortical
that differences caused by focal dysplasias surface of either one or both hemispheres
or heterotopias and polymicrogyria may but are typically localized to a limited area
contribute to the performance of individuals of the cortex (Aicardi, 1992; Friede, 1989).
with dyslexia (Hynd, Morgan, & Vaughn, Polymicrogyria can be asymptomatic, but it
1997). has also be associated with the incidence of
learning disabilities (Galaburda & Kemper,
Focal dysplasias or heterotopias are “ec- 1979; Galaburda et al., 1985). Generally,
topic gray matter in the cerebral hemi- polymicrogyria are associated with dyslexia
spheres” (Hynd et al., 1997, p. 49). There when the affected area is in the perisylvian
are two types of focal dysplasias: laminar region (Galaburda & Kemper, 1979; Hynd
heterotopias and nodular heterotopias. & Willis, 1988). Polymicrogyria is the result
Laminar heterotopias are gray matter in of disturbances in cortical gyration that
symmetrical ribbons in the centrum semio- takes place during the fifth to sixth gesta-
vale (Friede, 1989). Laminar heterotopias tional month following neuronal migration
are not associated with deficits in perfor- (Aicardi, 1992; Friede, 1989). It is not clear
mance when limited in scope. When dys- what causes these disturbances, but mater-
plasias are present to a significant degree nal influenza, severe trauma, secondary fe-
tal asphyxia, and genetic transmission have

Neurological Correlates of Reading Disabilities 249

been implicated in polymicrogyria (Hynd et Van Orden, Bookman, & Smith, 1987).
al., 1997). Galaburda (1990) stated that in addition to
the impact of testosterone on fetal brain de-
Neuronal Pruning Hypothesis velopment, a genetic predisposition toward
developing bilateral plana is likely involved.
Noting clinical observations of the high in-
cidence of autoimmune disorders and left- Genetics
handedness among dyslexics, Geschwind
(and later in collaboration with Galaburda) The recent national interest in the human
hypothesized that autoimmune function, genome has increased the emphasis on ge-
left-handedness, and dyslexia might be netic research in all fields. Not surprisingly,
related to androgen secretion during devel- research on dyslexia has followed this trend,
opment (Geschwind & Behan, 1982; but the investigation of dyslexia’s genetic
Geschwind & Galaburda, 1985). Gesch- basis is not new. Since the early 1970s, re-
wind assumed as the basis of his theory that search has indicated a dramatic increase in
dyslexia is associated with symmetry or re- risk for a childhood reading disability to oc-
versed asymmetry of the planum temporale cur when a parent also had a reading dis-
(Galaburda et al., 1985; Geschwind & Lev- ability. In considering genetic research on
itsky, 1968; Humphreys et al., 1990) and dyslexia, it is important to remember that
noted that the symmetrical plana of dyslexi- reading is a complex cognitive task that in-
cs were bilaterally larger than that of con- volves attention, memory, phonological
trols (Galaburda et al., 1985). processing, and rapid naming, among other
abilities. As a result, any findings related to
Researchers theorized that the symmetry individual chromosomal linkage might actu-
or reversed asymmetry of the plana is due to ally identify a chromosomal difference asso-
decreased rates of ontogenetic cell death in ciated with aspects of cognition or linguistic
the left planum temporale during the period abilities associated with the clinical manifes-
of corticogenesis that occurs between the tation of dyslexia.
fifth and seventh month of human fetal de-
velopment (Geschwind & Behan, 1982; However, it is important to emphasize that
Hynd & Semrud-Clikeman, 1989). Gal- research suggests particular chromosomal
aburda and colleagues (1985) suggested loci associated with reading and language
that this decreased rate of cell death during disabilities. Several studies implicate chro-
corticogenesis may be associated with a fa- mosomes 6 and 15 in reading disabilities
milial tendency toward autoimmune and al- (Grigorenko et al., 1997). Previous research
lergic disorders, while Geschwind’s hypoth- suggests that dyslexia is the result an autoso-
esis (Geschwind & Behan, 1982) pointed to mal dominant transmission with variable ex-
a high level of prenatal testosterone as the a pression and incomplete penetrance (Elbert
potential cause of dyslexia, left-handedness, & Seale, 1988). This means that each child
and autoimmune disorders. In fact, testos- of a parent with the genes hypothesized to
terone can decrease ontogenetic cell death cause dyslexia has a 25% probability of in-
(Kelley, 1993) and has been thought to af- heriting those genes. These genes are not sex
fect the development of the immune system linked, and in children who inherit the genes,
(Geschwind & Galaburda, 1985). the expression of dyslexia can range from be-
ing nonapparent to causing severe impair-
The important role of testosterone in the ment. Profiles of impairment may vary
development of these two disorders is across family members.
strengthened by studies which show a high-
er prevalence rate of developmental dyslexia Research also suggests that both genetic
in boys (Geschwind & Galaburda, 1985). and environmental etiological factors have
Later studies, however, have attributed this an impact on developmental dyslexia.
difference in prevalence rates on selection Specifically, children with dyslexia demon-
bias (Shaywitz, Shaywitz, Fletcher, & Esco- strate skill profiles that are similar to those
bar, 1990). Furthermore, evidence associat- of their nonimpaired siblings (DeFries,
ing left-handedness, immune disorders, and Singer, Foch, & Lewitter, 1978) and their
dyslexia has not consistently supported the parents (Foch, DeFries, McClearn, &
Geschwind hypothesis (Pennington, Lefly, Singer, 1977). Using behavioral genetic

250 CAUSES AND BEHAVIORAL MANIFESTATIONS

techniques, research using nuclear families pus callosum in individuals with develop-
has identified an allele for dyslexia that ex- mental dyslexia (Preis et al., 2000). Further-
plains 54% of the phenotypical variance in more, in the first reported postmortem cases
reading performance (Gilger, Pennington, & of a dyslexic’s brain, cortical abnormalities
DeFries, 1991). including polymicrogyri were found in
the temporal–parietal region (Galaburda,
Data collected in studies of twins suggest LeMay, & Kemper, 1978). Recently, re-
that 50% of the variance in reading prob- search has begun to focus on the relation-
lems experienced by children with dyslexia ship between reading difficulties, verbal
is due to heritable influences, but there is ability, and gyral morphology of the perisyl-
also a significant environmental component vian region (Heimenz & Hynd, 2000;
as well (Wadsworth, Olson, Pennington, & Leonard et al., 1993), which could be asso-
DeFries, 1992). Another study found that ciated with the relative size and position of
what is inherited is reading performance, receptive language areas such as the planum
rather than reading disabilities per se, with temporale.
data to demonstrate that parent reading
performance predicts child reading perfor- The morphology of these areas of the
mance regardless of whether or not dyslexia brain is likely related to genetic factors as
is present (Alarcon, DeFries, & Gillis, well as to variability in fetal development.
1994). In fact, there is consistent evidence that
reading disabilities tend to run in families,
Summary of Consistent Findings with 50% of the variance in reading prob-
lems explained by genetics (Wadsworth et
Developmental dyslexia is a complex behav- al., 1992). Several studies have implicated
iorally diagnosed disorders that reflects un- chromosomes 6 and 15 (Grigorenko et al.,
derlying cognitive deficits that are believed 1997). Perinatal environment has also been
to be related to neurobiological abnormali- implicated in the development of reading
ties. The primary behavioral deficit is an in- disabilities. The Geschwind hypothesis the-
ability to rapidly decode words (word at- orized that higher levels of testosterone in
tack), which results in poor and dysfluent utero could decrease pruning that normally
reading. The primary underlying cognitive occurs during corticogenesis and could af-
deficit is theorized to be related to phono- fect immune functioning, leading to devel-
logical processing and rapid naming speed opmental dyslexia, left-handedness, and au-
(Wolf & Bowers, 1999), although evidence toimmune disorders (Geschwind & Behan,
also exists that there is an orthographic or 1982; Geschwind & Galaburda, 1985). It
visual component in some individuals (Eden must be remembered, however, that aspects
et al., 1996; Pennington et al., 1987), which of the postuterine environment, such as ex-
might constitute a separate subtype. posure to language and reading, certainly
play an important role in the development
The neurobiological correlates believed to of reading problems.
underlie these cognitive deficits are centered
around the left temporal–parietal region. Whereas the neurobiological evidence
Differences in the asymmetry of the planum seems to support the notion that neurode-
temporale have consistently been found in velopmental variation during fetal ontogeny
association with reading disabilities (Gal- is associated with deficient language and
aburda & Kemper, 1979; Galaburda et al., reading skills found in dyslexics, it is not
1985; Hynd & Semrud-Clikeman, 1989). exactly clear what cognitive abilities are
Specifically, symmetry of the planum tem- most affected. In this context, it is helpful
porale (Galaburda et al., 1985; Humphreys to examine models of cognitive processes
et al., 1990) due to a larger right plana or a most representative of the dyslexic syn-
reversal of the normal pattern of left greater drome.
than right asymmetry (Hynd et al., 1990)
has been found in individuals with develop- Competing Models
mental dyslexia. Although the research find-
ing have been inconsistent, most studies The models described in this section are not
have found differences in the size of the cor- necessarily contradictory to those previous-

Neurological Correlates of Reading Disabilities 251

ly described. In fact, all three models pro- processing is the interpretation of abstract
vide complementary information that en- representations (series of letters that form
hances the predominant research focus in words) during the process of reading. Or-
dyslexia studies. These models include de- thographic processing is most closely relat-
scriptions of dyslexia from the perspectives ed to sight word reading where the individ-
of deficits in the visual processing system, a ual does not use decoding strategies to read
nondiscrepancy-based model, and one that words but, rather, knows the entire word
focuses on the role of orthographic process- “on sight.” Research on orthographic cod-
ing in the presentation of the reading ing suggests that it contributes significantly
deficits associated with dyslexia. to word-reading ability (Olson, Forsberg, &
Wise, 1994). Furthermore, this contribution
The Visual Processing Model appears to be beyond that of the contribu-
tions of phonological processing to the
The model proposed by Eden and colleagues reading process (Cunningham, Perry, &
(1996) focuses on the visual processing of Stanovich, 2001).
rapidly presented information. Although this
paradigm for the investigation of dyslexia Although there is evidence to suggest that
appears opposed to the more linguistic mod- phonological processing is influenced by ge-
els, it represents a perspective that posits netic factors, there is little evidence to sug-
global temporal-processing deficits, both au- gest the same for orthographic coding. In
ditory and visual, as the underlying factors in fact, Stanovich and colleagues have noted in
dyslexia. As a result, this model is comple- multiple articles that orthographic process-
mentary to the neurolinguistic model of ing is linked to exposure to print and other
dyslexia. According to this model, these environmental indicators of reading (e.g.,
deficits are related to magnocellular process- Cunningham & Stanovich, 1993; Olson,
ing in the visual pathway system. According Wise, Conners, Rack, & Fulker, 1989). This
to Livingstone and Hubel (1988), the mag- may explain, in part, why genetic factors
nocellular system is responsive to stimulus only account for 50% of the variance in
onset and offset, rather than motion or color reading performance.
sensitivity. Research by Felmingham and
Jakobson (1995) suggests evidence for mag- Conclusions and Future Directions
nocellular deficits and deficient flicker/mo-
tion detection abilities in individuals with An abundance of evidence supports a neuro-
dyslexia. Furthermore, Eden and colleagues biological basis for developmental dyslexia.
present evidence that the extrastriate area of Persistent differences in the patterns of nor-
the cortex receives mostly magnocellular mal symmetry and/or morphology have con-
projects fails to show normal activation pat- sistently been noted in the temporal–parietal
terns in adults with dyslexia. region of the left hemisphere, specifically the
planum temporale (Galaburda & Kemper,
As an additional component in the visual- 1979; Galaburda et al., 1985; Hynd & Sem-
processing model, poor eye movement may rud-Clikeman, 1989). Cortical abnormali-
also be implicated in dyslexia. There is evi- ties have also been documented (Galaburda,
dence to suggest that poor eye movement LeMay, & Kemper, 1978). Research is now
stability or fixation may also be deficient in turning toward examining the pattern of gy-
individuals with dyslexia, although these ral morphology in this region and its relation
deficits may be better understood as a result to performance on neurolinguistic measures
of linguistic deficits (Eden, Stein, Wood, & and diagnosis (Heimenz & Hynd, 2000;
Wood, 1994). This evidence confirmed ear- Leonard et al., 1993).
lier work by Pirozzollo and Rayner (1979).
There is evidence that the origin of the
Orthographic processing neurological abnormalities found in dyslex-
ia is a genetic one, as reading disabilities
Orthographic processing is more closely re- tend to run in families (Wadsworth et al.,
lated to the visual aspects of reading, de- 1992). Controversial theories have also im-
scribed by Eden and others, than to the plicated high perinatal levels of testosterone
phonological components. Orthographic as a possible causal factor of abnormal

252 CAUSES AND BEHAVIORAL MANIFESTATIONS

brain development in the language cortex perisylvian morphology, research using
(Geschwind & Galaburda, 1985). fMRI may facilitate targeting intervention
strategies to specific deficits observed in the
Though many researchers have concluded metabolic functioning in brain regions in
that the core deficit in dyslexia is phonolog- the future. If fMRI becomes part of the
ical processing (Huettner, 1994; Hynd et al., technologies used by to monitor progress in
1995; Lyon, 1996; Shaywitz & Shaywitz, the remediation of specific reading deficits,
1999; Siegel, 1993), competing models ex- this may lead to the identification of more
ist. They include the visual processing mod- effective intervention efforts. Furthermore,
el, which implicates the magnocellular path- if intervention can be demonstrated to be ef-
way of the visual system as dysfunctional, fective in changing how a child reads, it may
and the orthographic processing model, further the knowledge about the nature of
which implicates difficulties with accurate dyslexia.
perception and processing of visual sym-
bols. Both models have been described by A final area warranting mention as a fu-
Eden and colleagues (1996) and propose vi- ture area of study on dyslexia is further re-
sual deficits in reading as opposed to the au- search about the nature of the genetics in-
ditory deficits proposed in the phonological volved in dyslexia. Although the reviewed
model and double-deficit model. These the- research suggests a strong heritable compo-
ories are by no means mutually exclusive, nent for reading processes, there are many
and it is likely that neurolinguistic deficits unknowns yet to be explored. For example,
occur in both auditory and visual process- although certain genes have been targeted as
ing in individuals with reading disabilities, being involved in dyslexia, it is not known
and that these deficits have a neurological how these chromosomes cause the manifes-
basis. It is clear that research should contin- tation of reading deficits. Furthermore, the
ue to examine the relationship between and nature of the relationship between these
prevalence of both auditory and visual pro- chromosomal differences and the visual
cessing difficulties. processes involved in reading has yet to be
explored. In addition, there has been little
It should also be pointed out that vari- investigation on the genetic contributions to
ability in pattern of plana symmetry or dyslexia in minority populations. Without
asymmetry is not a sufficient cause of severe additional study, there may be more unan-
reading disability or dyslexia. This is be- swered questions about the nature of
cause symmetry and reversed asymmetry of dyslexia than questions with answers.
the plana appear in the normative popula-
tion (Geschwind & Levitsky, 1968; Hynd et Acknowledgment
al., 1990). As Galaburda and colleagues
(1985) suggest, it may be that cortical or Preparation of this chapter was supported in part by
subcortical dysplasias must also be present a grant to the third author (GWH) from the Nation-
in the brain to such a degree that they seri- al Institute for Child Health and Human Develop-
ously disrupt normal language processes so mental, National Institutes of Health (No. 26890-
essential to the development of fluent read- 06).
ing ability. Unfortunately for research pur-
poses, these cortical abnormalities cannot References
be imaged using MRI as the spatial resolu-
tion is not yet sufficient to reveal abnormal- Aicardi, J. (1992). Diseases of the nervous system in
ities in the microstructure of the brain. childhood. New York: MacKeith Press.

An additional area that merits further in- Alarcon, M., DeFries, J. C., & Gillis, J. J. (1994).
vestigation is the link between the brain- Familial resemblance for measures of reading
based deficits in dyslexia and intervention. performance in families of reading-disabled and
Although recent studies of dyslexia have control twins. Reading and Writing, 6, 93–101.
used fMRI technology to further describe
the involved brain–behavior relationships, American Psychiatric Association. (2000). Diagnos-
this technology has yet to be applied to the tic and statistical manual of mental disorders (4th
efficacy of intervention strategies. Although ed., text revision). Washington, DC: Author.
there is no way to intervene in childhood to
change planum temporale asymmetry or Barkley, R. A. (1997). Behavioral inhibition, sus-
tained attention, and executive functions: Con-

Neurological Correlates of Reading Disabilities 253

structing a unifying theory of ADHD. Psycholog- Galaburda, A. M., & Kemper, T. L. (1979). Cytoar-
ical Bulletin, 121, 65–94. chitectonic abnormalities in developmental
Cronin, V., & Carver, P. (1998). Phonological sensi- dyslexia: A case study. Annals of Neurology, 6,
tivity, rapid naming, and beginning reading. Ap- 94–100.
plied Psycholinguistics, 19, 441–460.
Cunningham, A. E., Perry, K. E., & Stanovich, K. E. Galaburda, A. M., LeMay, M., & Kemper, T. L.
(2001). Converging evidence for the concept of (1978). Right–left asymmetries in the brain:
orthographic processing. Reading and Writing, Structural differences between hemispheres may
14, 549–568. underlie differences. Science, 199, 852–856.
Cunningham, A. E., & Stanovich, K. E. (1993).
Children’s early literacy environments and early Galaburda, A. M., Rosen, G. D., & Sherman, G. F.
word recognition subskills. Reading and Writing, (1990). Individual variability in cortical organi-
5, 193–204. zation: Its relationship to brain laterality and im-
Cutting, L. E., & Denckla, M. B. (2001). The rela- plications to function. Neuropsychologia, 28,
tionship of rapid serial naming and word reading 529–546.
in normally developing readers: An exploratory
model. Reading and Writing, 14, 673–705. Galaburda, A. M., Sherman, G. F., Rosen, G. D.,
DeFries, J. C., Singer, S. M., Foch, T. T., & Lewitter, Aboitiz, F., & Geschwind, N. (1985). Develop-
F. I. (1978). Familial nature of reading disability. mental dyslexia: Four consecutive patients with
British Journal of Psychiatry, 132, 361–367. cortical abnormalities. Annals of Neurology, 18,
Denckla, M. B., & Rudel, R. G. (1976). Rapid “au- 222–233.
tomatized” naming of pictured objects, colors,
letters, and numbers by normal children. Cortex, Geschwind, N., & Behan, P. (1982). Left-handed-
10, 186–202. ness: Association with immune disease, migraine,
Duara, R., Kushch, A., Gross-Glenn, K, Barker, W., and developmental learning disorder. Proceedings
Jallad, B., Pascal, S., Loewenstein, D. A., Shel- of the National Academy of Sciences of the United
don, J., Rabin, M., Levin, B., & Lubs, H. (1991). States of America, 79, 5097–5100.
Neuroanatomic differences between dyslexia and
normal readers on magnetic resonance imaging Geschwind, N., & Galaburda, A. (1985). Cerebral
scans. Archives of Neurology, 48, 410–416. lateralization: Biological mechanisms, associa-
Eden, G. F., Van Meter, J. W., Rumsey, J. M., & Zef- tions, and pathology: I. A hypothesis and program
firo, T. A. (1996). The visual deficit theory of de- for research. Archives of Neurology, 42, 428–459.
velopmental dyslexia. Neuroimage, 4, S108–S117.
Eden, G. F., Stein, J. F., Wood, H. M., & Wood, F. Geschwind, N., & Levitsky, W. (1968). Human
B. (1994). Differences in eye movements and brain: Left–right asymmetries in temporal speech
reading problems in dyslexic and normal chil- region. Science, 161, 186–187.
dren. Vision Research, 34, 1345–1358.
Elbert, I. C., & Seale, T. W. (1988). Complexity of Gilger, J. W., Pennington, B. F., & DeFries, J. C.
the cognitive phenotype of an inherited form of a (1991) Risk for reading disability as a function of
learning disability. Developmental Medicine and parental history in three family studies. Reading
Child Neurology, 30, 181–189. and Writing, 3, 205–217.
Felmingham, K. L., & Jakobson, L. S. (1995). Visual
and visuomotor performance in dyslexic children. Grigorenko, E. L., Wood, F. B., Meyer, M. S., Hart,
Experimental Brain Research, 106, 467–474. L. A., Speed, W. C., Shuster, A., & Pauls, D. L.
Fletcher, J. M., Francis, D. J., Rourke, B. P., Shay- (1997). Susceptibility loci for distinct compo-
witz, S. E., & Shaywitz, B. A. (1994). The validi- nents of developmental dyslexia on chromosomes
ty of discrepancy-based definitions of reading dis- 6 and 15. American Journal of Genetics, 60,
abilities. Journal of Learning Disabilities, 25, 27–39.
555–561.
Foch, T. T., DeFries, J. C., MacClearn, G. E., & Heimenz, J. R., & Hynd, G. W. (2000). Sulcal/gyral
Singer, S. M. (1977). Familial patterns of impair- pattern morphology of the perisylvian language
ment in reading disability. Journal of Educational region in developmental dyslexia. Brain and Lan-
Psychology, 69, 316–329. guage, 74, 113–133.
Friede, R. L. (1989). Developmental neuropatholo-
gy (2nd ed.). Berlin, Germany: Springer-Verlag. Hinshelwood, J. (1900). Congenital word-blind-
Galaburda, A. M. (1990). The testosterone hypoth- ness. Lancet, 1, 1506–1508.
esis: Assessment since Geschwind & Behan,
1982. Annals of Dyslexia, 40, 18–38. Huettner, M. I. S. (1994). Neuropsychology of lan-
Galaburda, A. M. (1993). The planum temporale. guage and reading development. In P. A. Vernon
Archives of Neurology, 50, 457. (Ed.), The neuropsychology of individual differ-
Galaburda, A. M. (1995). Anatomic basis of cere- ences (pp. 9–34). San Diego, CA: Academic Press.
bral dominance. In R. J. Davidson & K. Hugdahl
(Eds.), Brain asymmetry (pp. 51–74). Cambridge, Humphreys, P., Kaufman, W. E., & Galaburda, A.
MA: MIT Press. (1990). Developmental dyslexia in women: Evi-
dence for a subgroup with a reversal of cerebral
asymmetry. Annals of Neurology, 28, 727–738.

Hynd, G. W., Hall, J., Novey, E. S., Eliopulos, D.,
Black, K., Gonzalez, J. J., Edmonds, J. E., Riccio,
C., & Cohen, M. (1995). Dyslexia and corpus
callosum morphology. Archives of Neurology,
52, 32–38.

Hynd, G. W., Hynd, C. R., Sullivan, H. G., &
Kingsbury, T. B. (1987). Regional cerebral blood
flow (rCBF) in developmental dyslexia: Activa-
tion during reading in a surface and deep dyslex-
ic. Journal of Learning Disabilities, 20, 294–300.

254 CAUSES AND BEHAVIORAL MANIFESTATIONS

Hynd, G. W., Morgan, A. E., & Vaughn, M. chopathology (pp. 390–435). New York: Guilford
(1997). Neurodevelopmental anomalies and mal- Press.
formations. In C. R. Reynolds & E. Fletcher- Manis, F. R., Doi, L. M., & Bhadha, B. (2000).
Janzen (Eds.), Handbook of clinical child neu- Naming speed, phonological awareness, and or-
ropsychology: Critical issues in neuropsychology thographic knowledge in second graders. Journal
(pp. 42–62). New York: Plenum Press. of Learning Disabilities, 33, 325–333.
Olson, R. K., Forsberg, H., & Wise, B. (1994).
Hynd, G. W., & Semrud-Clikeman, M. (1989). Genes, environment, and the development of or-
Dyslexia and brain morphology. Psychological thographic skills. In V. W. Berninger (Ed.), The
Bulletin, 106, 447–482. varieties of orthographic knowledge, 1: Theoreti-
cal and developmental issues. Neuropsychology
Hynd, G. W., Semrud-Clikeman, M., Lorys, A., and cognition (Vol. 8, pp. 27–71). Norwell, MA:
Novey, E., & Eliopulos, D. (1990). Brain mor- Kluwer Academic.
phology in developmental dyslexia and attention Olson, R., Wise, B., Conners, F., Rack, J., & Fulker,
deficit hyperactivity disorder. Archives of Neurol- D. (1989). Specific deficits in component reading
ogy, 47, 919–926. and language skills: Genetic and environmental
influences. Journal of Learning Disabilities, 25,
Hynd, G. W., & Willis, W. G. (1988). Pediatric neu- 562–573.
ropsychology. Needham Heights, MA: Allyn & Palmini, A., Andermann, F., Aicardi, J, Dulac, O.,
Bacon. Chaves, F., Ponsot, G., Pinard, J. M., Goutieres,
F., Livingston, J., Tampieri, D., Andermann, E.,
Kelley, D. B. (1993). Androgens and brain develop- & Robitaille, Y. (1991). Diffuse cortical dyspla-
ment: Possible contributions to developmental sia, or the “double cortex” syndrome: The clini-
dyslexia. In A. M. Galaburda (Ed.), Dyslexia and cal and epileptic spectrum in 10 patients. Neurol-
development: Neurobiological aspects of extra- ogy, 41, 1656–1662.
ordinary brains (pp. 21–41). Cambridge, MA: Palmini, A., Andermann, F., de Grissac, H.,
Harvard University Press. Tampieri, D., Robitaille, Y., Langevin, P., Desbi-
ens, R., & Andermann, E. (1993). Stages and
Kibby, M. Y., & Hynd, G. W. (2001). Neurobiolog- patterns of centrifugal arrest of diffuse neuronal
ical basis of learning disabilities. In D. P. Halla- migration disorders. Developmental Medicine
han & B. K. Keogh (Eds.), Research and global and Child Neurology, 35, 331–339.
perspectives in learning disabilities: Essays in Pennington, B. F., Lefly, D. L., Van Orden, G. C.,
honor of William M. Cruickshank (pp. 25–42) Bookman, M. O., & Smith, S. D. (1987). In
Mahwah, NJ: Erlbaum. phonology bypassed in normal or dyslexic devel-
opment. Annals of Dyslexia, 37, 62–89.
Kinsbourne, M. (1989). Mechanisms and develop- Pirozzollo, F. J., & Rayner, K. (1979). Cerebral or-
ment of hemisphere specialization in children. In ganization and reading disability. Neuropsy-
C. R. Reynolds & E. Fletcher-Janzen (Eds.), chologia, 17, 485–491.
Handbook of clinical child neuropsychology. Preis, S., Steinmetz, H., Knorr, U., & Jancke, L.
Critical issues in neuropsychology (2nd ed., pp. (2000). Corpus callosum size in children with de-
69–85). New York: Plenum Press. velopmental language disorder. Cognitive Brain
Research, 10, 37–44.
Kral, M., Nielson, K., & Hynd, G. W. (1998). His- Riccio, C. A., & Hynd, G. W. (1996). Neu-
torical conceptualization of developmental roanatomical and neurophysiological aspects of
dyslexia: Neurolinguistic contributions from the dyslexia. Topics in Language Disorders, 16, 1–13.
19th and early 20th centuries. In R. Licht, A. Semrud-Clikeman, M., Hynd, G. W., Novey, E. S.,
Bouma, W. Slot, & W. Koops (Eds.), Child neu- & Eliopulos, D. (1991). Dyslexia and brain mor-
ropsychology (pp. 1–16). Delft, The Netherlands: phology: Relationships between neuroanatomical
Eburon. variation and neurolinguistic tasks. Learning and
Individual Differences, 3, 225–242.
Larsen, J. P., Hoien, T., Lundberg, I., & Odegaard, Shaywitz, S. E., & Shaywitz, B. A. (1999). Cogni-
H. (1990). MRI evaluation of the size and sym- tive and neurobiologic influences in reading and
metry of the planum temporale in adolescents in dyslexia. Developmental Neuropsychology,
with developmental dyslexia. Brain and Lan- 16, 383–384.
guage, 39, 289–301. Shaywitz, S. E., Shaywitz, B. A., Fletcher, J. M., &
Escobar, M. D. (1990). Prevalence of reading dis-
Larsen, J. P., Hoein, T., & Odegaard, H. (1992). ability in boys and girls: Results of the Connecti-
Magnetic resonance imaging of the corpus callo- cut Longitudinal Study. Journal of the American
sum in developmental dyslexia, Cognitive Neu- Medical Association, 264, 998–1002.
ropsychology, 9, 123–134. Siegel, L. S. (1988). Evidence that IQ scores are ir-
relevant to the definition and analysis of reading
Leonard, C., Voeller, K. K. S, Lombardino, L., Mor- disability. Canadian Journal of Psychology, 42,
ris, M., Hynd, G. W., Alexander, A., Andersen, 201–215.
H., Garofalakis, M, Honeyman, J., Mao, J., Siegel, L. S. (1993). Phonological processing deficits
Agee, H., & Staab, E. (1993). Anomalous cere-
bral structure in dyslexia revealed with magnetic
resonance imaging. Archives of Neurology, 50,
461–469.

Livingstone, M. S., & Hubel, D. (1988). Segregation
of form, color, movement, and depth: Anatomy,
physiology, and perception. Science, 240, 740–
749.

Lyon, G. R. (1996). Learning disabilities. In E. J.
Mash & R. A. Barkley (Eds.), Child psy-

Neurological Correlates of Reading Disabilities 255

as the basis of a reading disability. Developmen- (1996). Cognitive profiles of difficult-to-remedi-
tal Review, 13, 246–257. ate and easily remediated poor readers: Early in-
Siegel, L. S. (1999). Issues in the definition and di- tervention as a vehicle for distinguishing between
agnosis of learning disabilities: A perspective on cognitive and experiential deficits as basic causes
Guckenberger v. Boston University. Journal of of specific reading disability. Journal of Educa-
Learning Disabilities, 32, 304–320. tional Psychology, 88, 601–638.
Stanovich, K. E. (1991). Conceptual and empirical Wadsworth, S. J., Olson, R. K., Pennington, B. F.,
problems with discrepancy definitions of reading & DeFries, J. C. (1992). Differential genetic etiol-
disability. Learning Disabilities Quarterly, 14, ogy of reading disability as a function of IQ.
269–280. Journal of Learning Disabilities, 33, 192–199.
Stanovich, K. E. (1999). Who is rational?: Studies Wagner, R. K., Torgesen, J. K., & Rashotte, C. A.
of individual differences in reasoning. Mahwah, (1994). Development of reading related-phono-
NJ: Erlbaum. logical processing abilities: New evidence of bi-
Stanovich, K. E., & Siegel, L. S. (1994). Phenotypic directional causality from a latent variable longi-
performance profile of children with reading dis- tudinal study. Developmental Psychology, 30,
abilities: A regression-based test of the phonolog- 73–87.
ical core variable-difference model. Journal of Wagner, R. K., Torgesen, J. K., Rashotte, C. A.,
Educational Psychology, 86, 24–53. Hecht, S. A., Barker, T. A., Burgess, S. R., Don-
Steinmetz, H., Ebeling, U., Huang, Y., & Kahn, T. ahue, J., & Garon, T. (1997). Changing relations
(1990). Sulcus topography in the parietal opercu- between phonological processing abilityies and
lar region: An anatomic and MR study. Brain word-level reading as children develop from be-
and Language, 38, 515–533. ginning to skilled readers: A five-year longitudinal
Temple, C. M. (1997). Reading disorders. In C. M. study. Developmental Psychology, 33, 468–479.
Temple (Ed.). Developmental cognitive neuropsy- Wolf, M., & Bowers, P. G. (1999). The double-
chology (pp. 163–223). Hove, East Sussex, UK: deficit hypothesis for developmental dyslexias.
Psychology Press. Journal of Educational Psychology, 91, 415–438.
Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. Zeffiro, T., & Eden, G. (2001). The cerebellum and
(1994). Longitudinal studies in phonological pro- dyslexia. Perpetrator or innocent bystander?:
cessing and reading. Journal of Learning Disabil- Comment from Thomas Zeffiro and Guinevere
ities, 27, 276–286. Eden to Nicholson et al. Trends in Neuro-
Vellutino, F. R., Scanlon, D. M., Sipay, E. R., Small, sciences, 24, 512–513.
S. G., Pratt, A., Chen, R., & Denckla, M. B.

16

Genetic Influences on Reading
and Writing Disabilities

Jennifer B. Thomson
Wendy H. Raskind

Dyslexia and dysgraphia are common com- earlier intervention for children at higher
plex disorders characterized by difficulty risk. Because the optimal developmental
learning to read and/or write, not attribut- window for acquisition of reading and writ-
able to general cognitive delay, psychiatric ing skills is narrow, a delay in triaging strug-
or neurological disorder, sensory impair- gling children for the special education ser-
ment, or inadequate instruction. There is vices they need may have an adverse effect
now considerable evidence that genetic fac- on their eventual learning outcome. Educa-
tors contribute to the development of tors may be reluctant to label younger chil-
dyslexia and dysgraphia (Raskind, 2001), dren as potentially “learning disabled”
although given the complex interactions be- when they may merely be at the lower end
tween genetic background and environmen- of the normal range of developing readers
tal factors, genetic information is not likely or writers. A suggestive family history of
to be the definitive diagnostic test for learning disability and a “high risk” geno-
dyslexia/dysgraphia. The phenotype (behav- type might spur earlier remediation for a
ioral expression) for dyslexia and dys- child who demonstrates a slow trajectory of
graphia is complex and is different for each acquisition of basic reading skills.
affected individual, and the occurrence and
manifestations of the disorder in an individ- In addition, genetic information could
ual may depend on the interaction of genetic have clinical applicability in the choice of a
factors (genes), environmental factors (e.g., specific instructional plan and prognosis for
amount and quality of educational expo- the intensity and duration of early interven-
sure), and stochastic processes (i.e., chance tion required. For example, our research
events). Nevertheless, delineating these ge- showing that phonological short-term mem-
netic influences will not only lead to a better ory has a genetic etiology (Wijsman et al.,
understanding of the biological mechanisms 2000) led to an instructional program in
of the disorder but will also be valuable for which word reading begins with creating a
assessment and intervention purposes. precise representation of syllables and
phonemes in spoken polysyllabic words
For example, knowledge of the genes in- rather than beginning with written words
volved in these disorders might facilitate (Berninger, 2000). Our research showing

256

Genetic Influences on Reading and Writing Disabilities 257

that the efficiency (rate) of phonological de- Estimates of the Heritability of
coding has a genetic etiology (Raskind, Reading/Writing Disabilities
Hsu, Berninger, Thomson, & Wijsman,
2000) led to treatment aimed at automatiz- Research strategies involving twins have
ing phonological decoding (Berninger, long been used to explore the relative con-
2000). Thus, a knowledge of genetic influ- tributions of genetic and environmental fac-
ences on dyslexia and dysgraphia may also tors. In the traditional twin design, differ-
contribute to development of effective in- ences between identical, or monozygotic
structional interventions tailored to geneti- (MZ), and fraternal, or dizygotic (DZ), co-
cally constrained processes in reading and twins are studied. Both types of twins share
writing. their home and school environments to a
substantial degree, but they differ in their
Terms genetic similarity. Because MZ twins arise
from the splitting of one early embryo, their
Every discipline has a vocabulary that is pe- genetic constitution is the same. DZ twins,
culiar to itself. Although this vocabulary fa- on the other hand, arise from two ova sepa-
cilitates communication between experts rately fertilized by two sperm; they have the
within the field, it may bar others from a same genetic relationship as other siblings
full understanding of the area. Appendix and have in common, on average, one-half
16.1 provides a short list of frequently used of their genes. For a dichotomous trait that
terms in human genetics. This list will help is entirely genetically determined, MZ twins
readers understand this chapter and publi- should be fully concordant, whereas DZ
cations about other genetic investigations. twins should be less often concordant.
For many of these terms, the reader may
need to refer to other words in the glossary Based on a categorical definition of
for additional definitions. dyslexia, the majority of early twin studies
found greater concordance among MZ than
Familial Aggregation DZ co-twins. In the first reported study of
reading disability in five sets of twins, two
The concept that reading disabilities may were concordant MZ pairs, one was a con-
have a hereditary basis is not new. Since the cordant opposite-sex DZ pair, and two were
first descriptions of dyslexia, researchers discordant opposite-sex DZ pairs (Hallgren,
have noted that reading disability clustered 1950). Hermann and Norrie (1958) report-
in families (e.g., Fisher, 1905; Hinshelwood, ed that 4 of 19 same-sex and 6 of 11 oppo-
1907; Stephenson, 1907; Thomas, 1905). site-sex DZ twin pairs were concordant for
Later, it was observed that there is a greater reading disability, whereas all 9 MZ twin
prevalence of reading difficulties in first- pairs were concordant. A weakness of this
degree relatives of individuals with dyslexia report is the lack of information regarding
as compared to controls (e.g., Decker & De- ascertainment or diagnostic procedures.
Fries, 1981; DeFries, Singer, Foch, & Lewit- Bakwin (1973) later reported that 26 of 31
ter, 1978; Finucci, Guthrie, Childs, Abbey, MZ twin pairs ascertained on the basis of
& Childs, 1976; Foch, DeFries, McClearn, history of reading problems in the proband
& Singer, 1977; Owen, Adams, Forrest, twin and only 9 of 31 same-sex DZ co-
Stolz, & Fisher, 1971; Wolff & Melngailis, twins were concordant for reading disabili-
1994). Familial aggregation is necessary but ty. The differences in concordance rates
not sufficient to prove a genetic etiology be- were significant. Significant differences be-
cause families tend to experience the same tween DZ and MZ concordance rates in the
environmental as well as genetic influences. larger of the studies and in aggregate pro-
Diseases such as lead poisoning or asbesto- vide evidence for a genetic etiology of read-
sis and behaviors such as speaking French ing/writing disabilities. However, the obser-
or eating tropical fruits are examples of fa- vation that concordance among MZ twins
milial traits that do not have a genetic etiol- is less than complete reflects the existence of
ogy. nongenetic factors in the development of
reading and writing disabilities.

Given the evidence that reading and writ-
ing skills are normally distributed in the

258 CAUSES AND BEHAVIORAL MANIFESTATIONS

population and that the threshold between coding (ability to analyze sound patterns in
categorically defined affected and nonaffect- spoken words) and orthographic coding
ed status is somewhat arbitrary, it is more (ability to distinguish between written real
precise to use quantitative data. For a genet- and pseudo-word homonyms) have been re-
ically determined continuous trait pheno- ported to be significantly heritable (approx-
type, the scores of DZ co-twins of probands imately 31–60%) (Castles, Datta, Gayan, &
selected for scoring in one tail of the normal Olson, 1999; Olson, Forsberg, & Wise,
distribution should regress halfway to the 1994; Olson, Wise, Connors, Rack, & Fulk-
population mean, whereas those of MZ co- er, 1989).
twins should not. On the other hand, for a
continuous trait that is entirely determined Although twin studies have provided in-
by environmental influence, the scores of formation regarding the heritability of a va-
co-twins in both MZ and DZ types of twin riety of phenotypes, there are some inherent
sets should regress equally to the population limitations. For research on learning disabil-
mean. In twin studies, an estimate of heri- ities, a major assumption has been that en-
tability can therefore be derived by compar- vironmental influences on MZ and DZ co-
ing the average regressions to the popula- twins’ reading/writing development are
tion mean for MZ and DZ co-twins of equivalent. It may be the case, however, that
probands (DeFries & Fulker, 1985). MZ twins share a more similar environment
than do DZ twins (Grayson, 1989; Joseph,
This method was first applied to a subject 2001; Kendler, Neale, Kessler, Heath, &
set of 64 MZ and 55 DZ twin pairs ascer- Eaves, 1993). The genetic etiology of co-
tained as part of a longitudinal study of morbid behaviors that modify the reading
reading and writing skills in children at- and writing phenotype may confound the
tending elementary schools in Colorado interpretation of studies on learning disabil-
(DeFries, Fulker, & LaBuda, 1987). Pro- ities in twins. For example, a substantial ge-
bands were selected on the basis of low netic contribution to attention/concentra-
achievement on a composite measure based tion skills has recently been reported (e.g.,
on the Reading Recognition, Reading Com- Faraone & Doyle, 2001; Willcutt, Penning-
prehension, and Spelling subtests of the ton, & DeFries, 2000). If MZ co-twins are
Peabody Individual Achievement test (Dunn more similar with regard to their ability to
& Markwardt, 1970), verbal intelligence of attend to written material, this may result in
no more than two-thirds standard devia- more similarity with regard to their atten-
tions below the mean (Wechsler, 1974), and tion to reading and writing tasks. Another
no evidence of neurological, sensory, or potential limitation in twin research lies in
emotional deficit. Average regressions of the generalization of these data to the popu-
1.16 and 0.77 standard deviation units to- lation at large. Are there etiological factors
ward the mean were found for DZ co-twins that are unique to the twinning situation for
and MZ co-twins, respectively, yielding an the variables under study? Does the greater
estimated heritability for reading/writing likelihood of participation in studies by MZ
disability of 0.29 (p = .003). A much higher than DZ twin pairs and by females than
heritability estimate of 0.58 was obtained in males contribute to a systematic bias? Over-
a more recent analysis of a larger twin sam- all, however, twin research supports the her-
ple from the same project—223 pairs of itability of reading disability and of the
MZ twins and 169 pairs of same-sex DZ phonological and orthographic processing
twins (Wadsworth, Olson, Pennington, & abilities underlying it.
DeFries, 2000).
Modes of Inheritance of
Reading and writing disabilities are het- Reading/Writing Disabilities
erogeneous disorders with multiple sub-
types, and it is possible that some subtypes The observation that dyslexia is highly fa-
may have an appreciable genetic etiology milial and moderately heritable does not
and others may not. Therefore, the heri- necessarily imply that it is the result of a sin-
tability of the underlying processes thought gle gene. There are several possible modes
to contribute to reading and writing ability
has also been evaluated. Both phonological

Genetic Influences on Reading and Writing Disabilities 259

of transmission of a genetically influenced ing Project twin study supported the exis-
trait. The phenotype can be caused by one tence of a major gene effect on reading dis-
or more independent, single major genes ability (e.g., Alarcón & DeFries, 1995),
(Mendelian), by several genes acting in com- twin studies do not provide precise informa-
bination (oligogenic), or by many genes tion about the mode of inheritance. To an-
each of which contributes a small amount swer this question, a variety of mathemati-
of variation (polygenic) (Smith & Goldgar, cal approaches have been developed to
1986). This distinction is of more than theo- model the inheritance patterns of a trait in
retical interest because it may be difficult to families. Detailed descriptions of the meth-
identify the multiple genes involved in a ods are beyond the scope of this chapter, but
polygenic trait when none has a big enough summaries of the major findings in reading
effect to be distinguished on its own. On the and writing disabilities are provided herein.
other hand, if an allele of one of the genes
has a more pronounced (major) effect on Segregation Analysis
the phenotype than do the other genes, the
gene may be identified even if the overall One way to evaluate the mode of inheri-
heritability of the trait is not high. tance of a trait in families is through segre-
gation analysis. The method of ascertain-
Traits with Mendelian inheritance may be ment of families for these studies is of
autosomal or sex linked and dominant, paramount importance because systematic
codominant, or recessive. Although some bias can lead to a model of inheritance that
observational family studies suggested a does not reflect the actual transmission pat-
simple autosomal dominant mode of trans- tern of the trait in the larger, unselected,
mission of reading disability (e.g., Drew, population (see Smith & Goldgar, 1986, for
1956; Hallgren, 1950) and a genetic linkage a discussion regarding ascertainment correc-
study in a single large family found evidence tion). In essence, segregation analysis evalu-
for involvement of a gene with autosomal ates a variety of possible modes of inheri-
dominant effect (Fagerheim et al., 1999), tance for their ability to explain the
the growing consensus is that the most observed familial pattern of the trait. In its
prevalent pattern of dyslexia in pedigrees simplest form, the proportion of affected
does not fit a simple Mendelian mode of in- children of probands (the segregation ratio)
heritance (e.g., Finucci et al., 1976; Hohnen is computed and tested to evaluate whether
& Stevenson, 1999; Omenn & Weber, the observed ratio fits the expected one
1978). In contrast to a simple Mendelian (Smith & Goldgar, 1986). For example, in
trait in which one gene’s alleles determine a the classic transmission of an autosomal
phenotype, in complex disorders the geno- dominant trait with no environmental influ-
types of a set of genes act in concert with ence and 100% penetrance, the expected
environmental factors to affect the proba- segregation ratio is .50 because it is expect-
bility that an individual will manifest the ed that half the offspring of matings in
phenotype; alterations of a single gene are which one parent is affected and the other is
not sufficient to cause the disorder. In addi- not affected would be affected if enough
tion, for most complex disorders, the popu- families are studied.
lation cannot simply be categorized as “af-
fected” or “unaffected” because there is a In complex segregation analysis (CSA), a
range of manifestations. In fact, when the mathematical equation describing a more
genes are cloned and studied in detail, even complete model of the inheritance pattern
many “simple” disorders are found to be of a single gene is generated. In this method
much more complex than previously the potential contributions of other genes
thought. Their phenotypic effects can be are combined into one polygenic term,
quite varied, even without the involvement terms for both shared and unshared envi-
of independent modifying genes and envi- ronment effects can also be included, and a
ronmental factors (e.g., Cohn, Bornstein, & single model is chosen that best fits the data
Jowell, 2000; Gayther et al., 1997; Quinzii (Bonney, 1986; Morton & MacLean,
& Castellani, 2000). 1974). An estimate of the number of genes
that are involved can be obtained using a
Although data from the Colorado Read-

260 CAUSES AND BEHAVIORAL MANIFESTATIONS

newly devised sophisticated Bayesian measure of phonological nonword memory
Monte-Carlo Markov-Chain stochastic using both complex and MCMC segrega-
method (MCMC) that models several tion analysis methods.
Mendelian loci simultaneously but does not
include a polygenic component (e.g., Daw, Linkage Analyses
Heath, & Wijsman, 1999; Heath, 1997;
Wijsman et al., 2000). If there is a gene of major effect that con-
tributes to a phenotype, it is possible to lo-
Hallgren (1950) was the first to use segre- cate its position in the genome by a process
gation analysis in his sample of 116 reading called linkage analysis. When the mode of
disabled children and their families, finding transmission of the phenotype is known,
no evidence for a sex-linked mode of inheri- powerful model-based linkage methods can
tance but good agreement with an autosomal be employed. Parameters estimated in segre-
dominant mode of transmission. Finucci and gation analyses, such as gene frequency, the
colleagues (1976) studied 20 families and relative effects of genetic and nongenetic in-
found that the mode of transmission did not fluences, and potential heterogeneity, can be
fit any one mode of single gene inheritance. incorporated into these linkage analyses.
Lewitter, DeFries, and Elston (1980) found On the other hand, when the mode of trans-
that although there was no evidence for a mission of a phenotype is not known, non-
single gene mode of transmission in a set of parametric methods, such as sib- and other
133 families, plausible models could explain affected relative-pair analyses, which are
several subsets of these families. For exam- relatively model free, must be used. Regard-
ple, the data were compatible with a major less of the analysis method, the question
recessive gene for reading disability in fami- asked during each step in a linkage study is
lies with female probands. In a study of four whether two loci segregate independently in
independently ascertained samples, Penning- meiosis.
ton and colleagues (1991) obtained results
consistent with a sex-influenced, dominant, During meiosis, chromosomal segments
or major gene effect in three of four samples derived from the individual’s parents are ex-
and with polygenic transmission in the changed in a process, “crossing over,”
fourth. The latter two studies highlight the which recombines the DNA that had been
importance of careful phenotyping of fami- inherited from the individual’s parents. The
lies because the identification of reliable sub- ovum thus contains DNA from both parents
types could help in the identification of dis- of the mother and the sperm contains DNA
tinct genetic etiologies. from both parents of the father. Therefore,
an individual in the next generation inherits
Because reading ability is normally dis- some alleles from each of the four grandpar-
tributed in the population (Shaywitz, Esco- ents. What makes linkage analyses feasible
bar, Shaywitz, Fletcher, & Makuch, 1992), is the existence of DNA sequence differ-
the boundary between abnormal and nor- ences between individuals. In essence, the
mal reading is indistinct and analyses using locus responsible for the phenotype of inter-
quantitative data may be more appropriate. est (affected or unaffected, in the case of a
In a population of 125 nuclear families as- categorical phenotype) whose place in the
certained through a reader without disabili- genome is unknown is evaluated with a se-
ties, Gilger, Borecki, DeFries, and Penning- ries of polymorphic loci (markers) whose
ton (1984) reported evidence that low places in the genome are known. When two
penetrant common alleles of one or more loci are far apart, for instance, on different
dominant genes contributed a significant chromosomes, their alleles segregate inde-
proportion of the variance to a continuous pendently from each other. Assume, for ex-
measure of reading ability based on the ample, that a phenotype being studied re-
Reading Recognition, Reading Comprehen- sults from a mutation (–) in a gene on one
sion, and Spelling subtests of the Peabody chromosome pair and there is a marker lo-
Individual Achievement Test (Dunn & cus on another chromosome pair with alle-
Markwardt, 1970). In addition, Wijsman les A and a (the genotype of the affected
and colleagues (2000) reported evidence for parent is +/– at the gene and A/a at the
the presence of one or more genes con- marker). Only one chromosome of each
tributing to performance on a continuous

Genetic Influences on Reading and Writing Disabilities 261

pair is transmitted to each offspring. The linkage hypothesis (Morton, 1955). It has
possible outcomes of independent segrega- been suggested that more stringent signifi-
tion of the phenotype and the marker locus cance levels be required to reach the same
are –/A and –/a and each is equally likely to conclusions for a complex trait (Lander &
be observed. In contrast, alleles at loci that Kruglyak, 1995; Thomson, 1994). Linkage
are close to each other on the same chromo- studies also have the capacity to detect ge-
some will stay together much more often netic heterogeneity in the sample and to es-
than would be predicted by chance. For ex- timate the proportion of families in which
ample, if the allele causing the phenotype is the trait locus appears to be linked to a spe-
on the same chromosome as allele A of the cific location. The linkage analysis can then
linked marker locus, offspring with the phe- be repeated in the subset of unlinked fami-
notype will more often inherit allele A than lies to map another locus.
allele a. If cosegregation of alleles of the
known locus and the phenotype of the un- Association Studies and Linkage
known locus is observed more frequently Disequilibrium Mapping
than 50% of the time, the chromosomal po-
sition of the unknown gene relative to the For the linkage analysis just described, the
known locus can be estimated. The distance specific marker alleles are not relevant to
between two locations in the genome is of- the phenotype. They merely allow the trans-
ten expressed as the fraction of gametes in mission of the phenotype to be analyzed
which there has been a recombination (i.e., concurrently with the segregation of a locus
the recombination fraction, ␪). in the genome in each family. Another cate-
gory of analytic methods relies on the phe-
In the affected relative pair method, usu- nomenon of linkage disequilibrium—that is,
ally done on sib pairs, constraints are not the association of the phenotype with a spe-
placed on the model of transmission. It is cific allele at a marker. Such an association
assumed that if the trait locus and the mark- will occur when the allele has an effect on
er are linked, the sib pair will be concordant the phenotype or when there has not been
for both the trait and an allele at the mark- sufficient time for the specific genetic
er. If the phenotype is linked to the marker, change responsible for the phenotype and
there should be a correspondence between an allele at a closely linked marker to reach
the phenotypical and genotypical similari- equilibrium. It takes time (in the form of
ties between the relatives regardless of the generations) for a new allele that arises by
mode of inheritance. mutation at one site to be separated by the
process of recombination from the specific
Model-based methods are more powerful alleles at a nearby site on the same chromo-
than sib-pair methods, at least in part be- some. For a while, these alleles will be trans-
cause information about all members of the mitted together in a unit called a haplotype.
family, affected and not affected, can be A state of equilibrium exists when the alle-
used. A ratio of the probability of linkage to les at both loci are distributed in the popula-
the probability of nonlinkage, under the tion as is predicted for the distance between
proposed model of transmission, is comput- the two loci. Because the frequencies of alle-
ed at several recombination frequencies for les at a locus may vary considerably be-
each family. The log of the ratios, called the tween different ethnic populations, the
LOD score (log of the odds of likelihood of analysis must control for ethnicity (Schaid
linkage), are calculated so that LOD scores et al., 1999; Terwilliger & Ott, 1992).
at various ␪ levels can be summed across
families. Given the large number of genes Genetic Mapping and Association Studies in
and markers in the genome, the a priori Reading and Writing Disabilities
likelihood of linkage of any two locations is
extremely low. Therefore, the hypothesis of Given the phenotypical heterogeneity of
linkage is accepted (at the p < .05 level) reading and writing disorders, it is not sur-
when a LOD score greater than 3.0 is ob- prising that linkage results have also been
tained (1,000:1 odds in favor of linkage). diverse. Potential locations for genes influ-
For a simple Mendelian trait, an LOD score encing reading and spelling disorders have
of less than –2.0 (100:1 odds against link-
age) is conventionally used to reject the

262 CAUSES AND BEHAVIORAL MANIFESTATIONS

been reported on chromosomes 1, 2, 6, 15, ed with deficits in oral prose reading in a re-
and 18, as reviewed later and partially sum- cent study of parent–proband trios (Morris
marized previously in a table (Raskind, et al., 2000). In addition, there is suggestive
2001). It is important to note that because evidence for the localization to chromosome
of differences in ascertainment procedures, 15 of a gene that modifies the expression of
measures used to define reading/spelling dis- dyslexia in a family whose dyslexia appears
orders, and analytic methods, these studies to be linked to chromosome 2 (Fagerheim
are not directly comparable. et al., 2000). Finally, in one family, three
of four individuals carrying a balanced
Smith, Kimberling, Pennington, and Lubs translocation involving chromosome band
(1983) were the first to report a significant 15q21 had dyslexia (Nopola-Hemmi et al.,
linkage finding for reading disability. In this 2000). However, in a second family, only
initial study of eight families chosen for ap- one of four carriers of a translocation in-
parent autosomal dominant transmission of volving the same region of chromosome 15q
dyslexia, a dichotomous classification of was affected by dyslexia. Not all researchers
reading disability was used based on oral have been able to replicate the chromosome
reading and written spelling at least 2 years 15 linkage findings (Bisgaard, Eiberg,
below expected grade level (determined by Moller, Niebuhr, & Mohr, 1987; Sawyer et
mathematics and general information tests) al., 1998), even when the original families
for children and history for adults. Using a were included in the sample studied (Car-
cytogenetically observable polymorphism at don et al., 1994; Lubs et al., 1991; Rabin et
the proximal portion of the long arm of al., 1993).
chromosome 15 as the marker, a LOD score
of 3.241 at a recombination fraction of 0.13 One of these groups reported suggestive
was found. The putative dyslexia locus was evidence for linkage to the distal portion of
assigned the name DYX1. However, when the short arm of chromosome 1 (1p34–36)
additional families were included and DNA in nine of the extended families with a
markers were used, the overall LOD score three-generation history of reading disabili-
decreased and there was evidence of signifi- ty, including the one that had previously
cant genetic heterogeneity, with about 20% provided much of the evidence for linkage
of the families showing linkage to chromo- to chromosome 15 (Lubs et al., 1991; Rabin
some 15 but the remainder not demonstrat- et al., 1993). Supporting evidence for a
ing linkage (Smith, Pennington, Kimberling, dyslexia locus on the short arm of chromo-
& Ing, 1990). A single large family provided some 1 was found in another set of extend-
most of the evidence for linkage. There was ed families with deficits in single word read-
some evidence for a chromosome 15 local- ing, phonological decoding, and rapid
ization when nonparametric multiple regres- automatized naming (Grigorenko et al.,
sion techniques were applied to the quantita- 2001). However, no study has obtained sig-
tive scores in a subset of the most severely nificant evidence for linkage to chromo-
affected sibships (Fulker et al., 1991). some 1p, and at least one study reported no
evidence for linkage to this region (Cardon
The chromosome 15 locus was subse- et al., 1994).
quently evaluated by other groups for link-
age to a variety of reading and writing phe- Evidence for a dyslexia locus on the short
notypes. Significant evidence for linkage arm of chromosome 6, DYX2, has also been
was found by parametric methods for a di- reported. This region had been targeted for
chotomous phenotype based on deficits in analysis because of earlier reports of an in-
single word reading in a sample of six ex- creased prevalence of autoimmunity in fami-
tended families containing at least four af- lies of children with learning disabilities
fected members (Grigorenko et al., 1997). (Geschwind & Behan, 1982; Pennington,
Results for a phenotype based on spelling Smith, Kimberling, Green, & Haith, 1987),
deficits, categorically defined, were consis- although subsequent research has not sup-
tent with linkage to chromosome 15, but ported this association (Gilger et al., 1998).
the LOD scores were too low to provide in- The first suggestion of linkage here was re-
dependent confirmation of this localization ported by Smith, Kimberling, and Penning-
(Schulte-Körne et al., 1998). Markers in this ton (1991) in a study of a subset of families
region have also been significantly associat- whose linkage to chromosome 15 (as report-

Genetic Influences on Reading and Writing Disabilities 263

ed previously) was questioned. Using the sib- dence for a locus on the short arm of chro-
pair analysis method specialized for DZ mosome 18 (18p11.2) for single-word read-
twins (DeFries & Fulker, 1985), Cardon and ing ability. Weaker evidence for linkage to
colleagues (1994, 1995) found evidence for the same site was found for measures of or-
linkage to 6p for a continuous scale discrim- thographic (irregular word reading and or-
inant score derived from performance on thographic choice) and phonological
measures of reading recognition, reading (phonological awareness and phonological
comprehension, and spelling. In a new sam- decoding) processing. This localization was
ple of sibs, modest supporting evidence was confirmed in an independent sample of fam-
obtained for a quantitative trait locus on 6p ilies from the United Kingdom for the phe-
for an orthographic task involving discrimi- notype of phonological awareness. The U.S.
nation of a real word and a nonword homo- sample provided suggestive evidence for a
phone (orthographic choice) and for phono- second locus on the long arm of chromo-
logical decoding (Gayán et al., 1999). In some 18. It is also interesting that some evi-
targeted studies, this localization was repli- dence for a locus on chromosome 2p15-p16
cated by two other groups. The phenotypes emerged in both the U.K. and U.S. samples,
providing the evidence included categorical- suggesting that DYX3 may be involved in
ly defined phonological segmentation (Grig- the broader population of dyslexia than just
orenko et al., 1997), real word reading, and the rare family with a monogenic form of
spelling/vocabulary (Grigorenko et al., reading disability (Fagerheim et al., 1999).
2001) subtypes and continuous measures of
ability to read irregular real words (ortho- This increasing list of genomic locations
graphic processing) and phonological decod- for genes involved in reading and spelling
ing of nonwords (Fisher et al., 1999). How- disabilities is not unexpected given the com-
ever, several other groups found no evidence plex nature of the phenotype. The relative
for linkage to 6p for a measure of phonolog- contribution of each of these putative genes
ical decoding (Field & Kaplan, 1998), for to reading and spelling disabilities is not
spelling disabilities (Schulte-Körne et al., known, nor is it clear how they interact to
1998), or for measures of phonological produce the variation in the phenotype that
awareness, phonological decoding, spelling, is seen within as well as between families.
and rapid automatized naming (Petryshen, To answer these questions, it will be neces-
Kaplan, Liu, & Field, 2000). sary to identify the genes. The genetics of
learning disabilities might be more tractable
In the first genome-wide scan for a if the phenotypes studied could be reduced
dyslexia locus, using a qualitative affecta- to simpler or more phenotypically precise
tion status, significant evidence for linkage subphenotypes. If genetic models can be
to a locus on chromosome 2q was found in constructed for these subphenotypes, it
a single large Norwegian family with an ap- would be possible to use model-based gene-
parent autosomal dominant mild to moder- mapping methods that are inherently more
ate reading and spelling disability (Fager- powerful than model-free methods and that
heim et al., 1999). This locus has been can be used to refine the localizations of the
assigned the name DYX3. Although mildly suggested dyslexia gene sites (Wijsman &
positive linkage results were obtained to the Amos, 1997). The University of Washington
pericentromeric region of chromosome 15, Multidisciplinary Learning Disabilities Cen-
where the DYX1 locus was originally speci- ter (UWMLDC) was established in 1995
fied, linkage to DYX2 was excluded in this and began ascertaining and phenotyping
family (Fagerheim et al., 1999, 2000). multigenerational families to address these
considerations.
A second genome-wide scan investigated
several quantitative reading-related pheno- Genetic Research at the University
types in two samples from the United States of Washington Multidisciplinary
and the United Kingdom, using two non- Learning Disabilities Center
parametric methods (Fisher et al., 2001).
Genotypes of parents and their children Probands were ascertained without regard
were included in the analyses, but quantita- to family size or family history of reading
tive data were available only for the sib-
lings. Both samples provided strong evi-

264 CAUSES AND BEHAVIORAL MANIFESTATIONS

and/or writing problems. These potential formation regarding these basic language
probands were administered an approxi- processes (phonological processing, ortho-
mately 4-hour battery of 23 tests that graphic coding, and rapid automatized
spanned the following domains: naming) was deemed necessary to under-
stand the various subphenotypes that may
1. Verbal comprehension. The Verbal exist within and between families.
Comprehension factor of the Wechsler
scales of intelligence (1981, 1992) was ad- Careful characterization of the phenotype
ministered to provide an overall indication was initially performed on 102 nuclear fam-
of verbal ability. Although there is contro- ilies containing 409 individuals (Berninger
versy regarding the use of verbal intelligence et al., 2001). Structural equation modeling
to qualify children for special education ser- was used to validate that four different lan-
vices within the school system, the use of guage processes—Verbal Comprehension,
verbal intelligence was warranted for genet- Orthographic Coding, Rapid Automatized
ic research for two reasons. First, the exclu- Naming, and Phonological Processing—
sion of children with verbal intelligence of contribute to the expression of the pheno-
less than 90 made it less likely that a sizable type but that different patterns of covari-
portion of the probands might have more ance across processes and different patterns
global impairment in cognition due to neu- of paths from these language processes to
rological, neurogenetic, or psychiatric disor- the academic achievement measures were
ders, rather than a specific disability in the found within child probands and between
area of reading or writing. Second, previous child probands and their affected parents. It
genetic research has provided evidence that was found that the phonological processing
reading disabilities defined by a discrepancy factor had significant paths to all reading
criterion may be more heritable than read- and writing skills except reading rate and
ing disabilities as defined by low perfor- handwriting in both probands and adults,
mance alone (Olson, Datta, Gayan, & De- but in adults only if the effect of verbal IQ
Fries, 1999), and reading disability in was removed. For the orthographic coding
groups of children with higher intelligence factor, significant paths to all reading and
may be more heritable than reading disabili- writing factors with the exception of read-
ty in groups of children with lower intelli- ing comprehension were found in both
gence (Wadsworth et al., 2000). probands and adults. Rapid naming had a
significant path only to reading rate in
2. Academic achievement. Several mea- probands and adults. Verbal IQ had signifi-
sures of reading accuracy, reading rate, cant paths to reading comprehension in
reading comprehension, handwriting, probands and adults, and to reading accura-
spelling, written composition, mathematics cy, reading rate, spelling, and composition
calculation, and mathematical problem in adults but not probands. These data
solving were also administered. If a child demonstrate that the same set of language
demonstrated verbal comprehension skills processing skills are used differently across
of at least SS = 90 and academic achieve- academic tasks and across ages.
ment of one standard deviation below ver-
bal ability, in addition to academic perfor- Additional phenotyping studies have also
mance below the mean, the child was been carried out with the data from these
included in our study as a proband and original 102 families. In a study of the co-
their family was invited to be tested (see morbidity of calculation and reading dis-
Berninger, Abbott, Thomson, & Raskind, abilities, it was found that children with
2001, for complete details regarding tests both disabilities were more likely to have a
administered and ascertainment proce- rapid automatized naming deficit than were
dures). other children in the study, but they did not
differ on measures of verbal IQ or phono-
3. Language processing. In addition to logical processing (Busse, Thomson, Ab-
verbal ability and academic achievement, a bott, & Berninger, 1999). This observation
large battery of language-processing mea- suggests that a deficit in phonological pro-
sures was also administered to gain infor- cessing may lead to a reading disability but
mation regarding the underlying processes that deficits in automaticity may lead to
involved in reading or writing disability. In-

Genetic Influences on Reading and Writing Disabilities 265

dual disability with calculation and reading. measures may be accounting for the ob-
In a further analysis, it was found that those served aggregation patterns of others. If a
children with the dual reading and calcula- promising aggregation pattern disappeared
tion disability actually had a triple deficit in when another measure was included as a
reading, calculation, and written expression covariate, perhaps the covariate measure
and were more severely affected in all three was accounting for the promising aggrega-
domains (as well as language processing tion pattern in the first measure. Using co-
markers) than those with a reading disabili- variate analyses, we found that there may
ty alone (Busse et al., 1999). be a genetic contribution to phonological
decoding rate in addition to the genetic con-
To evaluate the familial patterns of each tribution it shares with phonological decod-
measure in the battery, aggregation analyses ing accuracy, a genetic contribution to
of the quantitative data were performed phonological nonword memory in addition
(Raskind et al., 2000). Using a generalized to the genetic contribution it shares with
estimating equation (GEE)-based approach written spelling from dictation, a genetic
that can handle correlated data, such as var- contribution to written spelling from dicta-
ious measurements on the same subject or tion in addition to the genetic contribution
on related subjects, and that is not influ- it shares with phonological decoding accu-
enced by the different size of pedigrees racy, and a genetic contribution to self-regu-
(Liang & Zeger, 1986; Zhao, Grove, & lation of cognition (inattention ratings) in
Quaioit, 1992), correlations between nu- addition to the genetic contribution it shares
clear family members were computed for all with rapid automatized naming or oral
measures to identify patterns consistent reading rate (Hsu, Wijsman, Berninger,
with a genetic basis. If there is a genetic Thomson, & Raskind, 2002).
basis for a trait, the correlations between
relatives should reflect the degree of genetic Although the familial patterns were con-
relatedness. Therefore, the correlation be- sistent with an underlying genetic compo-
tween siblings should exceed the correlation nent, aggregation analyses do not provide
between parents and their children, and information about the mode of inheritance
both correlations should be positive. In con- or the number of genes involved. To investi-
trast, unless there is assortative mating or gate these questions, two candidate pheno-
inbreeding, the trait should show no corre- types, phonological nonword memory and
lation between the parents. Two measures digit span, were chosen for further investi-
were identified that met these criteria with gation in more labor-intensive segregation
a statistical significance at p < 0.05 for analyses (Wijsman et al., 2000). This re-
both sibling and parent–offspring correla- search diverged from other segregation
tions: rate or automaticity of phonological studies in that we evaluated only one com-
decoding and phonological memory for ponent process of reading/writing disability
nonwords. Three measures had slightly rather than attempting to analyze the genet-
weaker patterns with one correlation signif- ic contribution of reading or writing skills
icant at p < 0.05 and the other significant as a whole. In single-measure analyses, we
at p < 0.1: accuracy of phonological decod- found evidence supporting a major-gene
ing, short-term memory for digits, and one mode of inheritance for both the nonword
measure of written spelling from dictation. memory task and digit-span task. Results
Five additional measures showed at least obtained by reciprocal adjustment of the
suggestive evidence for an appropriate ag- measures suggested that the genes con-
gregation pattern with only one of the two tributing to the nonword memory score
correlations significant at p < 0.1: another completely account for the genetic basis of
measure of written spelling from dictation, the digit-span score but that there is an ad-
accuracy, fluency and comprehension of ditional genetic contribution to the non-
oral reading, and one measure of rapid au- word memory task that is not accounted for
tomatized naming. by the digit-span measure.

Aggregation analyses were then per- Unraveling the genetics of complex phe-
formed on pairs of related measures in or- notypes such as reading and writing disabil-
der to gain information regarding their in- ities will require large data sets. Mapping
terdependence. We reasoned that some broad genomic locations for contributing

266 CAUSES AND BEHAVIORAL MANIFESTATIONS

genes is an important first step. Subsequent Autosome: Any nuclear chromosome other than the
steps include refinement of the locations, sex chromosomes. There are 22 pairs of auto-
candidate gene analysis or positional somes.
cloning, mutational analysis, genotype/phe-
notype/environment correlations, and Chromosome: A threadlike structure in the cell’s
searches for modifying genes. Therefore, nucleus that contains DNA. There are normally
continued ascertainment and assessment of 46 chromosomes in each cell organized into 23
study subjects is a priority. The study sam- pairs.
ple currently consists of 183 families con-
taining 1,093 first-degree and more distant Complex trait inheritance: A pattern of inheritance
relatives. Additional phenotypes, such as that is not Mendelian and usually reflects in-
morphology (especially derivational suffixes volvement of alleles at more than one locus inter-
that convey grammatical information) and acting with environmental factors.
syntax, are being analyzed for the possibili-
ty of a genetic etiology, and segregation and Concordance: The presence of the same trait in a set
linkage analyses are being performed for of individuals, such as a pair of twins.
promising subtypes.
Dichotomous trait: A phenotype that is either pre-
Conclusion sent or absent.

Research on the genetic contribution to read- Discordance: The presence of a trait in only one
ing and writing disabilities has progressed member of a set of individuals.
substantially over the years, from early ob-
servational studies of its familial aggregation DNA (deoxyribonucleic acid): The molecule com-
to sophisticated linkage analyses made feasi- posed of four different elements (bases) that en-
ble through advances in the Human Genome code genetic information for the individual and
Project and development of powerful genetic that allows transmission of this genetic informa-
statistical methods. Several possible sites on tion from one generation to another.
five different chromosomes (1, 2, 6, 15, and
18) have been implicated in the expression of Dominant: One copy of the trait-causing form of
reading/writing disorders in a number of the gene is sufficient for the phenotype.
families, but no etiological genes have yet
been identified. The next decade promises to Epigenetic: A factor that can affect the phenotype
be an exciting time in the history of dyslexia without any change in the genotype.
research when continued enlargement of the
study populations and development of even Epistasis: Technically means that an allele of one
more powerful analytic methods culminate gene can block the phenotypical expression of all
in the cloning of these genes. This milestone alleles of another gene. Often used to imply a
will usher in a new phase of research into gene that has an effect on the expression of the al-
genotype/phenotype correlations, biological leles at another gene.
consequences of specific genetic changes,
and potential intervention strategies guided Expressivity: The degree to which a heritable trait is
by genetic profiles. exhibited in an individual. When different pheno-
types or severity of the phenotype may result
Appendix 16.1. Glossary from the action of a gene, the gene is said to man-
ifest “variable expressivity.”
Allele: One of the alternative versions of a gene or
marker. Gene: A sequence of DNA that contains all the in-
formation necessary for synthesis of an RNA
Ascertainment: The method of selection of individu- molecule, which, in turn, contains the informa-
als for inclusion in a study. tion to produce a protein product.

Association: In human genetics, this describes the Genetic heterogeneity: More than one genetic mech-
situation in which a particular allele is found anism can produce the same phenotype.
more or less often in a group of affected indivuals
than would be expected by chance in the general Genetic statistical modeling: A process that seeks to
population. describe in mathematical terms the inheritance
pattern of an observed phenotype.

Genome: All the DNA of a cell or individual.
Genotype: The alleles at a specific locus or portion

of the genome.
Haplotype: A set of closely linked loci that tend to

be inherited toegether as a unit.
Heritability: The proportion of the variation in a

trait that is attributable to genetic (or inherited)
factors.
Linkage analysis/Gene mapping: A statistical
method in which the chromosomal location of a
gene responsible for a trait is identified by track-
ing the inheritance of a gene or phenotypical trait
and a DNA marker in a family or group of
families. The closer together two loci are (the
marker and the trait gene), the more often they
will stay together during meiosis. When two loci
are unlinked they will be transmitted indepen-
dently.
Locus: A position in the DNA sequence.

Genetic Influences on Reading and Writing Disabilities 267

Major gene: A gene that is responsible for a de- References
tectable portion of the variation in a phenotype.
Alarcón, M., & DeFries, J. C. (1995) Quantitative
Marker: A site in the genome that has a known lo- trait locus for reading disability: An alternative
cation and more than one form (allele). It may be test. Behavior Genetics, 25, 253.
a gene or a DNA sequence that is not part of a
gene. Bakwin, H. (1973). Reading disability in twins. De-
velopmental Medicine and Child Neurology, 15,
Mendelian inheritance: The pattern of a phenotype 184–187.
in a family that can be explained by transmission
of a locus that follows Mendel’s first law of inde- Berninger, V. (2000). Dyslexia, the invisible, treat-
pendent segregation; that is, at a single locus, the able disorder: The story of Einstein’s Ninja Tur-
alleles that were inherited from the two parents tles. Learning Disability Quarterly, 23, 175–
are independently segregated during transmission 195.
to the offspring. The phenotype may be autoso-
mal (the locus for the gene is on an autosome) or Berninger, V. W., Abbott, R. D., Thomson, J. B., &
sex linked (the locus for the gene is on the X or Y Raskind, W. H. (2001). Language phenotype for
chromosome) and dominant, codominant, or re- reading and writing disability: A family ap-
cessive. proach. Scientific Studies of Reading, 5, 59–106.

Oligogenic inheritance: The effects of several genes Bisgaard, M. L., Eiberg, H., Moller, N., Niebuhr, E.,
combine to cause the phenotype. & Mohr, J. (1987). Dyslexia and chromosome 15
heteromorphism: negative LOD score in a Danish
p: The short arm of a chromosome. material. Clinical Genetics, 32, 118–119.
Penetrance: The proportion of individuals who have
Bonney, G. (1986). Regressive logistic models for
the trait-associated genotype who exhibit the trait familial disease and other binary traits. Biomet-
phenotype. This number is expressed as a percent rics, 42, 611–625.
and may differ with age of the individuals.
Phenotype: The phenotype is what is observed. It Busse, J., Thomson, J., Abbott, R., & Berninger, V.
can be a physical attribute, such as eye color, or a (1999, August). Cognitive processes related to
measurable attribute, such as performance on a dual disability in reading and calculation. Paper
test. It can be defined categorically (blue eyes or presented at the annual meeting of the American
brown eyes, affected or unaffected) or quantita- Psychological Association, Boston.
tively (blood pressure, verbal IQ).
Proband: The affected person through whom the Cardon, L. R., Smith, S. D., Fulker, D. W., Kimber-
other family members are ascertained. ling, W. J., Pennington, B. F., & DeFries, J. C.
Polygenic inheritance: The phenotype is determined (1994). Quantitative trait locus for reading dis-
by many genes at different loci, each with small ability on chromosome 6. Science, 266, 276–279.
additive effects.
Polymorphism: A common variation in the se- Cardon, L. R., Smith, S. D., Fulker, D. W., Kimber-
quence of DNA at the same site among individu- ling, W. J., Pennington, B. F., & DeFries, J. C.
als. (1995). Quantitative trait locus for reading dis-
q: The long arm of a chromosome. ability: correction [Letter]. Science, 268, 1553.
Qualitative trait: A trait that is categorical, often di-
chotomous. The individual either has the trait or Castles, A., Datta, H., Gayan, J., & Olson, R. K.
does not have it. (1999). Varieties of developmental reading disor-
Quantitative trait: A trait that shows a continuous der: Genetic and environmental influences. Jour-
and quantitative distribution in the population. nal of Experimental Child Psychology, 72,
Quantitative trait locus: A chromosomal region that 73–94.
contains a gene that contributes part of the vari-
ance of an oligogenic trait. Cohn, J. A., Bornstein, J. D., & Jowell, P. S. (2000).
Recessive: All copies of the gene must have the trait- Cystic fibrosis mutations and genetic predisposi-
causing form for the phenotype to be expressed. tion to idiopathic chronic pancreatitis. Medical
Recurrence risk: The probability that a genetic dis- Clinics of North America, 3, 621–631, ix.
order present in a member of a family will recur
in another member of the same family or a later Daw, E. W., Heath, S. C., & Wijsman, E. M.
generation. (1999). Multipoint oligogenic analysis of age-at-
Sex chromosomes: The X and Y chromosomes. onset data with applications to Alzheimer’s dis-
Segregation analysis: A statistical genetic method ease pedigrees. American Journal of Human Ge-
that assesses the pattern of a trait in families to netics, 64, 839–851.
determine the most likely mode of inheritance.
Decker, S. N., & DeFries, J. C. (1981). Cognitive
Acknowledgment abilities in families with reading disabled chil-
dren. Developmental Medicine and Child Neu-
Both authors are supported by NIH Grant No. P50 rology, 23, 217–227.
HD33812.
DeFries, J. C., & Fulker, D. W. (1985). Multiple re-
gression analysis of twin data. Behavior Genetics,
15, 467–473.

DeFries, J. C., Fulker, D. W., & LaBuda, M. C.
(1987). Evidence for a genetic aetiology in read-
ing disability of twins. Nature, 329, 538–539.

DeFries, J. C., Singer, S. M., Foch, T. T., & Lewitter,
F. I. (1978). Familial nature of reading disability.
British Journal of Psychiatry, 132, 361–367.

Drew, A. L. (1956). A neurological appraisal of fa-

268 CAUSES AND BEHAVIORAL MANIFESTATIONS

milial congenital word-blindness. Brain, 79, ovarian cancer associated with different germline
440–460. mutations of the BRCA2 gene. Nature Genetics,
Dunn, L. M., & Markwardt, F. C. (1970). Peabody 15, 103–105.
Individual Achievement Test. Circle Pines, MN: Geschwind, N., & Behan, P. (1982). Left-handed-
American Guidance Service. ness: association with immune disease, migraine,
Fagerheim, T., Raeymaekers, P., Tonnessen, F. E., and developmental learning disorder. Procedings
Pedersen, M., Tranebjaerg, L., & Lubs, H. A. of the National Academy of Sciences USA, 79,
(1999). A new gene (DYX3) for dyslexia is locat- 5097–5100.
ed on chromosome 2. Journal of Medical Genet- Gilger, J. W., Borecki, I. B., DeFries, J. C., & Pen-
ics 36, 664–669. nington, B. F. (1994). Commingling and segrega-
Fagerheim, T., Raeymaekers, P., Tonnessen, F. E., tion analysis of reading performance in families
Sandkuijl, L. A., Lubs, H. A., & Tranebjaerg, L. of normal reading probands. Behavior Genetics,
(2000). A genome wide search for dyslexia loci in 24, 345–355.
a large Norwegian family. American Journal of Gilger, J. W., Pennington, B. F., Harbeck, R. J., De-
Human Genetics, 67(Suppl.), A1712. Fries, J. C., Kotzin, B., Green, P., & Smith, S.
Faraone, S. V., & Doyle, A. E. (2001). The nature (1998). A twin and family study of the associa-
and heritability of attention-deficit/hyperactivity tion between immune system dysfunction and
disorder. Child and Adolescent Psychiatric Clin- dyslexia using blood serum immunoassay and
ics of North America, 10, 299–316, viii–ix. survey data. Brain and Cognition, 36, 310–333.
Field, L. L., & Kaplan, B. J. (1998). Absence of link- Grayson, D. A. (1989) Twins reared together: Mini-
age of phonological coding dyslexia to chromo- mizing shared environmental effects. Behavior
some 6p23-p21.3 in a large family data set. Amer- Genetics, 19, 593–604.
ican Journal of Human Genetics, 63, 1448–1456. Grigorenko, E. L., Wood, F. B., Meyer, M. S., Hart,
Finucci, J. M., Guthrie, J. T., Childs, A. L., Abbey, L. A., Speed, W. C., Shuster, A., & Pauls, D. L.
H., & Childs, B. (1976). The genetics of specific (1997). Susceptibility loci for distinct compo-
reading disability. Annals of Human Genetics, nents of developmental dyslexia on chromosomes
40, 1–20. 6 and 15. American Journal of Human Genetics,
Fisher, J. H. (1905). Case of congenital word blind- 60, 27–39.
ness (inability to learn to read). Ophthalmologi- Grigorenko, E. L., Wood, F. B., Meyer, M. S., Pauls,
cal Review, 24, 315–318. J. E. D., Hart, L. A., & Pauls, D. L. (2001). Link-
Fisher, S. E., Francks, C., Marlow, A. J., MacPhie, I. age studies suggest a possible locus for develop-
L., Newbury, D. F., Cardon, L. R., Ishikawa- mental dyslexia on chromosome 1p. American
Brush, Y., Richardson, A. J., Talcott, J. B., Journal of Medical Genetics, 105, 120–129.
Gayan, J., Olson, R. K., Pennington, B. F., Smith, Hallgren, B. (1950). Specific dyslexia (congenital
S. D., DeFries, J. C., Stein, J. F., & Monaco, A. P. word blindness): A clinical and genetic study. Acta
(2001). Independent genome-wide scans identify Psychiatrica et Neurologica, 65(Suppl.), 1–287.
a chromosome 18 quantitative-trait locus influ- Heath, S. C. (1997). Markov chain Monte Carlo
encing dyslexia. Nature Genetics, 30, 86–91 segregation and linkage anlysis for oligogenic
Fisher, S. E., Marlow, A. J., Lamb, J., Maestrini, E., models. American Journal of Medical Genetics,
Williams, D. F., Richardson, A. J., Weeks, D. E., 61, 748–760.
Stein, J. F., & Monaco, A. P. (1999). A quantita- Hermann, K., & Norrie, E. (1958). Is congenital
tive-trait locus on chromosome 6p influences dif- word blindness a hereditary type of Gerstmann’s
ferent aspects of developmental dyslexia. Ameri- syndrome? Psyciatria et Neurologia, 136, 59–73.
can Journal of Human Genetics, 64, 146–156. Hinshelwood, J. (1907). Four cases of congenital
Foch, T. T., DeFries, J. C., McClearn, G. E., & word-blindness occurring in the same family.
Singer, S. M. (1977). Familial patterns of impair- British Medical Journal, 2, 1229–1232.
ment in reading disability. Journal of Educational Hohnen, B., & Stevenson, J. (1999). The structure
Psychology, 69, 316–329. of genetic influences on general cognitive, lan-
Fulker, D. W., Cardon, L. R., DeFries, J. C., Kim- guage, phonological and reading abilities. Devel-
berling, W. J., Pennington, B. F., & Smith, S. D. opmental Psychology, 35, 590–603.
(1991). Multiple regression analysis of sib-pair Hsu, L., Wijsman, E. M., Berninger, V. W., Thom-
data on reading to detect quantitative trait loci. son, J. B., & Raskind, W. H. (2002). Familial ag-
Reading and Writing: An Interdisciplinary Jour- gregation of dyslexia phenotypes: Paired correlat-
nal, 3, 299–313. ed measures. American Journal of Medical
Gayán, J., Smith, S. D., Cherny, S. S., Cardon, L. R., Genetics (Neuropsychiatric Genetics Section),
Fulker, D. W., Brower, A. M., Olson, R. K., Pen- 114, 471–478.
nington, B. F., & DeFries, J. C. (1999). Quantita- Joseph, J. (2001). Separated twins and the genetics
tive-trait locus for specific language and reading of personality differences: A critique. American
deficits on chromosome 6p. American Journal of Journal of Psychology, 114, 1–30.
Human Genetics, 64, 157–164. Kendler, K. S., Neale, M., Kessler, R., Heath, A., &
Gayther, S. A., Mangion, J., Russell, P., Seal, S., Bar- Eaves, L. (1993). A twin study of recent life
foot, R., Ponder, B. A., Stratton, M. R., & East- events and difficulties. Archives of General Psy-
on, D. (1997). Variation of risks of breast and chiatry, 50, 789–796.

Genetic Influences on Reading and Writing Disabilities 269

Lander, E., & Kruglyak, L. (1995). Genetic dissec- Pennington, B. F., Gilger, J. W., Pauls, D., Smith, S.
tion of complex traits: Guidelines for interpreting A., Smith, S. D., & DeFries, J. C. (1991). Evi-
and reporting linkage results. Nature Genetics, dence for major gene transmission of develop-
11, 241–247. mental dyslexia. Journal of the American Medical
Association, 266, 1527–1534.
Lewitter, F. I., DeFries, J. C., & Elston, R. C.
(1980). Genetic models of reading disability. Be- Pennington, B. F., Smith, S. D., Kimberling, W. J.,
havior Genetics, 10, 9–30. Green, P. A., & Haith, M. M. (1987). Left-hand-
edness and immune disorders in familial dyslexi-
Liang, K. Y., & Zeger, S. L. (1986). Longitudinal cs. Archives of Neurology, 44, 634–639.
data analysis using generalized linear models.
Biometrika, 73, 13–22. Petryshen, T. L., Kaplan, B. J., Liu, M. F., & Field,
L. L. (2000). Absence of significant linkage be-
Lubs, H. A., Duara, R., Levin, B., Jallad, B., Lubs, tween phonological coding dyslexia and chromo-
M.L., Rabin, M., et al. (1991). Dyslexia sub- some 6p23–21.3, as determined by use of quanti-
types—Genetics, behavior, and brain imaging. In tative-trait methods: Confirmation of qualitative
D.D. Duane & D.B. Gray (Eds.), The reading analyses. American Journal of Human Genetics,
brain: The biological basis of dyslexia (pp. 66, 708–714.
89–117). Parkton, MD: York Press.
Quinzii, C., & Castellani, C. (2000). The cystic fi-
Morris, D. W., Robinson, L., Turic, D., Duke, M., brosis transmembrane regulator gene and male
Webb, V., Milham, C., Hopkin, E., Pound, K., infertility. Journal of Endocrinological Investiga-
Fernando, S., Easton, M., Hamshere, M., tion, 10, 684–689.
Williams, N., McGuffin, P., Stevenson, J., Kraw-
czak, M., Owen, M. J., O’Donovan, M. C., & Rabin, M., Wen, X. L., Hepburn, M., Lubs, H. A.,
Williams, J. (2000). Family-based association Feldman, E., & Duara, R. (1993). Suggestive
mapping provides evidence for a gene for reading linkage of developmental dyslexia to chromo-
disability on chromosome 15q. Human Molecu- some 1p34-p36. Lancet, 342, 178.
lar Genetics, 9, 843–848.
Raskind, W. H. (2001). Current understanding of
Morton, N. (1955). Sequential tests for the detec- the genetic basis of reading and spelling dis-
tion of linkage. American Journal of Human Ge- ability. Learning Disability Quarterly, 24,
netics, 7, 277–318. 141–157.

Morton, N., & MacLean, C. (1974). Analysis of Raskind, W., Hsu, L., Berninger, V., Thomson, J., &
family resemblance. III. Complex segregation Wijsman, E. (2000). Familial aggregation of phe-
analysis of quantitative traits. American Journal notypic subtypes in dyslexia. Behavior Genetics,
of Human Genetics, 26, 489–503. 30, 385–399.

Nopola-Hemmi, J., Taipale, M., Haltia, T., Lehesjo- Sawyer, D. L., Krishnamani, M. R. S., Hannig, V.
ki, A. E., Voutilainen, A., & Kere, J. (2000). Two L., Garcia, M., Kim, J. K., Haines, J. L., &
translocations of chromosome 15q associated Phillips, J. A. (1998). Genetic analysis of phono-
with dyslexia. Journal of Medical Genetics, 37, logical core deficit dyslexia (PCDD). American
771–775. Journal of Human Genetics, 63, A1776.

Olson, R. K., Datta, H., Gayan, J., & DeFries, Schaid, D. J., Buetow, K., Weeks, D. E., Wijsman,
J. (1999). A behavioral-genetic analysis of E., Guo, S. W., Ott, J., & Dahl, C. (1999). Dis-
reading disabilities and component processes. In covery of cancer susceptibility genes: Study de-
R. M. Klein & P. A. McMullen (Eds.), Converg- signs, analytic approaches, and trends in technol-
ing methods for understanding reading and ogy. Journal of the National Cancer Institute
dyslexia (pp. 133–151). Cambridge, MA: MIT Monograph, 26, 1–16.
Press.
Schulte-Körne, G., Grimm, T., Nothen, M. M.,
Olson, R., Forsberg, H., & Wise, B. (1994). Genes, Muller-Myhsok, B., Cichon, S., Vogt, I. R., Prop-
environment, and the development of ortho- ping, & Remschmidt, H. (1998). Evidence for
graphic skills. In V. W. Berninger (Ed.), The vari- linkage of spelling disability to chromosome 15.
eties of orthographic knowledge I: Theoretical American Journal of Human Genetics, 63,
and developmental issues (pp. 27–71). Dor- 279–282.
drecht, The Netherlands: Kluwer Academic.
Shaywitz, S. E., Escobar, M. D., Shaywitz, B. A.,
Olson, R., Wise, B., Conners, F., Rack, J., & Fulker, Fletcher, J. M., & Makuch, R. (1992). Evidence
D. (1989). Specific deficits in component reading that dyslexia may represent the lower tail of a
and language skills: Genetic and environmental normal distribution of reading ability. New Eng-
influences. Journal of Learning Disabilities, 22, land Journal of Medicine, 326, 145–150.
339–348.
Smith, S. D., & Goldgar, D. E. (1986). Single gene
Omenn, G. S., & Weber, B. A. (1978). Dyslexia: analyses and their application to learning disabil-
Search for phenotypic and genetic heterogeneity. ities. In S. D. Smith (Ed.), Genetics and learning
American Journal of Medical Genetics, 1, disabilities (pp. 47–65). San Diego, CA: College-
333–342. Hill Press.

Owen, F. W., Adams, P. A., Forrest, T., Stolz, L. M., Smith, S. D., Kimberling, W. J., & Pennington, B. F.
& Fisher, S. (1971). Learning disorders in chil- (1991). Screening for multiple genes influencing
dren: sibling studies. Monographs of the Society dyslexia. Reading and Writing, 3, 285–298.
for Research in Child Development, 36, 1–77.
Smith, S. D., Kimberling, W. J., Pennington, B. F., &

270 CAUSES AND BEHAVIORAL MANIFESTATIONS

Lubs, H. A. (1983). Specific reading disability: for Children—Third edition. San Antonio, TX:
Identification of an inherited form through link- Psychological Corporation.
age analysis. Science, 219, 1345–1347. Wijsman, E. M., & Amos, C. I. (1997). Genetic
Smith, S. D., Pennington, B. F., Kimberling, W. J., & analysis of simulated oligogenic traits in nuclear
Ing, P. S. (1990). Familial dyslexia: Use of genetic and extended pedigrees: summary of GAW10 con-
linkage data to define subtypes. Journal of the tributions. Genetic Epidemiology, 14, 719–735.
American Academy of Child and Adolescent Psy- Wijsman, E. M., Peterson, D., Leutenegger, A. L.,
chiatry, 29, 204–213. Thomson, J. B., Goddard, K. A., Hsu, L.,
Stephenson, S. (1907). Six cases of congenital word- Berninger, V. W., & Raskind, W. H. (2000). Seg-
blindness affecting three generations of one fami- regation analysis of phenotypic components of
ly. Ophthalmoscope, 5, 482–484. learning disabilities: Nonword memory and digit
Terwilliger, J. D., & Ott, J. (1992). A haplotype- span. American Journal of Human Genetics, 67,
based “haplotype relative risk” approach to de- 631–646.
tecting allelic associations. Human Heredity, 42, Willcutt, E. G., Pennington, B. F., & DeFries, J. C.
337–346. (2000). Twin study of the etiology of comorbidity
Thomas, C. J. (1905). Congenital “word-blindness” between reading disability and attention-
and its treatment. Ophthalmoscope, 3, 380–385. deficit/hyperactivity disorder. American Journal
Thomson, G. (1994). Identifying complex disease of Medical Genetics, 96, 293–301.
genes: Progress and paradigms. Nature Genetics, Wolff, P. H., & Melngailis, I. (1994). Family pat-
8, 108–110. terns of developmental dyslexia: Clinical find-
Wadsworth, S. J., Olson, R. K., Pennington, B. F., ings. American Journal of Medical Genetics, 54,
& DeFries, J. C. (2000). Differential genetic etiol- 122–131.
ogy of reading disability as a function of IQ. Zhao, L. P., Grove, J., & Quiaoit, F. (1992). A
Journal of Learning Disabilities, 33, 192–199. method for assessing patterns of familial resem-
Wechsler, D. I. (1981). Wechsler Adult Intelligence blance in complex complex human pedigrees,
Scale—Revised. San Antonio, TX: Psychological with as application to the nevus-count data in
Corporation. Utah kindreds. American Journal of Human Ge-
Wechsler, D. I. (1992). Wechsler Intelligence Scale netics, 51, 178–190.

III

EFFECTIVE INSTRUCTION

This page intentionally left blank

17

Effective Remediation of
Word Identification and Decoding
Difficulties in School-Age Children

with Reading Disabilities

Maureen W. Lovett
Roderick W. Barron

Nancy J. Benson

I don’t have any trouble with the sentences; it’s the words that get in my way.
—A 9-YEAR-OLD STRUGGLING READER (2000)

The cognitive psychology of reading and its struction for these children in their struggle
contributing processes enjoys a long and to learn to read. In this chapter, we review
honorable scientific history spanning more the literature on reading acquisition difficul-
than 100 years, from the pioneering interests ties in children, evidence on the most persis-
of Cattell (1886) and Huey (1908) to current tent underlying learning deficits that appear
research in the cognitive neuroscience of central to reading disabilities, and what is
reading and its disorders (e.g., Pugh, Mencl, known to date of what constitutes effective
Jenner, Katz, Lee, et al., 2000; Pugh, Mencl, remediation.
Jenner, Katz, Frost, et al., 2001; Simos et al.,
2002) and molecular genetic studies of the Although the existence of developmental
transmission of reading ability and disability reading disabilities has been acknowledged
across generations (Grigorenko, 2001). As for the past century, it is only within the
Pennington (1997) phrased it, reading can be past 10–15 years that reliable evidence from
considered “a cognitive science success sto- controlled evaluations has been reported on
ry” (p. 13). its remediation. In the mid–1980s, there
was almost no scientifically credible evi-
Similarly, it has been recognized for more dence to indicate that reading disabilities
than a century that a sizable minority of were amenable to intervention or that any
otherwise intelligent, healthy, and well- one remedial approach was better than any
developing children experience unexpected other (Gittelman, 1983; Hewison, 1982;
failures in learning to read (Hinshelwood, Lovett, 1992). This dearth of evidence re-
1917; Morgan, 1896). Surprisingly, it has garding the effectiveness of remediation
only been in the last two decades that there over so many decades led to inevitable ques-
has been a scientific literature worth citing tions about the amenability of the disorder
with respect to what constitutes effective in- to treatment (Lovett, 1997, 1999).

273

274 EFFECTIVE INSTRUCTION

Methodological and Design reading skills before item difficulty levels
Limitations of Past Intervention rise and ceilings are quickly reached (Lovett,
Research on Reading Disabilities Hinchley, & Benson, 1997). As can be ex-
pected, experimental measures with more
Few studies were reported in the mid to late trials per level of difficulty result in more
1980s that evaluated the efficacy of inter- visible gains and better opportunities to
ventions in a controlled research design in demonstrate treatment-related change over
which alternative intervention approaches the short term. At present, many questions
were systematically compared. Failures to remain open regarding the best measure-
include a control group and to compare two ment models for evaluating the success or
or more interventions rendered reported ev- failure of a given intervention (Lyon &
idence as little more than anecdotal in im- Moats, 1997). The lack of power inherent
port (Lovett & Barron, 2002). Only when to traditional measurement choices has per-
alternative approaches are evaluated in con- haps masked the effects of some potentially
trolled designs is there an opportunity to promising remedial approaches.
separate treatment-specific effects from gen-
eral treatment effects (e.g., halo effects due Characterizing the Core Deficits
to inclusion in a special program, individu- Contributing to Reading Acquisition
alized attention, access to a teacher/thera- Failure and Assessing Children’s
pist with specialized reading disabilities ex- Response to Remediation
pertise) and change due to maturation and
experience. Despite these challenges, the contributions
of intervention researchers in the past 15
Until relatively recently, the intervention years have been scientifically sophisticated
literature on reading disorders has been at a level unprecedented in the field only
characterized by serious measurement and two decades ago. With an ambitious re-
methodological problems. There has been search funding program announced by the
limited recognition of the fact that outcome National Institute of Child Health and Hu-
is necessarily a multidimensional and multi- man Development (NICHD) in the United
variate construct, and its measurement is States in the early 1990s, clear incentives to
complex. Even in recent years, measurement conduct rigorous research on remedial in-
issues continue to plague otherwise success- terventions for phonologically based read-
ful intervention protocols yielding positive ing disorders became available (Lyon, 1995;
and specific intervention effects: It is ex- Lyon & Moats, 1997; Moats & Foorman,
tremely difficult, for example, to adequately 1997); and large-scale scientific research on
measure reading comprehension skills in a what constitutes effective remediation for
child with word identification problems and childhood reading disabilities started in
to evaluate whether word identification earnest. Advances in our understanding of
gains have yielded improved text reading the neurobiological and genetic substrates
and comprehension abilities (Berninger & of reading disability and other developmen-
Abbott, 1994; Levy, Abello, & Lysynchuk, tal learning disorders in the mid-1990s and
1997). The efficacy of an intervention re- beyond (Filipek, 1995, 1999; Grigorenko,
quires assessment with respect to the trans- 2001; Grigorenko et al., 1997; Pennington,
fer, generalization, and maintenance of its 1997, 1999; Shaywitz et al., 1998) have fo-
effects, yet basic questions remain regarding cused increased research attention on the
their reliable and valid measurement (Lyon, nature and underlying causes of these devel-
1996; Lyon & Moats, 1997; Shanahan & opmental learning disorders and the extent
Barr, 1995). to which the underlying cognitive process-
ing deficits can be ameliorated with effective
It is also a concern that the outcome mea- intervention methods.
sures that have been used in many studies
have varied enormously in their power and Many children who experience serious
sensitivity to treatment-related change. difficulty learning to read have precursor
Many standardized measures with steep problems in highly specific aspects of speech
item gradients, for example, allow relatively and language development. Prospective re-
few chances for an improving reader with
disabilities to demonstrate newly acquired

Word Identification and Decoding Difficulties 275

search studies have confirmed the relation- processing is thought to underlie the strug-
ship between early specific speech and lan- gling reader’s defining deficits in acquiring
guage difficulties and later reading disabili- alphabetic and phonologically based read-
ties in childhood (Bishop & Adams, 1990; ing skills (Brady, 1997; Lovett, 1997, 1999;
Gathercole & Baddeley, 1987; Scarbor- Lovett & Barron, 2003; Torgesen, Wagner,
ough, 1990, 1998). There has been remark- & Rashotte, 1997). The significance of
able consensus in the learning disabilities lit- these deficits in phonological processing has
erature in the past two decades that a core been emphasized by findings of different
language-related deficit associated with and profiles of brain activation while doing
predictive of reading acquisition failure in- phonologically demanding tasks—a distinct
volves a domain of linguistic competence “neurobiological signature” for individuals
known as “phonological awareness.” Chil- with reading disorders revealed in function-
dren with significant reading difficulties al neuroimaging studies (Shaywitz et al.,
typically exhibit a range of signature deficits 1998) which compare disabled and able
in their explicit awareness of and ability to readers of similar age and background. Be-
manipulate the sound structure of spoken haviorally and neurobiologically, the weight
words (Brady, 1997; Snowling & Hulme, of evidence identifies phonological process-
1993). Whether reading disorders are de- ing deficits, disruptions in word identifica-
fined on the basis of a significant discrepan- tion learning, and difficulty acquiring the al-
cy between measured IQ and actual reading phabetic principle as defining features of
achievement or simply on the basis of signif- developmental reading disabilities.
icant underachievement in reading based on
age or grade-level expectations, the deficits Deficits in these areas of speech and lan-
identified as potentially causal to this class guage development have been characterized
of learning disorders appear to be concen- as an “arrest in development” (Bruck,
trated within the word identification and 1992); and they are known to persist into
phonological processing domains (Fletcher adulthood for individuals with childhood
et al., 1994; Francis, Shaywitz, Stuebing, histories of reading disability (Bruck, 1992,
Shaywitz & Fletcher, 1996; Stanovich & 1998; Felton & Brown, 1990; Scarborough,
Siegel, 1994). 1984; Shaywitz et al., 1999). In one study,
even the highest-functioning adults with
Disabled readers have been characterized dyslexia exhibited lower levels of phonemic
as having a dysfunction “in the phonologi- awareness than did third-grade children
cal component of their natural capacity for with lower reading and spelling achieve-
language” (Liberman, Shankweiler, & ment (Bruck, 1992). Recognition of the
Liberman, 1989, p. 1). Phonological aware- depth and longevity of these phonological
ness represents a multifaceted and complex processing deficits in individuals with devel-
set of processing and metalinguisitic capa- opmental reading disorders led to concerns
bilities recognized to have different develop- about whether phonologically based read-
mental trajectories (Barron, 1998; Goswami ing disabilities were amenable to remedia-
& Bryant, 1990). Reading disabled individ- tion (e.g., Wagner, Torgesen, & Rashotte,
uals often experience significant difficulty 1994). New incentives and funding oppor-
segmenting and differentiating individual tunities from the NICHD encouraged scien-
speech sounds in spoken words, blending tists to conduct rigorous research evaluating
individual speech sounds to form a spoken remedial interventions for phonologically
word, and effectively using phonological based reading disabilities (Lyon, 1995; Lyon
codes to aid working memory performance & Moats, 1997; Moats & Foorman, 1997)
(Liberman & Mattingly, 1985; Mann, and early intervention programs for young
1986; Stanovich, 1991, 1994; Wagner & children constitutionally and environmen-
Torgesen, 1987). It is thought that all these tally at-risk for reading acquisition failure.
difficulties may stem from a more basic
problem in the ability to form phonological Early Intervention Studies
representations and to encode phonological
information accurately (Brady, 1997; Elbro, Because of these incentives, in the past
Borstrom, & Petersen, 1998). decade, several controlled and comparative

This basic dysfunction in phonological

276 EFFECTIVE INSTRUCTION

research studies have been reported assess- leagues (1998) provide important evidence
ing the efficacy of different approaches to that explicit instruction in letter–sound cor-
the remediation and/or prevention of read- respondences can prevent reading under-
ing acquisition problems in the early ele- achievement in children at risk for reading
mentary grades (Foorman, Francis, Fletcher, failure because of poor phonological aware-
Schatschneider, & Mehta, 1998; Foorman ness at school entry or a lack of literacy ex-
et al., 1997; Scanlon & Vellutino, 1997; periences in the home environment (see also
Torgesen et al., 1997, 1999; Vellutino, Scan- Juel & Minden-Cupp, 2000; Torgesen et al.,
lon, Sipay, et al., 1996). Foorman and col- 1999; Vellutino et al., 1996).
leagues (1998) conducted an important and
influential study assessing the reading devel- Torgesen and colleagues (1999) also re-
opment of 285 children in first and second ported a landmark early intervention study.
grades across 66 classrooms in several Title At-risk children entering kindergarten were
1 schools in Texas. These investigators included in the study, and the participants
sought to understand how the nature and were children who scored in the bottom
type of letter–sound instruction in early 12th percentile on kindergarten measures of
reading programs would interact with indi- letter knowledge and phonological aware-
vidual differences in the children’s entry-lev- ness. Children were randomly assigned
el skills in phonological awareness. Subjects to one of four instructional conditions.
were children who scored in the bottom The first condition provided phonological
18% on an early literacy assessment con- awareness and synthetic phonics training
ducted by the school district. (the Lindamood Auditory Discrimination in
Depth program) including explicit instruc-
Three types of experimental classroom tion in articulatory-based phonological
programs were compared with the standard awareness and much practice in phonetic
curriculum in the district. The three experi- decoding and word identification (ADD).
mental instruction programs differed in the The second condition also provided explicit
type of phonics instruction offered: One phonics instruction but allocated far more
provided direct instruction in letter–sound instructional time on application to the
correspondences and practice with decod- reading and writing of connected text mate-
able text (direct code); another offered less rials and the acquisition of a sight vocabu-
direct instruction in letter–sound correspon- lary (embedded phonics, or EP). A third
dences embedded in authentic literature condition provided regular classroom sup-
samples (embedded code); and a third pro- port (i.e., tutorial group), and the fourth
vided implicit instruction in the alphabetic was a no-treatment control group. Children
code while children read authentic text (im- in the experimental conditions received
plicit code). All teachers received ongoing one-to-one instruction in 20-minute ses-
training experiences specific to their pro- sions 4 days a week over 2½ years. At the
gram. Children from the direct code instruc- end of this 2½-year instructional period,
tional condition demonstrated better word children who had received the ADD in-
identification skills and steeper learning struction were the strongest readers when
curves in word reading than did those re- average scores were assessed. Both their
ceiving implicit code instruction. This ad- nonword reading (Word Attack) and word
vantage was greater for those children who reading (Word Identification) skills fell
entered the programs with the lowest levels overall within the average range. Their ad-
of phonological awareness. Although the di- vantage relative to the other groups was not
rect code instruction did not normalize the consistently established, however, on word
reading achievement of all instructed chil- identification and reading comprehension
dren, Torgesen (2000) derived a population- measures at the end of second grade. Torge-
based failure rate from this study computed sen and his colleagues note, in addition,
at 6%. As Fletcher and Lyon (1998) discuss, that there was substantial variability in re-
however, a failure rate of 6% for any in- sponse to the instruction, with 38% and
structional method represents a significant 39% of the ADD group scoring less than
reduction in the 15–20% of students with the 30th percentile on the word attack and
reading difficulty in the United States cur- word identification subtests, respectively.
rently. The findings of Foorman and col- Similar to calculations for the Foorman and

Word Identification and Decoding Difficulties 277

colleagues (1998) study, population-based the natural course of reading disorder fol-
estimates suggest that if the strongest condi- lows a deficit rather than a developmental
tion were to be applied more broadly, ap- lag model over time (Francis et al., 1996;
proximately 4% of children would remain Shaywitz et al., 1999). When children who
relatively weak in decoding ability and 5% met criteria for reading disability in grades
in sight word reading at the end of second 2 through 6 were compared with normally
grade (Torgesen, 2000). These investigators developing children with average and
concluded that despite enormous advances above-average reading skills in those grades,
in our understanding of the nature and eti- the persistently poor readers continued to
ology of reading disabilities, there remained demonstrate deficits in phonological coding
limited knowledge about how to effectively in adolescence and to experience continuing
remediate the more severe forms of devel- problems in reading, spelling, and reading
opmental reading disability and help these rate (Shaywitz et al., 1999). These results
children to become independent and fluent were considered “sobering” by the authors,
readers (Torgesen et al., 1997, 1999; Torge- particularly as children in the persistent
sen, Alexander, et al., 2001). poor reader group all had received special
education services. There was no evidence
Can the Deficits of Older Reading that children in the persistent RD group had
Disabled Children Be Remediated? caught up in their reading skills by high
school. In fact, another study has revealed
Given the persistence of reading disorders that typical special education placements in
and the negative sequelae of cumulative grades 4 and 5 accelerate students’ reading
reading experience deficits and more wide- development only by .04 standard devia-
spread academic difficulties in the later tions over the rate of growth achieved by
grades (Shaywitz et al., 1999; Stanovich, these children in their regular classroom
1986), it might be expected that it would be placements (Hanushek, Kain, & Rivkin,
easier to prevent reading acquisition failure 1998). Can the deficits of older disabled
in at-risk children than to remediate it in readers be effectively remediated?
older children diagnosed with reading dis-
ability. There is evidence, for example, that Remediation Studies with
first-grade reading achievement is a remark- Older Children
ably strong predictor of high school reading
achievement (Cunningham & Stanovich, Results from remediation studies focusing
1997), reinforcing the conclusion that the on the phonological reading deficits of older
gap between able and less able readers children with reading disabilities have also
widens rather than decreases with time and been reported in recent years (Olson, Wise,
development (Rayner, Foorman, Perfetti, Ring, & Johnson, 1997; Torgesen, Alexan-
Pesetsky, & Seidenberg, 2001). Research on der, et al., 2001; Wise & Olson, 1995; Wise,
adolescent and adult outcomes of children Ring, Sessions, & Olson, 1997). Solid con-
with developmental reading disorders indi- verging evidence is now available demon-
cates that reading ability does improve with strating that significant improvement can be
age and intervention for most individuals attained on speech-based and phonological
with reading disabilities, but that the signa- reading measures for both older children
ture deficits of reading disorder typically with reading disabilities and young children
persist into adulthood, particularly in more at significant risk for reading disability
severe cases. Phonological deficits persist, (Foorman et al., 1998; Lovett et al., 1994;
even in cases with relatively good literacy Lovett, Lacerenza, Borden, Frijters, et al.,
outcomes (Bruck 1992, 1998; 2000; Olson et al., 1997; Torgesen et al.,
Scarborough, 1984; Shaywitz et al., 1999). 1997, 1999; Torgesen, Alexander, et al.,
Problems also frequently persist with 2001; Vellutino et al., 1996); these results
spelling accuracy, word recognition speed, confirm that phonologically based decoding
and reading rate. and beginning reading skills are “teachable
aspects of reading for most children”
Prospective longitudinal data from the (Moats & Foorman, 1997, p. 188). Addi-
Connecticut Longitudinal Study reveals that