Classic Studies

By GW5, the rostral forebrain or telencephalon begins as a single layer of proliferative neuroepithelial cells overlying the newly formed lateral ventricles [Kriegstein et al., 2006; Norman et al., 1995; Sidman and Rakic, 1982]. This primitive layer, known as the ventricular zone (VZ), consists of densely packed and morphologically homogeneous cells that are radially oriented and maintain contact with both the ventricular lumen and the pial surface of the brain. Neural progenitor cells or neuroblasts in the VZ have highly mobile nuclei that move up and down within the cytoplasm of the cell, undergoing division when the nuclei are located near the ventricular surface.

During the next stages of cortical development, several new layers are seen. A marginal layer forms superficial to the VZ as a cell-sparse zone comprised mostly of cytoplasmic processes of cells in the VZ by GW5. Soon thereafter, a new neuronal layer, known as the preplate, forms above the VZ, and then a second proliferative layer of loosely arranged cells, known as the subventricular zone (SVZ) forms between the VZ and preplate. Both the VZ and SVZ contain mitotic cells. As cortical neurogenesis proceeds, newly generated neurons migrate radially out of the proliferative zones by climbing up a scaffold formed by specialized radial glia whose cell bodies are located in the VZ (vzRG)-cells that have long processes that extend up to the cortical surface. The first and smaller wave of neurons is generated in GW8 and migrates into the preplate in GW9, splitting it into a superficial marginal zone and a deeper subplate zone. A second and larger wave begins generating neurons in GW12 and peaks in GW15-16; these cells begin to migrate outward in GW13 and trail off several weeks after the peak of neurogenesis.

During cortical development, the marginal zone contains pioneer neurons known as Cajal–Retzius cells, which express Reelin and other proteins that regulate neuronal migration, cortical lamination, and later cortical organization [D’arcangelo et al., 1995]. The upper part of the marginal zone persists to form layer one of the mature cortex. Subplate neurons participate in development of critical thalamocortical and corticocortical connections, and also express many regulatory proteins before disappearing late in gestation [Del Rio et al., 2000]. The cortical plate grows rapidly, with newly arriving neurons migrating past earlier-generated cells to form progressively more superficial layers. This process of “inside-out” migration ultimately forms a six-layered cortex in which a neuron’s final position is determined primarily by its birth date. The VZ becomes progressively smaller and is eventually replaced by a single layer of ependymal cells that line the lateral ventricles. The SVZ eventually disappears as well, except along the lateral wall of the lateral ventricles, where it persists and continues to provide olfactory and possibly other neurons into adulthood.

Recent Progress

Earlier studies of cortical development suggested that all neurons destined for the cortex are generated in the VZ of the neocortex and climb up radial glia cells to reach the cortical plate, while dividing cells seen in the SVZ were interpreted as the origin of glial cells. However, recent studies have uncovered a far more complex process and overturned several former theories.

Neurogenesis

First, vzRG were found to be primitive multipotential cells that serve as the primary neuronal founder cells, as well as classic radial glia cells [Heins et al., 2002; Malatesta et al., 2000; Miyata et al., 2001; Noctor et al., 2002]. During early stages of cortical development, radial founder cells undergo several sequential symmetric cell divisions to generate additional founder cells. Once some critical number of radial founder cells has been reached – a number that appears to differ between species – the process changes to a form of asymmetric division that generates one vzRG and one immature neuron. The immature neuron then migrates up the radial glia fiber to reach the cortical plate. The daughter cells generated from each radial founder cell are thought to generate a functional neuronal column or radial unit. The radial unit hypothesis [Rakic, 1995, 2000] proposes that the evolutionary expansion in cortical surface area in primates resulted from an increase in the number of founder cells prior to neurogenesis, with each radial unit generating a column of six layers of neurons. This would have the effect of increasing cortical and ventricular surface area without increasing cortical thickness.

Second, neuronal progenitor cells are found in the SVZ, as well as the VZ. The first type discovered were intermediate progenitor cells (IPCs), which divide symmetrically to produce immature neurons [Haubensak et al., 2004; Miyata et al., 2004; Noctor et al., 2004]. However, the outer portion of the SVZ is also populated with nonepithelial radial glia-like (oRG) cells that are able to self-renew and produce neurons, similar to classic vzRG [Hansen et al., 2010]. Both vzRG and oRG cells undergo asymmetric cell division to produce IPCs, which divide one or a few times to produce immature neurons. The two populations of outer SVZ progenitors – oRG and IPCs (both also called transit amplifying cells) – contribute to a massively expanded outer SVZ in humans and produce a majority of neurons destined for the human cortex [Hansen et al., 2010; Haubensak et al., 2004]. Further, most newborn neurons do not migrate directly to the cortex. Rather, they demonstrate several distinct phases of migration, including a phase of retrograde movement toward the ventricle before migration out to the cortical plate [Noctor et al., 2004].

Specification of neuronal cell type occurs very early in the process, which is regulated by expression of a series of proteins. For example, vzRG and probably oRG progenitors express the paired-class transcription factor Pax6, whereas IPCs express the T-domain transcription factor Tbr2 (also designated EOMES in humans), and immature postmitotic neurons, especially neurons of the preplate and early-generated neurons that reside in cortical layer six, express the related Tbr1 gene [Englund et al., 2005; Hevner et al., 2001].

Third, the cerebral cortex receives major contributions from both pallial (dorsal neocortical VZ) and subpallial sources. The latter population consists of neurons that originate in the VZ of the ventral telencephalon, especially the medial ganglionic eminence, and migrate along nonradial pathways through the SVZ and intermediate zone (IZ) to the neocortex, mostly along axonal tracts (see Figure 26-1). These ventral progenitors give rise to inhibitory interneurons that express the neurotransmitter gamma-aminobutyric acid (GABA) [Anderson et al., 1997; Lavdas et al., 1999; Letinic et al., 2002; Wichterle et al., 2001], and constitute 20–30 percent of cortical neurons. This class of neurons modulates cortical output [Cherubini and Conti, 2001; Krimer and Goldman-Rakic, 2001], regulates neuronal proliferation and migration later in corticogenesis [Owens and Kriegstein, 2002], and contributes to postnatal development of cortical circuits [Fagiolini and Hensch, 2000; Huang et al., 1999]. Several distinct subtypes of interneurons have been identified that differ by axonal and dendritic morphology [Lund and Lewis, 1993], chemical markers [Defelipe, 1993; Gonchar and Burkhalter, 1997], connectivity, and physiology [Cauli et al., 1997; Kawaguchi and Kubota, 1996]. The major subtypes of GABAergic interneurons identified by chemical markers include calbindin-, calretinin-, parvalbumin-, somatostatin-, neuropeptide Y- and nitric oxide synthase-positive cells.

Neuronal Migration

Neuroblasts are generated a long distance from their eventual position in the cerebral cortex, a distance that may represent more than a thousand cell body lengths in humans. They undergo a process of movement or migration from either the periventricular zone (glutaminergic neurons) or the ganglionic eminences (GABAergic neurons). Most radial migration occurs by attachment to and locomotion along radial glial fibers, which form an extensive radial lattice that guides radially migrating cells [Gupta et al., 2002; O’Rourke et al., 1992; Rakic, 1971, 1972, 1988]. Radial glia-independent modes of radial migration also occur, particularly by nuclear translocation during the earliest stages while the VZ and cortical plate remain close together [Book and Morest, 1990; Gupta et al., 2002; Morris et al., 1998; Pearlman et al., 1998]. But cells must also receive a polarity signal, adhere to the radial fiber or other extracellular proteins, extend leading processes, translocate their nucleus from the cell body into the leading process to re-establish the cell body in a new location, contract trailing processes, and receive a stop migration signal.

The peak time for neuronal migration is between GW12 and GW20, although migration may continue up to at least GW26 [Evrard et al., 1989; Sidman and Rakic, 1982]. The cortex is formed by an “inside-out” pattern, in which early-generated neurons form the initial cortical plate and later-generated neurons climb past them to progressively more superficial positions. Thus, the earliest neurons to migrate end up in layer six, and the last to migrate in layer two.

Cortical Organization

By approximately 22 weeks, distinctive layers first appear in the cortex. Further maturation consists of additional synaptogenesis, retraction of early axonal collaterals that did not establish appropriate connections, neurotransmitter biosynthesis, and other processes. Most of these processes continue well beyond the perinatal period. The marginal zone of the fetal cortex matures to become layer one, or the molecular layer of mature cortex. The pioneer Cajal–Retzius neurons disappear, leaving a zone of nerve fibers, dendrites, and synapses with only a few cells. The subplate is incorporated into the deep layers of the cerebral cortex. By GW27, all six layers of the mature cortex are visible [Norman et al., 1995]. The cortex also has a columnar or vertical organization that derives from radial (centrifugal) migration of the dominant cell types.

While neuronal identity was specified much earlier in development, migrating neurons only begin the process of differentiation into functional subtypes as their final positions are reached. The critical changes include the appearance of neuronal processes, development of specific neuronal cell membrane properties, and expression of proteins characteristic of a specific type of cell-to-cell transmission. Additional adaptations include further morphologic differentiation, and more complex molecular and physiologic differentiation and synaptic interconnections [Cowan, 1992]. Throughout much of this period, glial cells are dividing and differentiating as well. At least three major types of glial cells are known, including oligodendrocytes, type 1 astrocytes, and type 2 astrocytes. The interactions among these cell types are complex but important in their later role in neuronal migration.

Further Reading

This summary touches on only a few of the advances made over the past decade. Many reviews are available, such as a recent collection of papers in a symposium on “Patterning and Evolving the Vertebrate Forebrain” [Creuzet, 2009; Medina and Abellan, 2009; Merot et al., 2009; Moreno et al., 2009; Saghatelyan, 2009; Subramanian et al., 2009; Suzuki-Hirano and Shimogori, 2009].

Pathology

LIS is a diffuse brain malformation manifested by a smooth cerebral surface, abnormally thick cortex with four abnormal layers that includes a deep zone of diffuse neuronal heterotopia, and enlarged, dysplastic ventricles [Barkovich et al., 1991; Forman et al., 2005; Norman et al., 1995]. It encompasses the pathologic terms agyria (absent gyri) and pachygyria (broad gyri). SBH consists of symmetric and circumferential bands of gray matter located just beneath the cortex and separated from it by a thin band of white matter. This appearance led to the alternative – although incorrect – term, “double cortex,” for this malformation. The overlying cortex appears normal or mildly simplified as a result of abnormally shallow sulci [Barkovich, 2000; Barkovich et al., 1994, 1989b; Dobyns et al., 1996a].

Several different types of LIS have been recognized, based on pathological features. They are most readily distinguished based on the number of cortical layers, and include four-, three-, and two-layered forms [Forman et al., 2005]. In the common four-layered or classic form (LIS-4L), the gyral malformation can be most severe over either anterior or posterior brain regions. The cortex is 12–20 mm thick and composed of a normal marginal layer, a superficial cellular layer that corresponds to the cortical plate, a cell-sparse zone, and a deep cellular layer composed of heterotopic neurons [Forman et al., 2005]. The two major subtypes can also be distinguished by looking at the transition from the deep cellular layer to the underlying white matter. In the posterior predominant (LIS1 gene) form, the transition is gradual, with no striking features. In the anterior predominant (DCX gene) form, the deep cellular layer transitions to multiple small nodules of subcortical heterotopia, and then to white matter. A few patients with classic LIS have mild to moderate cerebellar hypoplasia.

SBH consists of a normal six-layered cortex, a thin zone of white matter underlying the cortex, and then a zone of dense heterotopic neurons that breaks up into nodules at their lower border, closely resembling the X-linked form of LIS described above [Forman et al., 2005]. LIS-4L and SBH comprise a single malformation spectrum, based on observations of rare patients with areas of LIS that merge into SBH, and of many families with LIS in affected males and SBH in affected females [Des Portes et al., 1997; Dobyns et al., 1996a; Pilz et al., 1999].

The rare two-layered form of lissencephaly (LIS-2L) has been seen only in severe LIS variants with complete agyria and severe brainstem and cerebellar hypoplasia [Forman et al., 2005; Lecourtois et al., 2010]. The molecular layer appears normal, and overlies a single thickened and disorganized cortex with randomly arranged and oriented neurons. The three-layered and two-layered forms of LIS have not been seen with SBH.

The rare three-layered form of LIS (LIS-3L), associated with X-linked lissencephaly with abnormal genitalia (XLAG), consists of a hypercellular marginal or molecular layer, a middle zone with a relative increase in pyramidal neurons, and a deep layer that is relatively thick and composed primarily of small and medium sized neurons [Forman et al., 2005]. Other features include an intermediate-thickness (8–12 mm) cortex, disorganization of the basal ganglia with cysts, gliotic and spongy white matter, and agenesis of the corpus callosum [Bonneau et al., 2002; Forman et al., 2005].

Several rare forms of LIS associated with severe congenital microcephaly have been described, currently designated as microlissencephaly (MLIS), based on birth head circumference smaller than −3 standard deviations at birth. This group may overlap with LIS-2L. The first syndrome with MLIS reported, now known as Norman–Roberts syndrome, consists of severe congenital microcephaly, lissencephaly with diffuse agyria, and several pathological features that differed from classic LIS [Dobyns et al., 1984; Norman et al., 1976]. These included a relatively thin deep cellular layer, no heterotopia of inferior olive or dentate nuclei, and grossly dysplastic cerebellar cortex with reduced granular cells. The authors have seen only one child possibly fitting this syndrome over many years, leading to some doubts regarding its existence. While we believe that this syndrome does exist, several published reports appear to describe different syndromes [Caksen et al., 2004; Iannetti et al., 1993]. In OMIM (number 257320), Norman–Roberts syndrome has been linked to mutations of the RELN gene. This entry is egregiously in error, as no pathological data are available from humans with mutations of RELN or Reelin pathway genes, and individuals with known RELN mutations have normal head size.

Another pathologically distinct three-layered form of LIS (LIS-3L) is associated with microcephalic osteodysplastic primordial dysplasia type 1 (MOPD1). The brain malformation consists of extreme microcephaly, foreshortened frontal lobes, frontal predominant lissencephaly with agyria or pachygyria, agenesis of the corpus callosum, and cerebellar vermis hypoplasia [Juric-Sekhar et al., 2010; Klinge et al., 2002; Meinecke and Passarge, 1991; Ozawa et al., 2005]. The brainstem and cerebellar involvement seems less severe than lissencephaly with cerebellar hypoplasia (LCH). The dysplastic cortex is characterized by extensive but thin glioneuronal heterotopia, a thin superficial layer that appears hyperconvoluted in places (thus somewhat resembling polymicrogyria), a thin second layer of disorganized medium to large neurons, and a thick third layer consisting of mostly small neurons [Juric-Sekhar et al., 2010]. The glioneuronal heterotopia, lack of white-matter cystic changes, the brainstem and cerebellar hypoplasia, and other features distinguish this from the XLAG-type of LIS-3L.

The most severe form of true lissencephaly may be the form reported as “familial lissencephaly with extreme neopallial hypoplasia” that we now designate as Barth-type microlissencephaly [Barth et al., 1982; Dobyns and Barkovich, 1999; Kroon et al., 1996]. The pathological features include extreme congenital microcephaly, severe LIS with diffuse agyria, disorganized cortex with four layers resembling classic LIS-4L, except for a relatively thin deep cellular layer, small collections of cells with scant cytoplasm (possible germinal cells) in the leptomeninges, absent olfactory nerves, severe optic nerve hypoplasia, small thalami resembling streaks, agenesis of the corpus callosum and other commissures, severe brainstem hypoplasia with absent cerebral peduncles, transverse pontine fibers, inferior olives and pyramids, and extreme cerebellar hypoplasia with absent folia.

Finally, use of the old terms “type 1” and “type 2” LIS is strongly discouraged in favor of the new descriptive terms used above. The old “type 1” was first used to refer to the four-layered form, but was later applied to anything that was not “type 2.” The old “type 2” form is now designated as cobblestone malformation, discussed in the next section. Only the most severely affected patients have agyria or pachygyria. A few reports have used the term “type 3” LIS to refer to a rare condition with severe fetal akinesia, microcephaly, and undersulcated brain surface [Allias et al., 2004; Encha Razavi et al., 1996]. The cortex in this form appears thin, rather than thick, and has evidence of neuronal cell loss, suggesting a prenatal-onset degenerative disorder that would best not be classified as a form of LIS. In addition, both severe congenital microcephaly and more severe forms of polymicrogyria are commonly misidentified as LIS.

Brain Imaging

Most of the distinguishing features seen on pathological examination can be viewed by brain imaging as well (Figure 26-2). The different types of LIS and SBH may be distinguished by both the pattern and the severity of the malformation. Recognition of these patterns has become essential for syndrome and molecular diagnosis, and for assessing prognosis and genetic risk [Dobyns et al., 1999b; Kato and Dobyns, 2003; Pilz et al., 1998b]. The different patterns of LIS and SBH associated with known syndromes and causal genes are summarized in Table 26-1.

Fig. 26-2 Subtypes of lissencephaly.

Midline sagittal (A–H) and high axial (A′–H′) magnetic resonance images of lissencephaly (LIS). associated with mutations of seven different genes. The most severely involved brain regions are marked by arrowheads on the axial images. A–A′, Classic LIS, with posterior more severe than anterior (p > a) gradient associated with an intragenic mutation of LIS1. B–B′, Severe classic LIS due to deletion 17p13.3, which results in loss of LIS1, YWHAE, and all of the intervening genes in a child with Miller–Dieker syndrome. C–C′, Classic LIS with a p > a gradient caused by an intragenic mutation of TUBA1A. D–D′, Moderate-severity lissencephaly with cerebellar hypoplasia (LCH), with complete agenesis of the corpus callosum, large dysplastic midbrain and tectum, and severe cerebellar hypoplasia associated with another intragenic mutation of TUBA1A. E–E′, Classic LIS with anterior more severe than posterior (a > p) gradient, caused by intragenic mutation of DCX. F–F′, Moderate LIS variant with p > a gradient due to mutation of ARX in a male patient; lower images (not shown) demonstrate that the temporal lobes are even more severely affected. G–G′, LCH with mild frontal pachygyria and severe cerebellar hypoplasia with absent folia due to homozygous mutation of RELN. H–H′, LCH with very mild frontal pachygyria and severe cerebellar hypoplasia with nearly absent folia due to homozygous deletion of VLDLR. I–I′, Diffuse SBH associated with heterozygous mutation of DCX. J–J′, Posterior predominant SBH associated with mutation of LIS1.

(A–H, Reprinted from Dobyns WB. The clinical patterns and molecular genetics of lissencephaly and subcortical band heterotopia. Epilepsia 51[Suppl. 1]:5–9, 2010, with permission from Wiley–Blackwell. I, Courtesy of Dr. William B Dobyns, University of Washington, Departments of Pediatrics and Neurology, Seattle, Washington. These images were selected from patients LR08-316 [A–A′], LR08-401 [B–B′], LR07-008 [C–C′], LR07-244 [D–D′], LR06-020 [E–E′], LR04-245 [F–F′], LP95-137a2 [G–G′], LR08-330 [H–H′], LP94-049 [I–I′], and LP94-051 [J–J′].)

Table 26-1 Lissencephaly Syndromes, Inheritance and Genes

ACC, agenesis of the corpus callosum; AD, autosomal-dominant; AR, autosomal-recessive; LCH, lissencephaly with cerebellar hypoplasia; MOPD, microcephalic osteodysplastic primordial dysplasia; SpAD, sporadic with presumed autosomal-dominant inheritance caused by new mutations; XL, X-linked.

For all forms of true LIS, the brain surface appears smooth with areas of absent gyri (agyria) and abnormally wide gyri (pachygyria), and abnormally thick cerebral cortex [Barkovich et al., 1991; Dobyns and Truwit, 1995]. In normal brains, most gyri are approximately 1–1.5 cm wide and the normal cortex 3–4 mm thick, but thicker in the primary motor cortex and thinner in the primary visual cortex. In all types of LIS, gyri are typically 3 cm wide or more, and the cortex 8–20 mm thick, although this varies with the type (see Figure 26-2A–G, with Figure 26-2H being an exception). Several distinct types of LIS are associated with agenesis of the corpus callosum, moderate to severe cerebellar hypoplasia, designated LCH, or both [Ross et al., 2001].

In SBH (see Figure 26-2I–J), the brain surface appears superficially normal, except that the sulci or crevices between gyri tend to be very shallow (less than 1 cm instead of 1–3 cm), and the cortex is normal and not thick [Barkovich et al., 1994; Dobyns et al., 1996a]. Just beneath the cortex, however, often separated from it by just a few millimeters of white matter, lies a smooth band of neurons that never reached the true cortex. The inner margin of the band is usually smooth, while the outer margin may appear smooth (with thick bands) or follow interdigitations of the overlying cortex (with thin bands).

The most common or classic form of LIS seen on brain imaging corresponds to LIS-4L, and occurs in a continuous series with SBH. The severity varies from complete or nearly complete agyria (grades 1 and 2, see Figure 26-2B); to mixed agyria-pachygyria (grade 3, see Figure 26-2A and C); and then to extensive pachygyria (grade 4, see Figure 26-2D and E); mixed pachygyria and SBH (grade 5, not shown); and finally, SBH alone (grade 6, see Figure 26-2I and J). Both LIS-4L and SBH may be seen with either a frontal or posterior predominant distribution. Thus, the gradient of LIS and SBH can be anterior more severe than posterior (a > p), posterior more severe than anterior (p > a), or anterior equal to posterior (a = p). In LIS-4L, the cortex is usually 12–20 mm thick. Common associated malformations include thick and rounded hippocampi, enlarged posterior portions of the lateral ventricles, and flat anterior portion of the corpus callosum.

Several other types of LCH may be differentiated, based on brain imaging. A few patients with classic LIS have moderate vermis-predominant cerebellar hypoplasia, which corresponds to the previous LCH group a [Ross et al., 2001]. The next consists of intermediate severity LIS with mixed agyria and pachygyria that appears most severe in central (posterior frontal and anterior parietal) regions (see Figure 26-2D). Subtle asymmetry is often seen, as well as hypogenesis or agenesis of the corpus callosum. Another more severe variant of LCH consists of diffuse agyria with a cortex that appears only moderately thick (not shown). The lower margin of the cortex may be smooth or follow an undulating course. Associated abnormalities include severe hypoplasia of the hippocampus, variable agenesis of the corpus callosum, and severe brainstem and cerebellar hypoplasia. This severe variant corresponds to LIS-2L, and combines the prior LCH groups c, d, and f [Forman et al., 2005; Kumar et al., 2010; Ross et al., 2001].

A less severe LCH variant, which was previously designated as LCH group b [Ross et al., 2001], consists of frontal-predominant lissencephaly consisting of mild to moderate pachygyria, mildly thick 8–10 mm cortex, very small globular hippocampi, mildly small brainstem, and diffuse severe hypoplasia of the cerebellum with absent or nearly absent folia (see Figure 26-2G and H). This pattern has been associated with mutations of Reelin pathway genes.

The LIS variant seen in children with the XLAG syndrome corresponds to LIS-3L [Bonneau et al., 2002; Forman et al., 2005]. It is characterized by mixed agyria and pachygyria that appears most severe over the temporal lobes, next most severe over the parietal and occipital lobes, and least severe over the frontal lobes; the cortex is only moderately thick, usually 8–10 mm (see Figure 26-2F). Associated abnormalities include agenesis of the corpus callosum that is most often complete, abnormal basal ganglia with indistinct borders and cysts, and diffuse abnormal (high T2, low T1) signal of white matter. The brainstem and cerebellum appear normal.

Several rare LIS variants are associated with severe congenital microcephaly. In Norman–Roberts syndrome, the LIS resembles classic LIS by imaging, except for the very small head size [WB Dobyns, unpublished data]. In MOPD1, the LIS appears as frontal-predominant agyria or pachygyria with agenesis of the corpus callosum and probably mild cerebellar vermis hypoplasia. In Barth microlissencephaly syndrome, the LIS is extremely severe with minimal gray–white differentiation seen, absent corpus callosum, pencil-thin brainstem, and extremely small afoliar cerebellum. These syndromes are reviewed in more detail below.

Clinical Features

Children with the most common types of LIS (or SBH) typically appear normal as newborns, although a few have a history of polyhydramnios, apnea, hypotonia, poor feeding, and mildly elevated newborn bilirubin levels that may reflect poor swallowing. With most types of LIS, seizures are uncommon during the first days of life. Most affected children come to medical attention during the first year of life due to:

In all affected children, the major medical problems encountered are feeding problems and gastroesophageal reflux, epilepsy of many different types, and recurrent aspiration and pneumonia due to the feeding problems and epilepsy.

A few children feed poorly from the first weeks of life, but this often improves unless they have one of the severe LIS variants. Most feed reasonably well for the first several years, except that many have difficulty during intercurrent illnesses. Feeding often worsens later, especially after about 3 years, with increased aspiration, decreased feeding tolerance, and recurrent pneumonia. These problems are frequently related to worsening epilepsy and gastroesophageal reflux, whether obviously symptomatic or not. They lead to placement of gastrostomy tubes and operations to reduce reflux (fundal plications) in many affected children, although the age varies widely within the first decade [Dobyns et al., 1992].

Prognosis

For children with the most common forms of classic LIS, the mortality rate is greater than 50 percent by 10 years, and few children live past 20 years. Children with severe LIS syndromes, especially Miller–Dieker syndrome, severe forms of LCH, XLAG, MOPD1, or Barth microlissencephaly syndrome have a more severe course and higher mortality rate. However, these data apply to children with severe forms of LIS, and specifically do not apply to children with rare partial forms of LIS, SBH, or Baraitser–Winter syndrome, or to the RELN-associated type of LCH, as all have better cognitive function and much longer survival.

In contrast, most patients with SBH and rare patients with partial forms of LIS have mild to moderate mental retardation, although both normal intelligence and severe mental retardation have been seen. Other clinical features are minimal pyramidal signs and dysarthria [Barkovich et al., 1994; D’Agostino et al., 2002; Dobyns et al., 1996a]. Seizures usually begin in childhood but may appear much later, and multiple types occur that may be difficult to control. Cognitive development may slow after onset of seizures. The frequency and severity vary greatly. EEG investigations usually demonstrate generalized spike-and-wave discharges or multifocal abnormalities [Battaglia, 1996; D’Agostino et al., 2002; Palmini et al., 1991]. Neurologic outcome usually correlates with the thickness of the subcortical band heterotopia and simplification of the gyral pattern, as seen on magnetic resonance imaging (MRI).

Syndromes, Genetics, and Molecular Basis

While older reports suggested the possibility of extrinsic (nongenetic) causes of LIS, such as prenatal cytomegalovirus exposure or perfusion failure [Dobyns et al., 1992; Norman et al., 1976], clinical experience suggests that all or almost all LIS is genetic. Mutations of the six known LIS genes, including ARX, DCX, LIS1, RELN, TUBA1A, and VLDLR, account for more than 80 percent of patients [Boycott et al., 2005; Gleeson et al., 1998; Hong et al., 2000; Keays et al., 2007; Kitamura et al., 2002; Reiner et al., 1993]. However, several of these are associated with one or more specific LIS or SBH syndromes with different presentations. The contributions of these six genes and the specific mutational mechanisms have been derived from separate studies over many years, making comparisons difficult. An update of the prior results in light of current knowledge is shown in Table 26-2 [Dobyns, 2010]. A review of the major LIS syndromes and their molecular mechanisms is presented in the following sections.

Genetic Counseling

All forms of LIS and SBH are genetic, and genetic testing is currently available for more than 80 percent of patients, although this varies with the phenotype (see Table 26-2) [Dobyns, 2010; Haverfield et al., 2009]. When LIS or SBH is suspected but the exact syndrome diagnosis is uncertain, the most productive order of testing begins with analysis for deletions of chromosome 17p13.3 that include the LIS1 gene, using chromosome microarray or other methods, followed by sequencing of LIS1, DCX, and TUBA1A, and sometimes ARX if the phenotype is compatible. However, details of the phenotype can be used by experienced clinicians to change the order of tests, such as testing DCX first in females with SBH, or ARX first for males with XLAG. Genetic testing for LIS and SBH is important, as several syndromes are associated with high risks of recurrence. This risk is especially great for parents who are carriers of a rearrangement of chromosome 17, and for mothers who carry ARX or DCX mutations, as both are X-linked. When the causal gene is unknown or unavailable for testing, empiric recurrence risks based on the known or most likely pattern of inheritance are appropriate. These are listed in Table 26-1.

Pathology

The brain malformations seen in WWS, MEB, or FCMD consist of poor gyral development, cerebral and cerebellar cortical dysplasia, hydrocephalus, brainstem hypoplasia with dysplasia of the inferior olives and dentate nuclei, and hypoplasia of the corticospinal tracts in brainstem and spinal cord [Bordarier et al., 1984; Dobyns et al., 1985; Towfighi et al., 1984; Williams et al., 1984]. Studies of affected human fetuses demonstrate a striking cerebral and cerebellar cortical dysplasia with discontinuities in the basal lamina at the pial surface (glia limitans) that allow abnormal migration of neurons and glia beyond the cortical plate and into the leptomeninges [Miller et al., 1991; Squier, 1993; Takada et al., 1987]. The same changes are seen in several animal models [Moore et al., 2002; Willer et al., 2004]. The cerebellar cortical cysts result from entrapment of islands of meningeal tissue between adjacent folia [Muntoni et al., 2009].

The postnatal cortical malformation consists of an undersulcated and subtly pebbled surface that includes areas resembling agyria, pachygyria, or polymicrogyria, although the histological appearance is very different. The pebbled surface appearance prompted the name “cobblestone cortex,” first proposed by Haltia [Dubowitz, 1996]. The cortex is thick and dysplastic, with no recognizable layers. The original glia limitans is found in the middle of the cortex, which is disrupted by abnormal vascular channels and fibroglial bands throughout that extend into and often obstruct the subarachnoid space. The cortical malformation is most severe with a smooth surface in WWS, and progressively less severe in the other syndromes in this probably continuous series [Dobyns et al., 1985, 1989; Dubowitz, 1994, 1996; Haltia et al., 1997; Takada et al., 1988; Walker, 1942]. The white matter is poorly myelinated, with numerous heterotopic neurons and microscopic cysts. The cerebellum is small and dysplastic, with cerebellar cortical cysts. Other brain abnormalities include obstructive hydrocephalus, brainstem hypoplasia with dysplasia of the inferior olives and dentate nuclei, and hypoplasia of the corticospinal tracts in brainstem and spinal cord. In a few patients, absent septum pellucidum, absent corpus callosum, or occipital cephaloceles may be noted. The brain malformation is frequently associated with eye malformations. Skeletal muscle biopsies show changes of CMD and hypoglycosylated α-dystroglycan [Brockington and Muntoni, 2005; Mercuri et al., 2006; Muntoni et al., 2002].

Only limited human pathological data are available for the cobblestone malformations that are not associated with CMD. Analysis of a Gpr56 knockout mouse shows discontinuities in the pial surface basal lamina and migration of neurons and glia into the leptomeninges, just as in the CMD-associated cobblestone malformations [Li et al., 2008]. The first neuropathology data in humans were recently reported in abstract form, and confirm the cobblestone-type cortical malformation with widespread glioneuronal heterotopia [Fallet-Bianco et al., 2010].

Brain Imaging

The brain imaging changes of CMD-associated cobblestone malformation present a continuous series of malformations that begins with the severe changes of WWS and ends with normal brain imaging. However, most patients cluster into one of the defined syndromes, which consist of WWS, an intermediate group classified as MEB, and progressively less severe forms classified as FCMD, CMD with cerebellar dysplasia and cysts or CMD with microcephaly, and CMD with normal brain imaging. The key imaging features are most easily recognized for WWS, and are collectively pathognomonic for this disorder [Clement et al., 2008; Dobyns et al., 1989; Jissendi-Tchofo et al., 2009; Mercuri et al., 2006; van der Knaap et al., 1997].

The imaging abnormalities seen in WWS begin with macrocephaly and prominent forehead, resulting from existing or prior hydrocephalus, and reduced size and partial obliteration of extra-axial spaces, which is especially prominent between the cerebral hemispheres (Figure 26-3C). The cerebral surface is undersulcated, usually with diffuse agyria. The cerebral cortex is moderately thick, usually about 7–10 mm unless thinned by hydrocephalus. The cortical–white matter border is jagged, with frequent vertical (perpendicular to the cortical–white matter border) striations, which differs from the chaotic striations typical of true polymicrogyria. Just beneath the cortex, streaks of laminar subcortical heterotopia are seen that differ from typical subcortical band heterotopia based on their beaded and discontinuous appearance. The white matter has very abnormal signal (bright on T2 and dark on T1 MRI sequences; dark on computed tomography [CT] scan) and may have small cysts. The white-matter volume may be normal, or thinned by hydrocephalus. The third and lateral ventricles are enlarged and may be very large and rounded, reflecting active hydrocephalus. Very rarely, the ventricles may be small. The corpus callosum is present, although frequently thin.

Fig. 26-3 Subtypes of cobblestone cortical malformations.

Midsagittal (left column), parasagittal (middle column), and axial (right column) magnetic resonance images from patients with four different cobblestone malformation syndromes. A–A″, The top row demonstrates brain abnormalities in Fukuyama congenital muscular dystrophy. These images show mild frontal pachygyria with 7- to 8-mm-thick cortex, normal to mildly simplified gyral pattern posteriorly, widely open sylvian fissures, especially on the left, moderately enlarged lateral ventricles, intact but stretched corpus callosum, normal brainstem, and moderate cerebellar hypoplasia and atrophy. These images are from a Japanese child with Fukuyama’s congenital muscular dystrophy (FCMD), who most likely is homozygous for the Japanese founder mutation consisting of addition of a 3-kb retrotransposon into the 3′ untranslated region of the FCMD gene [Kobayashi et al., 1998]. B–B″, The second row shows changes in muscle-eye-brain disease. These images show moderate frontal pachygyria with an 8- to 12-mm-thick cortex that extends into the parietal lobe, patches of abnormal white matter with high T2 signal intensity, hypoplastic brainstem, especially in the pons, moderate cerebellar hypoplasia, and a few small cysts in the high white matter. The patient was a boy with a homozygous mutation of POMGnT1 in exon 20: c.1813delC and H573fs (patient CC described by Yoshida and colleagues [Yoshida et al., 2001] and Taniguchi and associates [Taniguchi et al., 2003]). C–C″, The third row shows changes in severe Walker–Warburg syndrome. These images show diffuse agyria with an irregular surface, beaded subcortical heterotopia (black arrow in C″), diffuse abnormal signal of white matter with very high T2 signal intensity, very thin corpus callosum, and moderate to severely enlarged lateral ventricles. Posterior fossa abnormalities include severe hypoplasia of the brainstem, enlarged tectum (top white arrow in C), classic Walker–Warburg kinking at the junction of the pons and midbrain (bottom long white arrow in C), severe cerebellar hypoplasia (bottom short white arrow in C), small fluid collection in the posterior fossa, and overall small posterior fossa. No mutation has been identified, although only POMT1 has been tested. D–D″, The bottom row shows the changes in bilateral frontoparietal polymicrogyria, with extensive involvement of the frontal and parietal lobes, extended sylvian fissure, patches of very bright T2 signal of white matter (again, more severe frontally), mildly enlarged lateral ventricles, thin corpus callosum, and hypoplasia of the brainstem and cerebellum. The patient was a girl with a missense mutation of the GPR56 gene in exon 3: 112C3T and R38W [Piao et al., 2004].

(Courtesy of Dr. William B Dobyns, University of Washington, Departments of Pediatrics and Neurology, Seattle, Washington. These images were selected from patients LR01-058 [A], LP95-146 [B], LR00-181 [C], and LP93-017 [D]).

The brainstem and cerebellum in WWS (see Figure 26-3C) are remarkably dysplastic, with a kink at the midbrain–pons junction in which the midbrain is angled dorsally with respect to the pons [Jissendi-Tchofo et al., 2009]. The lower brainstem appears small, with moderate to severe hypoplasia of the medulla and pons, near-absence of the basis points, and ventral midline clefts of the ventral pons. But the midbrain and especially the tectum are abnormally large. The cerebellum is small, with the vermis more severely involved than the hemispheres, a dysplastic foliar pattern, and often small cysts within or near the cerebellar cortex. The posterior fossa may be enlarged and a few patients have small occipital meningoceles.

The imaging abnormalities in MEB are similar but consistently less severe than in WWS (see Figure 26-3B). The macrocephaly and hydrocephalus are common in MEB as well, while partial obliteration of extra-axial spaces is less extensive. The cerebral surface is again undersulcated, but with frontal-predominant pachygyria rather than agyria; posterior-predominant pachygyria occurs but appears to be rare. Some areas resembling polymicrogyria may be seen. The jagged cortical–white matter border and vertical striations are similar, while fewer streaks of subcortical heterotopia are seen. The white matter has very abnormal signal, similar to WWS in infants, but over time this evolves from diffuse to patchy to minimal signal changes. The frequency of hydrocephalus and thinning of the white matter and corpus callosum is probably similar.

The brainstem and cerebellum in MEB (see Figure 26-3B) are also dysplastic but less severely so than in WWS. The brainstem lacks the kink, but the lower brainstem, especially the pons, appears small while the midbrain and tectum are enlarged. The cerebellar changes are similar but, on average, less severe than in WWS, except that cerebellar cortical cysts may be more common. This may be due to a high frequency of cerebellar cysts in MEB patients with mutations of the POMGnT1 gene [Muntoni et al., 2009].

The imaging abnormalities in FCMD (see Figure 26-3A) are similar to MEB but less severe. Hydrocephalus is uncommon. The cortical malformation is less severe but still frontal-predominant, although some patients have severe temporal lobe agyria. The brainstem and cerebellum most often appear normal. In contrast, patients with mental retardation and cerebellar cysts have relatively normal forebrain structures and cerebral cortex, although review of some published images shows that the cortex is mildly dysplastic. The cerebellum mimics the appearance seen in MEB.

The brain imaging abnormalities in the cobblestone-like disorders without CMD (see Figure 26-3D) are similar to those in MEB, including the frontal-predominant pachygyria, moderately thick cortex with vertical striations, patchy white-matter abnormalities, hydrocephalus, and brainstem- and vermis-predominant cerebellar hypoplasia and dysplasia [Chang et al., 2003; Van Maldergem et al., 2008]. The brainstem is usually not as thin as is seen in MEB.

Clinical Features

In all of the cobblestone malformation syndromes associated with CMD, the phenotype consists of moderate to profound mental retardation, severe hypotonia, mild distal spasticity, and poor vision. WWS is the most severe, FCMD and mental retardation with cerebellar cysts the least severe, and MEB in the middle. This is in part a semantic issue, as MEB is the diagnosis used for any intermediate phenotype. Several less severe CMD-associated phenotypes without brain malformations have been associated with different mutations of the same genes. The other (non-CMD) cobblestone syndromes have brain phenotypes that resemble FCMD or MEB, but without the eye or muscle manifestations [Clement et al., 2008; Godfrey et al., 2007; Klein et al., 2008; Muntoni and Voit, 2004].

Prognosis and Management

Children with most of these syndromes have both cobblestone brain malformation and progressive muscle disease, which lead to combined hypotonia (from the brain and muscle disease) and spasticity (from the brain malformation). The combination makes orthopedic and rehabilitation management more difficult. Although the causative genes may be the same, it is important to distinguish among the various clinical syndromes, as the prognosis and thus management differ. WWS is a severe disorder associated with profound mental retardation and congenital hypotonia. Few children with WWS live past the age of 3 years. MEB and FCMD are less severe, with variable survival from later in the first decade into the teens and beyond; some patients have survived into the fifth decade. Some children with WWS reported to live longer than 3 years may be better classified as having MEB.

Optimal management begins with early recognition and accurate counseling regarding the prognosis, as some parents of children with WWS choose to limit life-sustaining interventions, such as ventilation during severe illnesses. Most children with WWS and some with MEB have congenital hydrocephalus that can only be managed with a shunt. Children with WWS who have occipital cephaloceles may not have hydrocephalus at birth, but excision of the cephaloceles is often followed by hydrocephalus. Thus, placement of a shunt should be considered at the time of cephalocele removal. Seizures occur but are rarely as severe as those seen in patients with LIS or SBH. In WWS and MEB, congenital glaucoma and buphthalmos require care by an ophthalmologist. Treatment of other eye anomalies that can interfere with vision, such as retinal nonattachment or detachment, cataracts, and corneal opacity, should be assessed on an individual basis in view of the overall poor prognosis. Children with FCMD have no or only minor eye anomalies.

The muscle disease progresses slowly, so that frequent evaluations to assess orthopedic and other rehabilitation needs are important. Careful seating is essential, owing to severe hypotonia, and contractures need to be managed with physical therapy and splinting when needed. In WWS, the CMD probably progresses slowly over time, but this is not apparent clinically, owing to the severe hypotonia and short survival period. Children with FCMD who learn to walk typically lose this skill several years later.

Syndromes, Genetics, and Molecular Basis

All of the cobblestone malformation syndromes with muscle involvement share the common feature of hypoglycosylated α-dystroglycan on skeletal muscle biopsy, and all of the known genes code for proteins known or suspected to be involved with glycosylation, particularly O-mannosylation of α-dystroglycan, which has led to the general term “dystroglycanopathy” for these and related CMD syndromes with normal brain structure [Brockington and Muntoni, 2005; Hewitt and Grewal, 2003; Muntoni et al., 2002; Ross, 2002]. Alpha- and β-dystroglycan are components of the dystrophin-associated glycoprotein complex that links the actin-associated cytoskeleton and extracellular matrix, and are encoded by the same precursor peptide. Alpha-dystroglycan is a highly glycosylated peripheral membrane protein that binds many extracellular matrix proteins through attached carbohydrate groups. In the dystroglycanopathies, especially the cobblestone-CMD group, glycosyl groups are absent or reduced, resulting in decreased binding of ligands such as laminin-2, agrin and perlecan in skeletal muscle, and neurexin in the brain [Barresi and Campbell, 2006].

Accordingly, all of the genes associated with cobblestone syndromes with muscle involvement code for proteins that regulate glycosylation, particularly O-mannosylation of α-dystroglycan. At least two of the non-CMD cobblestone syndromes are also congenital disorders of glycosylation [Cantagrel et al., 2010; Morava et al., 2009; Van Maldergem et al., 2008], and another is associated with a heavily glycosylated protein [Jin et al., 2007]. The cobblestone syndromes and associated genes are listed in Table 26-3, with estimated mutation frequencies shown in Table 26-4. Biochemical deficiency has been demonstrated in muscle for some of the proteins involved [Zhang et al., 2003].

Table 26-3 Cobblestone Syndromes, Inheritance and Genes

AR, autosomal-recessive; CDG, congenital disorder of glycosylation; CEDNIK, cerebral dysgenesis, neuropathy, ichthyosis and keratoderma; CHIME, colobomas, heart defects, ichthyosiform dermatosis, mental retardation, and either ear defects or epilesy; CMD, congenital muscular dystrophy.

* Cobblestone cortical malformation has not been confirmed for this syndrome.

Genetic Testing

All of the syndromes associated with the cobblestone malformation, with or without CMD, have autosomal-recessive inheritance. Six genes associated with cobblestone malformations and CMD, and 4–5 associated with cobblestone malformation with normal muscle studies have been identified (see Table 26-3). All have autosomal-recessive inheritance. Testing is available for most of the cobblestone-CMD syndromes and for GPR56. Testing serum transferrin isoforms for CDG will detect two (or more) other syndromes.

Antenatal Diagnosis

Several features of the cobblestone syndromes can be detected by prenatal testing. Almost all patients with WWS or MEB have enlarged lateral ventricles, and many have hydrocephalus. Almost all have cerebellar hypoplasia, and a few have gross microphthalmia. Each of these may be detected by prenatal ultrasounds in the mid- to late second trimester. Most fetuses with WWS should be detected, although no large series has been reported. The yield is likely to be lower with less severe syndromes. Molecular genetic testing is possible only when the specific mutation has been identified in the proband.

Periventricular Nodular Heterotopia

PNH consist of nodular masses of gray matter that line the ventricular walls and protrude into the lumen, resulting in an irregular outline (see Figure 26-4). They are recognized as a relatively common malformation that may occur as a single nodule or heterotopion, or as multiple contiguous or noncontiguous nodules or heterotopia [Dobyns et al., 1996a; Dubeau et al., 1995]. When the nodules are single or few in number, they may be associated with seizures or learning problems, but are unlikely to help with diagnosis. When the lesions are bilateral and numerous, a genetic basis is likely. When the nodules are diffuse and contiguous, they may be associated with other brain malformations, especially hypogenesis of the corpus callosum, cerebellar hypoplasia, or polymicrogyria [Parrini et al., 2006; Wieck et al., 2005].

Subcortical Nodular and other Types of Heterotopia

Although less common and less well recognized, other types of heterotopia are known. Subcortical nodular heterotopia consist of a large mass of nodules expanding a portion of one cerebral hemisphere (see Figure 26-4C and D). This lesion is rarely, if ever, bilateral and may be associated with ipsilateral PNH, agenesis of the corpus callosum, and cerebellar vermis hypoplasia or Dandy–Walker malformation. Overlying cortical dysplasia may be seen [Barkovich, 2000; Barkovich and Kjos, 1992a]. Despite the large size of the heterotopia, development may be normal or only mildly abnormal, and onset of seizures may not occur until the third decade of life. Some patients have overlying cortical dysplasia, which leads to more severe problems, including mental retardation. Transmantle heterotopia, also called transmantle dysplasia, consist of a column of gray matter without nodular features that extends from the ventricular wall upwards to the cortical surface, where it is surrounded by a typically small area of focal cortical dysplasia (see Figure 26-4E–F). Smooth or laminar heterotopia appear to be rare, and/or “ribbon-like”, undulating heterotopia even more rare, in the periventricular region. No examples of familial recurrence have been reported with these malformations.

Excessive white-matter neurons have been reported in patients undergoing epilepsy surgery and in postmortem studies. These abnormalities most often are seen in the temporal lobe and less often in the frontal lobes. Isolated heterotopic neurons in the white matter in normal brains also are most numerous in the temporal lobes [Emery et al., 1997; Rojiani et al., 1996]. The clinical correlates of the pathologic disorder are not well understood, although these neurons appear to be increased in patients undergoing temporal lobe resections for seizures [Hardiman et al., 1988]. MRI studies may reveal an abnormal white–gray matter junction if the number of ectopic cells is large. No familial recurrence or specific gender or racial predominance has been observed.

Pathology

Macroscopically, PMG appears as an irregular or pebbled cortical surface. The overfolding and microgyration are easier to see in sections cut perpendicular to the cortical surface [Harding and Copp, 2002], as shown in Figure 26-5. The distribution varies significantly from unilateral forms, to bilateral symmetric and asymmetric forms. The perisylvian cortex is the most frequently affected area and involved sylvian fissures may appear extended and orientated posteriorly, as shown in Figure 26-6C. The midline cortex (e.g., the cingulate gyrus) and hippocampus are usually spared [Harding and Copp, 2002]. The cortex often appears thickened to 8–12 mm, but when viewed microscopically, the cortex is overfolded and not truly thick. PMG may occur at the periphery of many porencephalic or hydranencephalic defects [Friede and Mikolasek, 1978]. While PMG most often occurs as an isolated malformation, it co-occurs with several other brain malformations, including microcephaly and megalencephaly, gray-matter heterotopia, ventriculomegaly, and abnormalities of the septum pellucidum, corpus callosum, brainstem, and cerebellum.

Fig. 26-5 Microscopic features of polymicrogyria.

Photomicrograph of a section of cortex of a patient with PMG stained with hematoxylin and eosin. The section is cut perpendicular to the cortex. Note the abrupt transition from the normal six-layered cortex on the left to the abnormal four-layered microgyric cortex on the right. The polymicrogyric area is characterized by overfolding with microgyri and microsulci. There is fusion of the marginal layer between adjacent gyri (1). The remaining layers are composed of two neuron-rich layers (2) and (4), separated by a cell-sparse layer (3).

(Image courtesy of Dr. Jeffrey Golden, University of Pennsylvania.)

Fig. 26-6 Subtypes of polymicrogyria.

Polymicrogyria (PMG) may be difficult to see in newborns. A–A′, A noncontrast axial CT scan (A) shows nonspecific cortical irregularity and possibly deep perisylvian and insular infolding (white arrows), while a T1-weighted axial magnetic resonance imaging (MRI) scan (A′) clearly shows bilateral perisylvian PMG. The MRI confirms the abnormal appearance of the insular region, which is lined by overfolded irregular gray matter with stippling of the gray–white junction, typical of PMG (black arrows). The entire insular and perisylvian is affected but the rest of the cortex appears normal. B–B′, Classic schizencephaly. Both coronal T1 (B) and axial T1 (B′)-weighted MRI images show full-thickness clefts lined by irregular gray matter (arrows). Panel B shows bilateral closed-lip schizencephaly, while panel B′ shows right open-lipped schizencephaly and absent septum pellucidum. C–C″, Bilateral perisylvian PMG. Right and left parasagittal T1-weighted MRIs (C and C″) show abnormal extension of both sylvian fissures, with PMG involving adjacent cortex. The coronal T1-weighted MRI (C′) shows that the sylvian fissures are lined by overfolded and irregular gray matter typical of PMG (arrows). The bottom two rows show several less common PMG variants. D, Right unilateral perisylvian PMG, with microgyri surrounding an extended left sylvian fissure (arrow). E, Bilateral frontal PMG with subtle irregularity of the cortex throughout the frontal lobes, but not extending beyond the central sulcus. F, Bilateral frontoparietal PMG, with PMG maximal in the frontal lobes and extending posteriorly well into the parietal lobes. G, Diffuse parasagittal PMG, a very rare pattern. H, Bilateral generalized PMG involving all of the brain equally. This scan also shows diffuse high signal throughout the white matter. I, Bilateral posterior (parieto-occipital) parasagittal PMG, with microgyri surrounding deep abnormal sulci extending forwards from the occipital poles (arrows). J, Bilateral perisylvian PMG (white arrows), with bilateral periventricular nodular heterotopia (black arrows). K, Periventricular nodular heterotopia adjacent to the trigones and posterior horns, with overlying PMG.

(Courtesy of Dr. Richard Leventer, Children’s Neuroscience Centre and Murdoch Children’s Research Institute, Royal Children’s Hospital, Melbourne, Australia; and Dr. William B Dobyns, University of Washington, Departments of Pediatrics and Neurology, Seattle, Washington.)

The clefts of SCH can be unilateral or bilateral, and “open-lipped” or “closed-lipped,” as shown in Figure 26-6B. In open-lipped clefts, the walls of the clefts do not appose each other. In closed-lipped clefts, the walls of the cleft are apposed and often fused, although a line of continuity between the lateral ventricle and subarachnoid space is usually visible; this is known as the “pia ependymal seam” [Yakovlev and Wadsworth, 1946a]. Other brain malformations may accompany SCH. Most are rare, except for absence of the septum pellucidum, which is present in approximately 70 percent of patients, raising the suggestion that SCH and septo-optic dysplasia may be related malformations in the spectrum of “septo-optic dysplasia plus” [Barkovich et al., 1989a; Kuban et al., 1989; Miller et al., 1998].

The extent of PMG visible microscopically may be greater than that suspected by macroscopic inspection, especially if present in the depths of sulci. PMG may show a variety of histological patterns, but all show abnormal cortical lamination, excessive folding, and fusion of adjacent gyri [Harding and Copp, 2002]. Two main forms of PMG are described: unlayered and layered, the latter described as “true” or “structured” PMG [Niewhuijse, 1913]. Controversy remains over which form is the more common [Harding and Copp, 2002]. Both forms may be found in the same patient, suggesting that they may be variations of the same malformation [McBride and Kemper, 1982].

The layered form of PMG shows four layers: a molecular layer and two neuronal layers, separated by an intermediate cell-sparse layer that contains myelinated fibres and few neurons [Harding and Copp, 2002; McBride and Kemper, 1982]. At the boundaries of layered PMG, an abrupt transition to normal cortex may be seen [Kuzniecky et al., 1993]. Layered PMG has been described in association with bilateral perisylvian PMG and in congenital cytomegalovirus infection [Crome and France, 1959; Kuzniecky et al., 1993; Marques Dias et al., 1984].

The unlayered form of PMG contains only a molecular layer and a neuronal layer without true lamination. Although no distinct layers are found, ultrastructural examination reveals different neuronal types positioned at appropriate cortical depths, as well as heterotopic nodules of pyramidal and nonpyramidal neurons in the deeper zones [Ferrer, 1984; Ferrer and Catala, 1991]. The microgyri are fused at the molecular layer, and the neurons often appear immature and contain little cytoplasm. Increased numbers of blood vessels and astrocytes are seen and are described as forming a “scar” [Ferrer and Catala, 1991]. The boundaries of unlayered PMG and normal cortex are not abrupt [Ferrer, 1984]. The unlayered form of PMG is associated with porencephalic cysts, Aicardi’s syndrome, and the clefts in SCH [Ferrer et al., 1986; Ohtsuki et al., 1981; Yakovlev and Wadsworth, 1946a, 1946b]. The cortical malformation in Zellweger’s syndrome is sometimes referred to as PMG, but is more complex than typical unlayered PMG and commonly involves abnormalities of the brainstem, cerebellum, and white matter, as well [Liu et al., 1976].

Brain Imaging

Using CT (see Figure 26-6A) and low-field-strength MRI, PMG is difficult to discern and may only appear as mildly thickened cortex [Barkovich et al., 1987; Byrd et al., 1989; Kuzniecky and Andermann, 1994]. It is for this reason that PMG is frequently misdiagnosed as pachygyria or lissencephaly. The only role for CT in the evaluation of PMG is to assess for evidence of intracerebral calcifications, which are seen in PMG and result from congenital cytomegalovirus infection or “band-like calcification with simplified gyration and polymicrogyria,” also known as “pseudoTORCH” syndrome [Briggs et al., 2008]. Using high-quality MRI at 1.5T or greater with appropriate age-specific protocols, it is now possible to differentiate PMG reliably from other MCDs, provided that the individual interpreting the MRI has knowledge of the imaging features of MCDs [Raybaud et al., 1996]. Sagittal imaging extending laterally to involve the sylvian fissures is of great value, as PMG often affects the opercular regions [Kuzniecky and Andermann, 1994]. In some circumstances, special techniques, such as inversion recovery sequences, volume averaging, focal coils, and curvilinear reformatting, may be required to identify very focal forms of PMG, although these are unusual.

Polymicrogyric cortex often appears mildly thickened (usually 6–10 mm) on imaging, due to cortical overfolding rather than true cortical thickening. This compares with lissencephaly in which the cortex is usually 10–20 mm thick [Thompson et al., 1997]. With thick slices, the cortex may appear mildly thickened, with an irregular or “stippled” gray–white junction; with better imaging (such as inversion recovery) using thin contiguous slices, microgyri and microsulci may be appreciated [Thompson et al., 1997]. In younger children with PMG, the cortex may not appear particularly thickened. This is thought to be due to the immature state of myelination in subcortical and intracortical fibres [Takanashi and Barkovich, 2003]. T2 signal within the cortex is usually normal, although delayed myelination or high T2 signal in the underlying white matter may be seen [Thompson et al., 1997]. Diffusely abnormal white matter signal should raise the question of an in utero infection, such as cytomegalovirus, or a peroxisomal disorder [Barkovich and Lindan, 1994; Barkovich and Peck, 1997; van der Knaap and Valk, 1991; van der Knaap et al., 2004]. The subarachnoid space may be enlarged over PMG, and contain excessive or anomalous venous drainage, especially in the sylvian regions [Thompson et al., 1997]. Other developmental anomalies may include gray-matter heterotopia, ventricular enlargement or dysmorphism, and abnormalities of the corpus callosum and cerebellum [Wieck et al., 2005].

CT or MRI scanning is usually sufficient to diagnose SCH and to determine whether the SCH is associated with an open or closed lip, although MRI is the modality of choice. The gray matter lining the cleft has the imaging appearance of PMG, with apparent mild cortical thickening, an irregular surface, and stippling of the gray–white interface. Subtle SCH may be recognizable by a “puckering” or “dimple” on the outer side of the lateral ventricle, at the point at which the cleft reaches the ventricular margin (seen in Figure 26-6B). By definition, SCH clefts are always lined by gray matter. SCH is frequently asymmetric, and the contralateral hemisphere should be closely evaluated for the presence of a milder SCH or PMG without a cleft. Agenesis of the septum pellucidum and hypoplasia of the optic nerves are common, present in as many as 30 percent of patients [Barkovich et al., 1987, 1989a].

Several types of white-matter abnormalities can be seen with PMG or PMG-like malformations. In typical PMG and SCH, prominent perivascular spaces are common. Extensive white-matter signal changes characteristic of gliosis lining a cleft suggest that the lesion is porencephaly rather than SCH. More extensive white-matter signal changes are seen in cobblestone cortical malformations.

PMG has been described in a number of topographic patterns. Most of these are bilateral and symmetrical, the most common of which is bilateral perisylvian PMG, although the perisylvian form may be asymmetric or unilateral. Other bilateral symmetric forms are generalized, bilateral frontal, and parasagittal parieto-occipital PMG [Barkovich et al., 1999; Guerrini et al., 2000, 1997]. PMG has also been described in association with PNH [Wieck et al., 2005]. The imaging features of these patterns of PMG are shown in Figure 26-6 and described in Table 26-6.

Table 26-6 Clinical and Imaging Features of the Polymicrogyria Syndromes

FLAIR, fluid-attenuated inversion-recovery; PMG, polymicrogyria.

Clinical Features

The clinical sequelae of PMG are highly variable, and depend on several factors. The most consistent predictors of a poor developmental outcome include microcephaly, especially severe microcephaly of −3 SD or smaller; abnormal neurologic examination, especially spasticity; widespread distribution of PMG, especially when bilateral and frontal; and additional brain malformations, such as heterotopia or cerebellar hypoplasia.

PMG is reported as an occasional component in many different conditions, including metabolic disorders, chromosome deletion syndromes, and multiple congenital anomaly syndromes. These patients may have a variety of clinical problems, other than those attributable to the PMG. Some patients with PMG or SCH have fewer clinical problems than would be expected for the location and extent of cortex involved (e.g., perisylvian PMG with deletion 22q11.2), while others have more severe problems (e.g., perisylvian PMG with deletion 1p36.3). PMG may involve eloquent cortical areas representing language or primary motor functions, yet these functions may occasionally be retained with minimal or no disability. This is especially true with lesions such as unilateral perisylvian PMG [Avellanet et al., 1996; Bisgard and Herning, 1993; Cho et al., 1999]. Also, functional imaging studies have often shown retained activity within polymicrogyric cortex [Araujo et al., 2006].

Etiology, Genetics, and Molecular Basis

Few topics in the field have generated as much discussion as the etiology and pathogenesis of PMG. The main points of discussion involve the timing (before or after neuronal migration has completed) and nature of the cause (developmental or destructive). The emerging impression is that several different etiologies contribute to the pathogenesis of PMG, which accordingly may not be a single malformation per se.

Initial theories of PMG suggested that it was the result of vascular defects, such as arterial ischemia or impaired venous drainage [Kundratt, 1882; Marburg and Casamajor, 1944]. The concept of PMG being secondary to a vascular defect has remained a prevalent theme in the literature, with multiple authors relating the observation that PMG often occurs in vascular territories, as in bilateral perisylvian PMG, in which the lesions are in middle cerebral artery territory, or at the peripheries of ischemic porencephalic defects, as evidence of a vascular etiology in the pathogenesis of PMG [De Leon, 1972; Evrard et al., 1989; Ferrer, 1984; Ferrer and Catala, 1991; Yakovlev and Wadsworth, 1946a, 1946b]. The observation of a layer of presumed laminar necrosis in layered PMG supports the theory that PMG is due to postmigratory perfusion failure or hypoxia, as may be seen in other examples of ischemic lesions [Levine et al., 1974; Richman et al., 1974]. In both layered and unlayered PMG, neurons are found in locations that could only be expected if migration had continued to near- or total completion, as appropriate neuronal cell types are often found at the appropriate depth within the cortex in both layered and unlayered PMG.

Numerous causes, both genetic and nongenetic, have since been reported in association with PMG. Nongenetic causes other than hypoxia or hypoperfusion relate mainly to congenital infections, primarily cytomegalovirus, with a few anecdotal reports of toxoplasmosis, syphilis, and varicella zoster virus [Barkovich and Lindan, 1994; Crome and France, 1959; De Leon, 1972; Evrard et al., 1989; Harding and Baumer, 1988]. The pathogenesis of PMG in these infections is uncertain, yet some have suggested that vascular compromise by inflammation or obliteration of the microvasculature may be involved [Gressens et al., 2003; Toti et al., 1998]. No convincing reports of PMG occurring secondary to toxin exposure, radiation, or medications have appeared, other than one report of maternal ergotamine use, in which the PMG was (anecdotally) attributed to placental vasoconstriction [Barkovich et al., 1995].

Numerous reports have described a genetic basis for PMG, based on association with structural chromosome abnormalities, familial recurrence or parental consanguinity, metabolic disorders, other known genetic diseases, or multiple congenital anomaly syndromes. Several chromosome disorders and X-linked forms consistently resemble typical PMG, while some of the other forms appear to have atypical neuropathological or brain imaging changes. Thus, single-gene inheritance of typical PMG appears to exist but is uncommon, while several “look-alikes” have autosomal-recessive inheritance.

PMG has been described with a growing number of structural chromosomal abnormalities, now including deletion 1p36.3, 4q21.21–q22.1, 6q26–q27, 21q2, and 22q11.2, and duplication 2p16.1–p23.1; the most common of these appears to be deletion 22q11.2 [Dobyns et al., 2008; Jansen and Andermann, 2005; Robin et al., 2006]. All types of single-gene inheritance have been reported. The most common seems to be X-linked perisylvian PMG with three or more genes involved, including the SRPX2 gene in Xq23 and two putative loci mapped to Xq27 and Xq28 [Borgatti et al., 1999; Guerreiro et al., 2000; Santos et al., 2008; Villard et al., 2002]. Several families with autosomal-dominant inheritance of unilateral perisylvian PMG have been reported [Caraballo et al., 2000; Chang et al., 2006; Guerreiro et al., 2000]. A rare form of autosomal-recessive bilateral posterior PMG has been reported and mapped in one family to 6q16–q22 [Ben Cheikh et al., 2009; Ferrie et al., 1995]. Other reports of autosomal-recessive PMG have appeared but seem to describe several rare heterogeneous forms [Ciardo et al., 2001; De Bleecker et al., 1990; Guerreiro et al., 2000; Hung and Wang, 2003; Sztriha et al., 1998; Sztriha and Nork, 2002]. Several syndromes and genes associated with PMG have been reported but, even collectively, account for a small proportion of patients. These are listed in Table 26-7.

Table 26-7 Polymicrogyria Syndromes, Inheritance and Genes

ACC, agenesis of the corpus callosum; AD, autosomal-dominant; AR, autosomal-recessive; PMG, polymicrogyria; SpAD, sporadic with presumed autosomal-dominant inheritance caused by new mutations; XL, X-linked.

(Data from numerous reports [Abdel-Salam et al., 2008; Abdollahi et al., 2009; Aligianis et al., 2005; Baala et al., 2007; Briggs et al., 2008; Brooks et al., 2005; Brooks et al., 1999; Cantagrel et al., 2007; Glaser et al., 1994; Graham et al., 2004; Hevner, 2005; Ho et al., 1984; Jaglin et al., 2009; Mitchell et al., 2003b; O’Driscoll et al., 2010; Roll et al., 2006; Shigematsu et al., 1985; Sisodiya et al., 2001; Solomon et al., 2009; Warburg et al., 1993; Yamaguchi and Honma, 2001].)

PMG has also been reported with several metabolic diseases with severe phenotypes, including Zellweger’s syndrome [Barkovich and Peck, 1997; Kaufmann et al., 1996; van der Knaap and Valk, 1991; Zellweger, 1987], neonatal adrenoleukodystrophy [Van Der Knaap and Valk, 1991; Wanders et al., 1995], fumaric aciduria [Kerrigan et al., 2000], mitochondrial diseases [Keng et al., 2003; Samson et al., 1994], glutaric aciduria type 2 [Bohm et al., 1982], maple syrup urine disease [Martin and Norman, 1967], and histidinemia [Corner et al., 1968]. However, the histopathology differs from classic PMG for several of them, especially the peroxisomal disorders (Zellweger’s syndrome and neonatal adrenoleukodystrophy) and glutaric aciduria type 2. Based on a mouse model, it is proposed that the cortical dysplasia in Zellweger’s syndrome is secondary to glutamate receptor dysfunction and defective N-methyl-d-aspartate (NMDA) glutamate receptor-mediated calcium mobilization during neuroblast migration [Gressens et al., 2000].

PMG has been attributed to abnormalities during late neuronal migration or early cortical organization, i.e., between 13 and 26 weeks’ gestation [Barkovich et al., 2005]. Until recently, there was little molecular evidence to support this assertion. Mutations in the TUBB2B gene have, however, recently been identified in four patients with asymmetric PMG with functional studies, suggesting that this gene is required for neuronal migration [Jaglin et al., 2009], although pathological data suggest atypical PMG.

The etiology of SCH has also been controversial; support for nongenetic causes is strong, while genetic forms appear to be very rare and atypical. In SCH, substantial evidence supports an association between PMG and congenital cytomegalovirus infection or in utero ischemic (vascular) injuries, the latter including twin–twin transfusion syndrome [Curry et al., 2005; Iannetti et al., 1998; Landrieu and Lacroix, 1994; Pati and Helmbrecht, 1994; Sener, 1998; Sherer and Salafia, 1996; Van Bogaert et al., 1996]. While very rare, several families with recurrence of SCH have been reported [Granata et al., 1997; Hilburger et al., 1993; Hosley et al., 1992; Tietjen et al., 2005]. A few reports of EMX2 mutations in SCH were never confirmed, and an association seems doubtful [Granata et al., 1997; Merello et al., 2008; Tietjen et al., 2005].

None the less, genetic causes are likely to account for a significant proportion of patients with PMG, although other causes, such as cytomegalovirus infection and ischemia, may account for significant proportion as well, especially for SCH. PMG undoubtedly has a heterogeneous etiology, with both environmental and genetic causes resulting in a similar phenotypic endpoint: a cortex with excessive folding and abnormal lamination.

Pathology

FCD shows a wide spectrum of severity in its gross morphology, topography, and microscopic features. The key feature in FCD is the presence of abnormal cortical lamination, although this may be subtle. With more severe forms of FCD, abnormal neuronal and neuroglial cell types are seen. At the mildest end of the spectrum is “microdysgenesis,” which is poorly defined. It usually refers to subtle developmental cortical abnormalities, including neuronal heterotopia, undulations of cortical layering, or neuronal clusters among cell-sparse areas [Hardiman et al., 1988]. Microdysgenesis has been found at autopsy more commonly in individuals with epilepsy compared to controls without epilepsy or other neurological disorders, as well as in surgical specimens from patients with medically intractable epilepsy [Hardiman et al., 1988; Meencke and Veith, 1992; Raymond et al., 1995]. Despite this, it is still unclear what degree of “microdysgenesis” may fall within the normal spectrum [Meencke and Veith, 1999].

Several classification systems for FCD have been proposed using neuropathological criteria. Most divide FCD according to both the degree of dysplasia (architectural or cytoarchitectural dysplasia) and the presence or absence of balloon cells, which correspond to FCD types 1 and 2, respectively [Colombo et al., 2003b; Kuzniecky and Andermann, 2003; Tassi et al., 2002]. A proposal to simplify the classification and terminology for FCD suggests that the term “microdysgenesis” be abandoned [Palmini et al., 2004]. The “Palmini classification system” shown in Table 26-8 proposes that subtle cortical abnormalities characterized by ectopic or heterotopic neurons of normal morphology be classified as “mild MCD.” These authors suggest that the term FCD only be applied to lesions with architectural abnormalities, such as dyslamination or the presence of abnormal cells within the cortex.

Table 26-8 Proposed Classification System for Focal Cortical Dysplasia

(Summarized from Palmini, et al: Terminology and classification of the cortical dysplasias, Neurology 62[6 Suppl 3]:S2–S8, 2004.)

But even this more severe end of the FCD spectrum presents difficulties. The pathological features of the tubers of tuberous sclerosis and FCD type IIB have significant overlap. In fact, the term “balloon cells” was first used in the neuropathology literature in the description of the tubers of tuberous sclerosis. Traditionally, however, the tubers of tuberous sclerosis and FCD with balloon cells have been regarded as separate lesions [Kuzniecky and Andermann, 2003].

The extent of FCD is highly variable, ranging from small focal areas involving part of a gyrus to a multilobar distribution in one hemisphere. On macroscopic examination, the cortex may appear normal. Subtle irregularities of the gyral pattern and mildly increased cortical thickness may be localized to the cortex at the depth of a sulcus, a malformation known as “bottom of the sulcus” dysplasia [Barkovich et al., 2005; Besson et al., 2008]. Beginning with these small lesions, a complete series of progressively more extensive forms of FCD has been described, including focal transmantle, sublobar, lobar, posterior quadrantic, and finally hemispheric dysplasias. These demonstrate more obvious abnormalities of cortical morphology, with overlap between hemispheric dysplasia and HMEG [Adamsbaum et al., 1998; Barkovich et al., 1997; Barkovich and Peacock, 1998; D’Agostino et al., 2004].

Given this similarity, the term hemimegalencephaly should perhaps be reserved for patients with significant malformations of noncortical structures, such as enlarged white-matter volume and enlarged, dysmorphic ventricles, although precise criteria to distinguish hemispheric FCD from HMEG have not been developed.

A spectrum of abnormalities may be seen on microscopic examination, including disruption of cortical lamination, neuronal heterotopia within the cortex or subcortical white matter, poor differentiation of neuronal and glial cells, and the presence of abnormal cells, such as balloon cells and large bizarre neurons, examples of which are shown in Figure 26-7A–D and as diagrams in Figure 26-7E [Colombo et al., 2003a, 2003b; Mischel et al., 1995; Prayson et al., 1999, 2002; Tassi et al., 2002; Taylor et al., 1971]. Areas of FCD have been shown to be intrinsically epileptogenic both in vivo, using corticography or depth electrodes, and in vitro, using cortex resected from patients with intractable epilepsy [Avoli et al., 2003; Mattia et al., 1995; Palmini et al., 1995; Tassi et al., 2002]. Single-cell studies have suggested that the dysmorphic or cytomegalic neurons, and not the balloon cells, contribute to epileptogenesis [Cepeda et al., 2005].

Fig. 26-7 Microscopic features and subtypes of focal cortical dysplasia.

Photomicrographs of cresyl violet (A, C, and D) and hematoxylin and eosin (B) stains of FCD specimens. A and B, Balloon cells (arrows) with characteristic ovoid shape, enlarged cell soma, and laterally displaced nuclei. C and D, Dysplastic neurons (thin arrows) have an enlarged cell soma but, in contrast to balloon cells, have a distinctly neuronal morphology with a single apical or basal process (thick arrow) extending from the cell body. Scale bar = 50 μm. E, The histologic changes in FCD are demonstrated in a diagram that shows a normally laminated cortex on the left, contrasting with the middle and right images showing increasing severity of cortical dysplasia. The most severe dysplasia is marked by dyslamination, malorientation of neurons, and the presence of large dysmorphic neurons and balloon cells. F and G, Coronal T2-weighted MRI scans from two unrelated patients with intractable childhood epilepsy show typical FCD. Panel F shows areas of irregular gyral formation with deep sulci and mild cortical thickening in the right frontal lobe (arrow). The rest of the right hemisphere also showed mild atrophy and hypoplasia. Histology showed mild dyslamination and the presence of giant neurons consistent with FCDIb. Panel G shows an area of gyral irregularity and increased subcortical signal in the right frontal lobe (arrow). Histology showed marked dyslamination and the presence of both dysmorphic neurons and balloon cells consistent with FCDIIB. The bottom row shows T2-weighted images of several less common forms of FCD, including bottom of the sulcus, transmantle, and posterior quadrantic cortical dysplasia. H, A deep sulcus in the left frontal lobe with cortical thickening at its base, consistent with “bottom of the sulcus” dysplasia. I, A “transmantle sign,” consisting of an area of cortical high signal and thickening, plus a line of high signal from the lesion to the lateral ventricle. This is designated focal transmantle dysplasia. J, Irregular and thickened cortex with mixed high and low signal in a wedge of tissue radiating out from a dysmorphic left frontal horn consistent with extensive transmantle dysplasia. K, A large area of thickened irregular cortex with high signal and blurring of the gray–white junction involving the left temporal, parietal, and occipital regions, consistent with posterior quadrantic dysplasia.

(A–D, reprinted from Lamparello: Developmental lineage of cell types in cortical dysplasia with balloom cells, P. Brain et al 2007;130:2267–2276; E, reprinted from Kuzniecky R, Epilepsia: Magnetic resonance imaging in developmental disorders of the cerebral cortex: 35 Suppl 6:S44–56, 1994. F–K, courtesy of Dr. Richard Leventer, Children’s Neuroscience Centre and Murdoch Children’s Research Institute, Royal Children’s Hospital, Melbourne, Australia.)

Brain Imaging

FCD is rarely visible by CT [Kuzniecky et al., 1995], and may not be visible even with high-quality MRI. Subtle abnormalities in gyration, cortical thickness, and the gray–white junction are best seen using thin-slice T1-weighted images, and may be a clue to underlying FCD [Yagishita et al., 1998]. Some forms of FCD may show increased signal on fluid-attenuated inversion-recovery (FLAIR) and T2-weighted images [Bronen et al., 1997; Colombo et al., 2003a, 2003b; Mackay et al., 2003]. White-matter signal may be abnormal in the region of FCD, producing intractable seizures [Eltze et al., 2005; Palmini et al., 2004], but it is not clear whether this represents dysplastic white matter, or a consequence of abnormal or advanced myelination secondary to frequent seizure activity [Mitchell et al., 2003a].

Mild MCD (microdysgenesis) and FCD type I may not be detectable on MRI. If abnormalities are seen, typical features consist of subtle cortical thickening and irregular sulcation or gyration. FCD type I may also be associated with lobar hypoplasia or atrophy and hippocampal sclerosis. The most striking imaging features of FCD are seen in FCD type II (see Figure 26-7F and G). The lesions in these patients typically show increased cortical thickness, blurring of the gray–white junction, abnormal sulcal and gyral patterns, and high signal at the base of the lesion and in the underlying white matter on T2 and FLAIR sequences.

Several specific named patterns of FCD have been described, based primarily on brain imaging features. FCD has also been shown to occur at the base of a sulcus, with cortical thickening and poor gray–white differentiation, often with a linear band of high signal from the base of the lesion to the lateral ventricle, known as the “transmantle sign,” shown for FCD types IB (see Figure 26-7H) and IIB (see Figure 26-7I). In focal transmantle dysplasia, a wedge of dysplastic tissue extends from the lateral ventricle up to the cortical surface (see Figure 26-7J). Histology shows features of FCD with balloon cells plus white-matter astrogliosis, and MRI shows a wedge of disorganized tissue with increased T2 signal [Barkovich et al., 1997]. Sublobar dysplasia is characterized by a deep infolding of the cortex with a thickened cortex and possible poor gray–white differentiation in the malformed region [Barkovich and Peacock, 1998]. Associated brain abnormalities include ventricular dysmorphism and callosal and cerebellar dysgenesis. Tissue has not been available for pathological examination. Another form of FCD affecting one posterior quadrant of the brain has been designated posterior quadrantic dysplasia (see Figure 26-7K)[D’Agostino et al., 2004]. This form of FCD is alternately known by the clumsy term “hemihemimegalencephaly.” These lesions have collectively been called “bottom of the sulcus” dysplasias [Barkovich et al., 2005].

Presurgical localization of these lesions often requires advanced MRI techniques and analysis, such as the use of surface coils, volume averaging, curvilinear reformatting, or 3T imaging. Functional studies, including single-photon emission computed tomography (SPECT) and FDG-PET scanning, are also often required to maximize the likelihood of identifying and defining the boundaries of FCD lesions. New MRI methods for lesion detection are being evaluated, including multichannel coils, high field strength (>3T), arterial spin labeling, susceptibility-weighted imaging, and diffusion tensor/spectrum imaging [Madan and Grant, 2009].

Clinical Features

Apart from tuberous sclerosis, no particular dysmorphic, neurocutaneous, or multiple congenital anomaly syndromes have been described in which FCD is a feature. The most common clinical sequelae of FCD are seizures. Developmental delay, cognitive disability, and focal neurological deficits are only observed with extensive dysplasias [Barkovich and Kjos, 1992b; Mackay et al., 2003; Wyllie et al., 1994]. Seizures from FCD may arise at any age from in utero until adulthood, although most patients present in childhood [Du Plessis et al., 1993; Wyllie et al., 1994]. Recent studies have shown consistent clinical differences between patients with types I and II FCD, respectively. Patients with type II FCD usually have extratemporal lesions, present at a younger age of onset, and have higher seizure frequencies [Fauser et al., 2006; Krsek et al., 2008, 2009b; Lerner et al., 2009]. Seizures may be simple partial, complex partial, or secondarily generalized, depending on the location of the FCD and age of the patient. Younger children may present with asymmetric infantile spasms. The seizure disorder may be intractable and life-threatening [Desbiens et al., 1993], so that surgical resection may be required. Much of the developmental delay and many of the cognitive disabilities associated with FCD may be due to the effects of repeated seizure activity. As complete as possible surgical resection of the FCD is consistently the most important variable for long-term seizure control in epilepsy secondary to FCD unresponsive to anticonvulsants [Kim et al., 2009; Krsek et al., 2009a; Lerner et al., 2009], and accordingly, surgery is being performed at increasingly younger ages in an attempt to protect the child against the deleterious effect of uncontrolled epilepsy and multiple medications [Guerrini et al., 2008; Harvey et al., 2008].

Etiology, Genetics, and Molecular Basis

Apart from FCD due to tuberous sclerosis, the etiology of FCD is largely unknown. No good evidence exists for environmental causes, and no genes or chromosomal loci have been identified for the common patterns of FCD. The highly focal and variable nature of FCD, especially FCD type II, and the close pathological resemblance to tuberous sclerosis led to the hypothesis that somatic mosaic mutations of genes in the mammalian target of rapamycin (mTOR) pathway, which includes the TSC1 and TSC2 genes that cause tuberous sclerosis, were involved [Crino, 2007]. In support of this hypothesis, loss of heterozygosity of the TSC1 gene was found in resected tissue of 11 of 24 patients with FCD, but no changes were found in TSC2 [Becker et al., 2002]. However, the same TSC2 mutation was found in nine family members with phenotypes varying from tuberous sclerosis to much milder conditions. Four had classic brain and skin abnormalities consistent with tuberous sclerosis, 3 had seizures and isolated cerebral abnormalities, and 2 had seizures and a normal MRI scan [O’Connor et al., 2003]. Evidence is thus accumulating that the phenotypic spectrum for patients with mutations in the tuberous sclerosis genes may be wide, and may include some patients not fulfilling current criteria for tuberous sclerosis, who thus have mild expression or a “forme fruste” of tuberous sclerosis.

Further data supporting this link come from human tissue studies of lesions resected from patients with both tuberous sclerosis and FCD. Morphological comparisons between abnormal cells found in FCD and normal cells involved in cortical development have shown that cytomegalic neurons have similarities to subplate cells, while balloon cells have similarities to radial glia. This led to a hypothesis that some forms of FCD may be the consequence of retained prenatal cells and neurons that establish abnormal connections and lead to seizures, a theory the authors termed the “dysmature cerebral developmental hypothesis” [Cepeda et al., 2006]. The mTOR pathway is involved in the control of protein synthesis, cell growth and proliferation, and synaptic plasticity. It has been found to be upregulated in surgical specimens from tuberous sclerosis patients, and its role in other forms of FCD and some developmental tumours is being investigated, including therapeutic trials designed to downregulate the mTOR pathway with the drug rapamycin [Wong, 2010].

One prevailing hypothesis states that balloon cells result from a defect of proliferation or differentiation of primitive neurons or glial cells, as is thought to be the case with those seen in tuberous sclerosis. For this reason, Barkovich and colleagues classify FCD with balloon cells as an MCD secondary to non-neoplastic abnormal cellular proliferation, and FCD without balloon cells as an MCD secondary to abnormal cortical organization [Barkovich et al., 2001a, 2005].

Two autosomal-recessive syndromes with relative or absolute macrocephaly and FCD have been identified in different Amish families. The cortical dysplasia–focal epilepsy syndrome (in which some affected children have pervasive developmental disorder or autism) and a severe developmental encephalopathy that resembles Pitt–Hopkins syndrome have been associated with mutations of CNTNAP2 [Strauss et al., 2006; Zweier et al., 2009]. Notably, heterozygous deletions of this gene have been associated with epilepsy and schizophrenia [Friedman et al., 2008; Mefford et al., 2010]. Another developmental syndrome, consisting of megalencephaly, severe psychomotor retardation, infancy-onset focal seizures, muscle hypoplasia, and distinctive craniofacial dysmorphism in Amish children, is caused by a homozygous mutation of LYK5 [Puffenberger et al., 2007]. Several affected children have died, and neuropathological study revealed megalencephaly, ventriculomegaly, cytomegaly, and extensive linear vacuolization and neuronal ectopia within white matter, accompanied by diffuse reactive astrocytosis.

Researchers in the field of human MCD have made assumptions as to the likely timing of the etiologies of different malformations, based on their appearance using both pathological and neuroimaging techniques. For example, heterotopic gray matter is assumed to result from disordered neuronal migration, as the heterotopic neurons appear to have arrested their migration to the cortex prematurely. Assumptions such as these were proposed well before a basic understanding of the genetic and molecular mechanisms of normal cortical development were established. In many instances, these assumptions have proved correct, yet it is now appreciated that many cortical malformations are likely secondary to abnormalities occurring at stages other than that of neuroblast migration. Understanding of the genetic and molecular basis of cortical development is advancing rapidly, with new genes or new roles for known genes being discovered. A more complete understanding of the key genes involved in human cortical development will be required in order to understand the possible timing and molecular basis of many malformations of cortical development.