PERSPECTIVE
NATURE NEUROSCIENCE VOLUME 9 | NUMBER 10 | OCTOBER 2006 1213
CHILDHOOD DEVELOPMENTAL DISORDERS
From genes to behavior in developmental dyslexia
Albert M Galaburda, Joseph LoTurco, Franck Ramus, R Holly Fitch & Glenn D Rosen
All four genes thus far linked to developmental dyslexia
participate in brain development, and abnormalities in brain
development are increasingly reported in dyslexia. Comparable
abnormalities induced in young rodent brains cause auditory
and cognitive deficits, underscoring the potential relevance of
these brain changes to dyslexia. Our perspective on dyslexia is
that some of the brain changes cause phonological processing
abnormalities as well as auditory processing abnormalities;
the latter, we speculate, resolve in a proportion of individuals
during development, but contribute early on to the phonological
disorder in dyslexia. Thus, we propose a tentative pathway
between a genetic effect, developmental brain changes, and
perceptual and cognitive deficits associated with dyslexia.
One goal of cognitive neuroscience should be to establish transparent
pathways between genes and behavior and between genetic variants or
mutations and behavioral disorders. This effort starts with the discovery
of genes associated with a particular behavior, but must continue with
the characterization of all the downstream steps to that behavior, which
requires major, collaborative, cross-level research. Developmental dys-
lexia, a relatively common form of specific learning disability, is slowly
giving way to this type of discovery. We can now propose a pathway,
albeit still speculative and incomplete, between genetic variants (or
gene functions) and a complex developmental behavioral disorder, with
intervening abnormalities of brain development.
Following reports of acquired disorders of reading from injury affect-
ing the occipital and parietal lobes, researchers looked for congenital
anomalies involving those same posterior portions of the left hemi-
sphere to explain developmental dyslexia
1
. Subtle cortical neuronal
migration anomalies were observed in several post-mortem brains
2
.
Findings consistent with congenital brain malformations were reported
in a few cases of dyslexia and developmental language disorders. In
one autopsy case
3
, for instance, abnormal cortical folding affected the
parietal lobes, and neuron number was high in the subcortical white
matter. Magnetic resonance imaging identified 8 dyslexic cases with
the neuronal migration anomaly periventricular nodular heterotopia
4
,
and 13 children with perisylvian polymicrogyria (a form of abnormal
neuronal migration) and developmental language deficits
5
.
The most consistent findings in the original cases
2
are nests of
neurons, termed ectopias, in cortical layer 1 and occasional focal
microgyria affecting language areas. Additional defects in the thala-
mus and cerebellum
6,7
, together with the cortical changes, are pro-
posed to explain the dyslexic phonological deficits, as well as the
auditory discrimination and motor deficits that occur in some cases.
Animal models were developed to study possible causal relationships
between the brain abnormalities and behavior. However, the causes
of the malformations in the dyslexic brain remained unclear until
four candidate dyslexia susceptibility genes were reported—DYX1C1,
KIAA0319, DCDC2 and ROBO1—which are involved in neuronal
migration and other developmental processes. Experimental interfer-
ence with these genes leads to neuronal migration anomalies
8–14
.
Cognitive phenotype of dyslexia
The defining symptom of developmental dyslexia is a severe and spe-
cific difficulty in reading acquisition that is unexpected in relation to
other cognitive abilities and educational circumstances
15
. At the cog-
nitive level, there is widespread agreement that a large majority of dys-
lexic children suffer from what is commonly termed a “phonological
deficit,” that is, a deficit in some aspects of the mental representation
and processing of speech sounds
16
. Evidence for this phonologi-
cal deficit comes from three main behavioral symptoms: (1) poor
phonological awareness—the ability to consciously pay attention to
and mentally manipulate speech sounds, (2) poor verbal short-term
memory—the ability to temporarily maintain phonological represen-
tations active, and (3) slow lexical retrieval—the ability to retrieve the
phonological form of words for speech articulation
17,18
.
Other behavioral symptoms are sometimes associated with dyslexia,
including various types of auditory (prominently rapid auditory pro-
cessing), visual and motor deficits. It is likely that purely visual (but
not ocular) problems can explain reading disability in a minority of
dyslexic children, although the various theories of visual dyslexia still
need to be further specified and reconciled
19,20
. The associated audi-
tory and motor deficits are often proposed to be the underlying cause
of the phonological deficit
7,19,21
. However, their prevalence, at least in
older children and adults, is too low to explain the phonological defi-
cit in a straightforward way, and they are not specific to dyslexia
7,22
.
However, it could be argued that, as with most developmental disor-
ders, constellations of symptoms change with maturation, with some
symptoms remaining unchanged, others improving, and still others
Albert M. Galaburda and Glenn D. Rosen are in the Department of
Neurology, Division of Behavioral Neurology, Harvard Medical School,
Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston,
Massachusetts 02215, USA. Joseph LoTurco is in the Department of
Physiology and Neurobiology, University of Connecticut, Storrs, Connecticut
06269, USA. Franck Ramus is at the Laboratoire de Sciences Cognitives et
Psycholinguistique, École Normale Supérieure, 46 rue d’Ulm, 75230 Paris
Cedex 05, France. R. Holly Fitch is in the Department of Psychology, Behavioral
Neurosciences, University of Connecticut, Storrs, Connecticut 06269, USA.
e-mail: agalabur@bidmc.harvard.edu
Published online 26 September 2006; doi: 10.1038/nn1772
©
2
0
0
6

N
a
t
u
r
e

P
u
b
l
i
s
h
i
n
g

G
r
o
u
p


h
t
t
p
:
/
/
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
n
e
u
r
o
s
c
i
e
n
c
e

PERSPECTIVE
1214 VOLUME 9 | NUMBER 10 | OCTOBER 2006 NATURE NEUROSCIENCE
worsening, so that years later initial associations among symptoms,
including causal relationships, are no longer detectable. Therefore, the
possibility still exists that auditory deficits in the first year of life (or
earlier?), beginning before or at the time when phonological structures
are being established, may contribute to anomalies in phonological
development. However, this hypothesis is still speculative in the absence
of developmental studies beginning in infancy, focusing specifically
on this question. This effort will be helped along if early detection of
dyslexia susceptibility by genetic testing or by anatomical markers using
state-of-the-art in vivo imaging can be reliably implemented.
Dyslexia commonly coexists with specific language impairment
(SLI), developmental coordination disorder and dyscalculia, to name
a few
23–25
. This suggests at least partly shared etiological factors, for
which neuroimaging and genetic data should be highly informative.
For instance, it has been suggested
26
that the grey matter abnor-
malities observed in dyslexia partly overlap with those observed in
SLI. Common genetic factors would also be expected, although no
overlaps have as yet been detected.
Neural correlates support every observed cognitive deficit seen in
dyslexia. Both cytoarchitectonic and larger-scale gray- and white-
matter abnormalities in left perisylvian cortical areas involved in
phonological processing provide a sound basis for the phonological
deficit
2,21
. The additional abnormalities in the thalamus and cerebel-
lum provide equally sound bases for sensory and motor deficits
6,7
.
Animal modeling of these brain abnormalities (see below) indicates
that they interact during development.
Cellular and molecular mechanisms
Familial occurrence and twin studies have long suggested a genetic
basis for dyslexia. For over twenty years, it has been possible to pinpoint
chromosomal sites related to dyslexia susceptibility (for example, refs.
27–29). DYX1C1, KIAA0319, DCDC2 and ROBO1 are reported to be
dyslexia candidate susceptibility genes
8–14
. The proteins encoded by
these genes, though diverse, may be functionally linked, either directly
or by virtue of their similarity to other proteins, in pathways involved
in neuronal migration and axon growth. Developmental processes,
such as neuronal migration and axon growth, share several features
and requirements, including dependence on coordinated changes in
cell adhesion and cytoskeletal restructuring. ROBO1 has clear roles
in axon growth and neuronal migration, and the proteins in the DCX
family, of which DCDC2 is a member, have well-documented involve-
ment in neuronal migration to neocortex and may also be involved in
the development of the corpus callosum. In addition, KIAA0319 shares
extracellular motifs with proteins involved in cell adhesion.
DCDC2 is one of an eleven-member group of proteins distinguished
by the presence of tandem or single dcx domains. The first characterized
gene of this family, DCX, was identified after the discovery of muta-
tions in a gene that cause double cortex syndrome and lissencephaly
in humans
30,31
. Another member of the DCX family, Dclk, genetically
interacts with Dcx in mice, and two functioning copies of Dcx and Dclk
are necessary both for growth of axons across the corpus callosum and
for neuronal migration in cerebral cortex
32,33
. Dcdc2 has not yet been
tested directly for a possible role in axon growth across the callosum.
However, as with Dclk and Dcx, local loss of function induced by RNAi of
Dcdc2 results in the interruption of normal neuronal migration in neo-
cortex (Fig. 1)
12
. A comparison of the biochemical and cellular functions
of proteins in the DCX family
34
shows that DCDC2 shares functional
features with DCAMKL1 (also known as DCLK) and DCX.
KIAA0319 is the second of the two candidate dyslexia susceptibility
genes on chromosome 6p (refs. 10,11). The gene codes for an integral
membrane protein with a large extracellular domain, a single trans-
membrane domain and a small intracellular C-terminus. One pro-
tein—polycystin 1—with a similar extracellular domain, is involved in
adhesion between kidney cells. A role in cellular adhesion for KIAA0319,
or its homolog in rodents, has not yet been directly tested. However,
the phenotype in RNAi studies in developing brain indicates a role for
KIAA0319 in neuronal migration, perhaps by changing the relationship
between migrating neurons and radial glia
11
. The change in associa-
tion between radial glia and migrating neurons following interference
with the mouse homologue of Kiaa0319 (called D130043K22Rik) is
suggestive of a change in adhesion between migrating neurons and
radial glial fibers.
ROBO1 and DYX1C1 were both identified as candidate dyslexia sus-
ceptibility genes from chromosomal translocations in small pedigrees
or by linkage dysequilibrium association studies
9,13,35
. The functional
alleles that initially implicated DYX1C1 in dyslexia do not significantly
associate with reading disorders in other, larger populations, calling
into question the general relevance of these DYX1C1 alleles to dyslexia
(for example, refs. 36,37). The domain structure of DYX1C1 does not
clearly link it to cell adhesion or to cytoskeletal dynamics. However, in
vivo RNAi studies indicate that it is involved in neuronal migration
14,38
.
Soon after transfection, neurons expressing the interfering construct are
arrested in the intermediate zone, and by adult life, many of them have
migrated, albeit late, and have achieved abnormal laminar placement.
Additional neurons have migrated through breaches in the basal lamina
to produce ectopias in the molecular layer, which resembles the picture
seen in post-mortem dyslexic brains (Fig. 2).
The role of invertebrate and vertebrate homologues of ROBO1 in
neuronal development is well understood. Initially found in a genetic
screen for axon-patterning mutants in Drosophila, the roundabout
gene and vertebrate homologs are important in axon growth across
the commissures in the brain and spinal cord
39,40
. The ligands of ROBO
proteins—slits—have been implicated in both axon growth and cell
migration. For example, slit directs the migration of neurons to neocor-
tex by repelling cells away from the proliferative zones—the ventricular
zone and ganglionic eminence—of the developing forebrain
41
. Despite
this exciting beginning, we want to stress that the functional signifi-
cance of the specific genetic variants of the four genes associated with
Figure 1 Protein domains and possible functions. KIAA0319 and ROBO1
serve as transmembrane adhesion molecules and receptors that guide axons
to appropriate targets. DCDC2, and perhaps DYX1C1, are proposed to act
as downstream targets that then serve to modulate changes in cytoskeletal
dynamic processes involved in the motility of developing neurons. Critical
future studies must now address whether there are links between the
functions of these proteins in migration and axonal pathfinding.
©
2
0
0
6

N
a
t
u
r
e

P
u
b
l
i
s
h
i
n
g

G
r
o
u
p


h
t
t
p
:
/
/
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
n
e
u
r
o
s
c
i
e
n
c
e