Abstract

In epilepsy patients with cortical dysplasia (CD), this study determined the probable ontogenetic timing of pathogenesis based on the number, location and appearance of neurons. Magnetic resonance imaging (MRI) determined gray and white matter volumes of affected and non-affected cerebral hemispheres, and gray and white matter neuronal-nuclear protein (NeuN) densities and sizes were assessed in epilepsy surgery patients (0.2–38 years) with CD (n = 25) and non-CD etiologies (n = 14), and compared with autopsy cases (n = 13; 0–33 years). Pathology group, seizure type and age at surgery were compared against MRI and NeuN data. CD patients demonstrated increased MRI cerebral (3%) and gray matter (8%) volumes of the affected compared with non-affected cerebral hemisphere, and increased layer 1 (131%), upper cortical (9–23%) and white matter (28–77%) NeuN densities compared with autopsy cases. Non-CD cases showed decreased cerebral volumes of the affected hemisphere (14–18%) without changes in NeuN densities. Compared with autopsy cases, in CD and non-CD patients, cortical neurons were hypertrophied. Patients with a history of infantile spasms had a 40% increase in the size of layer 1 neurons compared with cases without spasms. By age, regardless of pathology group, there were logarithmic increases in MRI cerebral and white matter volumes, logarithmic increases in the size of lower gray and superficial white matter neurons, and logarithmic decreases in gray and white matter neuronal densities. These results support the concept that there were more neurons than expected in layer 1, gray, and white matter of CD patients compared with non-CD and autopsy cases. In addition, the location and appearance of neurons are consistent with the hypothesis that CD is the consequence of abnormalities occurring late in corticoneurogenesis that involve excessive neurogenesis with retention of pre-plate cells in the molecular layer and subplate regions.

Introduction

Cortical dysplasia (CD) was recognized as a pathologic substrate associated with epilepsy in 1971 (Taylor et al., 1971). Epilepsy from CD was originally thought to be uncommon, but with modern neuroimaging, CD has become increasingly recognized as a frequent substrate for seizures, especially in children (Mathern et al., 1999; Cook et al., 2004). While much has been learned over the past decade about the clinical, electrographic, magnetic resonance imaging (MRI) characteristics and neuronal properties associated with CD in epilepsy patients, less is known about the ontogenetic mechanisms that result in CD, although it is widely presumed that CD pathogenesis involves cortical maldevelopment (Mischel et al., 1995; Raymond et al., 1995; Battaglia et al., 1996; Kerfoot et al., 1999; Barkovich et al., 2001; Kuzniecky and Barkovich, 2001; Tassi et al., 2002; Cepeda et al., 2003; Mackay et al., 2003; Najm et al., 2004).

Neocortical development involves more prolonged and complex developmental processes in primates and humans compared with other mammals, and understanding the progression of corticoneurogenesis should provide clues to the timing of CD pathogenesis (Kostovic and Rakic, 1980, 1990; Marin-Padilla and Marin-Padilla, 1982; Rakic, 1988; Marin-Padilla, 1999; Zecevic and Rakic, 2001; Rakic and Zecevic, 2003). Telencephalon development begins with progenitor cell proliferation and symmetric cell divisions in the subventricular germinal zones. With successive divisions, some post-mitotic cells differentiate into pioneer neurons and migrate along radial glia, another self-renewing progenitor, to the cortical surface. The earliest migrating cells form the preplate, which is split by subsequent generations of migrating neurons into the marginal zone (future molecular layer) and subplate. The marginal and subplate zones are physically larger with more diverse cell types in primates and humans than rodents. In human prefrontal cortex, for example, the subplate is up to five times thicker (10–13 mm) than the cortical plate at 26–29 weeks gestation, and composed of multiple cell types, including large multipolar and polymorphous neurons (Kostovic and Rakic, 1990). The pyramidal and other neurons that form the eventual cortical plate migrate in successive waves, and form the gray matter in an inside-out gradient (i.e. earliest neurons in lower gray matter). Many cells of the preplate, including Cajal-Retzius, radial glia and subplate cells, eventually degenerate, and this process coincides with secondary gyral folding during prenatal cortical development (after 30 weeks gestation). Human cortical plate development continues postnatally, especially in the first months and years of life, with elongation of the cortical surface associated with a decline in gray and white matter neuronal densities, pyramidal maturation and synaptogenesis, and expansion of the white matter volume with myelination.

Based on an understanding of normal human prenatal corticoneurogenesis, the timing and possible mechanism(s) responsible for CD and epilepsy should be deducible from the neuronal abnormalities observed in postnatal surgical specimens (Rakic, 1988). For example, if the pathogenesis of CD involves arrest or severe abnormalities of early corticogenesis, before or during preplate formation, then the resulting neocortex at surgery should be severely malformed and microcephalic and should be similar to the four-layer cortex observed in children with lissencephaly (Uher and Golden, 2000; Volpe, 2000; Miyata et al., 2004). If CD involves abnormalities that obstruct neuronal migration during mid-corticoneurogenesis, then the resulting cortex should be smaller and should demonstrate numerous ectopic or heterotopic clusters of neurons in the white matter with fewer cells in the cortical plate. This would be similar to the pattern seen in the methylazoxymethanol (MAM) or radiation rodent models of CD, and fetal irradiation victims (Okajima et al., 1978; Roper, 1998; Colacitti et al., 1999; Baraban et al., 2000). If CD involves processes that occur later in corticoneurogenesis with overproduction of late generated gray matter neurons and lack of preplate cell degeneration, then the cortex at surgery should display more cells than expected in the upper gray, white matter and molecular layer with some cells demonstrating morphologic features consistent with Cajal-Retzius, radial glia and human subplate cells.

This study was designed to determine if the likely mechanisms responsible for CD in patients undergoing epilepsy surgery involved early, middle or late components of corticoneurogenesis based on the number, location and appearance of neuronal and cortical abnormalities as assessed by volumetric MRI and NeuN neuronal density measurements. Based on qualitative histopathological review of CD tissue, we hypothesized that patients with CD would show pathologic abnormalities supporting the view that it resulted from disturbances involving late corticoneurogenesis with over production of neurons and postnatal preservation of preplate cells.

Materials and Methods

Pre-surgery Evaluation

Patients with medically intractable seizures localized to part or all of a cerebral hemisphere were evaluated by the University of California, Los Angeles (UCLA) Pediatric Epilepsy Surgery Program, and the clinical protocols have been previously published (Mathern et al., 1999). Informed consent was obtained to use clinical data for research studies. In brief, the standardized pre-surgery evaluation included detailed history and neurologic examinations, interictal and ictal scalp EEG, and, when appropriate, intracarotid amobarbital injections (Wada test) and/or functional MRI for evaluation of memory and/or speech representation. Neuroimaging studies included high-resolution MRI and 18fluoro-2-deoxyglucose (FDG) positron emission tomography (PET). The epileptogenic region for surgical resection was anatomically defined based on convergent EEG and neuroimaging abnormalities. At surgery, electrocorticography further defined the brain regions to be removed based on background slowing, location of interictal spikes and polyspikes, and ictal discharges, as previously described (Mathern et al., 2000; Cepeda et al., 2003).

Pathologic Classification

Surgery patients were classified into categories based on histopathology of the resected specimen, and pre-surgery neuroimaging (MRI and FDG-PET) (Farrell et al., 1992; Vinters et al., 1992; Duong et al., 1994; Mathern et al., 1999).

Cortical Dysplasia (CD; n = 25)

These patients had severe cortical dysplasia on histopathology (defined as cortical dyslamination, heterotopia, excessive ectopic neurons, cytomegalic neurons, balloon cells, etc.) as previously described (Mischel et al., 1995). Neuroimaging showed hemimegalencephaly, pachygyria, large subcortical heterotopia or focal dysplasia. We excluded patients with polymicrogyria because these cases often do not contain cytomegalic neurons or balloon cells, and clinical studies support the notion that there may be different pathogenic mechanisms involved than the type of CD originally described by Taylor (Taylor et al., 1971; Barkovich et al., 2001). In addition, we excluded acquired dysplasia from perinatal strokes because again the mechanism of CD generation was likely to be different (Marin-Padilla, 1999).

Non-cortical Dysplasia (Non-CD; n = 14)

These cases consisted of patients with mostly static or progressive destructive cerebral pathologies, such as Rasmussen encephalitis (n = 6), temporal lobe epilepsy with or without tumors (n = 5), or evidence of old cerebral ischemia/infarction (n = 3). Neuroimaging often showed loss of brain structure, such as cortical atrophy, loss of white matter, etc.

MRI Assessed Cerebral, White and Gray Matter Volumes

The UCLA pre-surgery MRI for surgical cases was performed on a General Electric 1.5 T Signa scanner (Milwaukee, WI). Sequences included high-resolution coronal T1-weighted spoiled gradient echo pulse sequences (SPGR; TR = 13 ms, TE = 2.8ms, TI = 300 ms, flip angle = 25°, FOV = 24 cm, 1.5 mm coronal thickness slices, 78–124 slices per patient, matrix 256 × 256, NEX = 1) and coronal T2-weighted images (TR = 2000 ms, TE = 120 ms, FOV = 24 cm, 5 mm thickness slices, matrix 192 × 192, NEX = 2). Volumetric analysis of the MRI scans was performed with custom-designed commercial software (Silhouette; CEDARA, Ontario, Canada; www.cedera.com). Coronal T1 SPGR images were used for most of the analyses, and the operator was blinded to the pathology groups. The images were transferred into the Silhouette program, and each coronal section was segmented into CSF and brain parenchyma to obtain the total volume of each cerebral hemisphere. The volumes from each MRI coronal section were summated, and the cerebellum and brain stem volumes removed to obtain cerebral hemispheric volumes. The MRI images were then segmented according to the T1-weighted signal intensity difference into gray and white matter regions, and the borders were defined semi-automatically or manually. Volume measurements were calculated from the number of voxels included in the defined white matter structures. The volume containing white matter was subtracted from total cerebral hemisphere volume to obtain gray matter and basal ganglia volumes. In surgical cases <2 years of age, which often have poor delineation of the T1-weighted gray–white matter signal differentiation, coronal T2-weighted images were used to define the gray/white matter border. In these younger cases, the border was defined manually in each slice, and volumes calculated in the same manner as the T1-weighted images for older patients. For data analysis, each hemisphere was defined as the one affected by the CD or non-CD pathological process (i.e. side of surgery) or the non-affected non-operated side.

Tissue Selection and Classification

At surgery, one or more 1.5–2 cm blocks of neocortex and adjoining white matter involving the crown of a gyrus were microsurgically removed and the blocks were immersion fixed for immunocytochemistry (ICC). The remaining brain tissue from the surgical resection was processed for routine histopathology. By design, the ICC tissue blocks from the CD cases were selected from the regions closest to the worse abnormalities identified by FDG-PET and MRI to maximize sampling of the most dysplastic cortex. The sections from non-CD cases came from areas with the least cortical abnormalities by neuroimaging but part of the planned resection in order to sample the most normal tissue to compare with non-seizure autopsy cases. Neocortical blocks from 13 autopsy cases of similar ages without known neurological disease were also collected for comparison with the CD and non-CD cases. Death in the autopsy group was from acute cardiac, septic, or traumatic causes, and brain tissue was collected between 3 and 11 h after death (mean ± SD = 6.6 ± 2.3 h). The location of the sample blocks by gyrus was identified and recorded for both the surgical and autopsy cases.

NeuN Immunocytochemistry Processing

Neuronal-nuclear protein (NeuN) was chosen over traditional Nissl stains because of the specificity of this antibody in identifying differentiated neurons, and the fixation protocols were regimented so that autopsy and surgical tissue were processed in a similar fashion. Surgical and autopsy ICC tissue blocks were immediately immersion fixed in freshly prepared phosphate-buffered 4% paraformaldehyde for 24–48 h, and then cryoprotected overnight in 20% buffered sucrose and stored at −80°C. Cryostat-cut sections (30μm) were collected from the cryostat and placed in individual 3 ml wells containing 0.05 mol/l Tris–HCl-buffered saline (TBS; pH 7.4). The free-floating sections were processed the same day as follows, with 10 min TBS rinses (three changes) between each step: 5 min in 3% hydrogen peroxide/10% methanol in TBS; 60 min in a blocking solution of 2% normal horse serum in TBS; overnight in primary antisera against NeuN (mouse anti-neuronal nuclei, Chemicon International, Temecula, CA, catalog no. MAB377, 1:2000 dilution) diluted in 2% normal blocking serum; 60 min in diluted biotinylated anti-mouse antibody (ABC Kit, Vector Laboratories, Burlingame, CA); and 30 min in a solution of excess avidin and biotinylated horseradish peroxidase (ABC Kit, Vector Laboratories). The sections were developed for 7–8 min in 0.5 mg/ml 3,3′-diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxide. After sufficient coloring, the reaction was halted by washing in several rinses of cold PBS, the sections were mounted on subbed slides, air dried, treated for 35 s in 0.1% osmium tetroxide in 0.1 mol/l phosphate buffer (pH 7.4), dehydrated and coverslipped (Mathern et al., 1995, 1997).

NeuN Defined Neuronal Densities

Nine regions per tissue section were selected in a standardized manner for NeuN cell counts (Fig. 1). Because of the neocortical dyslamination associated with CD (a definition of CD), we selected cell density sample sites based on pre-determined distances from the pial surface or bottom of the cortical ribbon instead of identified neocortical cell layers. An ocular grid consisting of 10 × 10 boxes was positioned over the tissue section with the pial surface at the top. For layer 1, a 10 × 10 box at ×40 magnification (31 × 31 μm) was positioned with the superior line parallel to the pial surface, and all NeuN-labeled cells within the box were counted except those touching the upper and right borders of the grid. The neocortical gray matter sample sites were labeled levels 1 (superior) to 6 (inferior), and their location determined by measuring the distance from the bottom of layer 1 to the junction of the neocortex and white matter and dividing by six. A 5 × 5 box at ×40 magnification (15.2 × 15.2 μm) was positioned at each location, and NeuN-positive cells within the box counted (see Fig. 1). In the gray matter, the distance between each 5 × 5 box varied by 18.6 to 21.7 μm from case to case. For the NeuN cells in the superficial white matter, a 3 × 10 box at ×10 magnification (37.2 × 124 μm) was positioned just below the neocortex–white matter junction, and cells were counted. At a distance of 24.8 μm below the bottom of that box, another 5 × 10 box (62 × 124 μm) was positioned, and cell counts performed in the deep white matter. Cell counts from each area were calculated as the number of NeuN cells per 10 000 μm2.

Figure 1.

NeuN-stained (30 μm thick) section of cortex and underlying white matter from an autopsy case illustrating the sampling sites for neuronal density measurements. The outlined boxes and corresponding labels to the left indicate the nine regions sampled (see text).

Figure 1.

NeuN-stained (30 μm thick) section of cortex and underlying white matter from an autopsy case illustrating the sampling sites for neuronal density measurements. The outlined boxes and corresponding labels to the left indicate the nine regions sampled (see text).

It must be emphasized that NeuN-labeled cell densities, as used in this study, are estimates of the number of neurons per unit area (i.e. packing density) and not an absolute calculation of the total number of neurons per hemisphere. Experimental techniques used to determine absolute neuronal quantities within a hemisphere or brain region of autopsy cases were not practical in this surgical study because the entire hemisphere or area of pathology was not available for sampling. Likewise, it is nearly impossible to correct for tissue volume changes that occur from fixation shrinkage, although it can be assumed that it should be the same for all pathology groups including autopsy cases with our protocol. However, neuron densities, as used in this study, are reliable relative estimates of packing densities, and statistical differences between groups of patients that are similarly processed and counted can be accurately determined (Mathern et al., 1995, 1997).

Cortical Thickness

To assess the average thickness of the neocortex, an image computer was used as previously described (Mathern et al., 1997). The same NeuN-stained sections used for neuronal counts were imaged using a video charge-coupled device camera (SPOT RT CCD; v3.2; Diagnostic Instruments, Inc.; Sterling Heights, MI) attached to a Zeiss microscope interfaced with a PC. Once captured, the image was analyzed using image system software (Image-Pro Plus, v4.1; Media Cybernetics, Silver Spring, MD). The operator imaged the tissue section at low magnification and outlined for the computer straight portions of the cortical ribbon in a shape as close as possible to a rectangle or trapezoid. Once outlined, the computer measured the perimeter (P) and area (A). The average cortical thickness (CT) of the region of interest was calculated using the quadratic equation: CT = (P − (P2 − 16 × A)½)/4. One investigator performed these measures blinded to the pathology classification, and as previously indicated for neuronal density measurements, neocortical widths should be considered relative estimates.

NeuN Neuronal Size

The same imaging system was used to assess NeuN-labeled neuronal size. Images at ×50 were captured sampling regions from layer 1, the upper (levels 2–3) and lower (levels 5–6) gray matter, and superficial and deep white matter regions (Fig. 1). The operator outlined for the computer all individual NeuN-labeled cells within the captured image and the computer calculated the average area per cell for that section and region (μm2). Typically, 20 or more neurons were measured for each sample site.

Data Analysis

Data were entered into a database and analyzed using a statistical program (StatView 5; SAS Institute, Inc., Cary, NC). Differences between autopsy, non-CD and CD patient groups involving continuous dependent variables were compared statistically using an analysis of covariance (ANCOVA) that included the log of age at surgery or autopsy as co-independent variables (see Tables 2 and 3). Post-hoc statistical analyses used the Games–Howell test that controls for multiple comparisons of unequally sized samples and heterogeneous variances. Comparisons using nominal variables were performed using χ2 tests. Results were considered different at a minimal level of significance of P < 0.05.

Results

Patient Population

A total of 66 NeuN-stained sections from 52 patients were analyzed in this study (Table 1). Three autopsy cases, three non-CD and four CD patients contributed two or more sections from different gyri. The middle temporal gyrus was the most frequently sampled (n = 27), followed by the inferior frontal gyrus (n = 19), the middle frontal gyrus (n = 15) and the inferior parietal gyrus (n = 5). The distribution of brain region sample sites was not statistically different by patient category (χ2; P = 0.18). For non-CD and CD patients, the most frequent operation was hemispherotomy (n = 23), followed by temporal or frontal lobectomy (n = 8) and focal (non-lobar) resections (n = 7). There were no statistical differences in operative procedures between CD and non-CD patient groups (χ2; P = 0.19). Of the 25 CD patients, four (16%) had hemimegalencephaly and three (12%) had tuberous sclerosis. All CD patients except one had cytomegalic neurons and/or balloon cells in the surgical specimen consistent with a diagnosis of severe CD as previously defined (Mischel et al., 1995).

Table 1

Characteristics of patient groups

Patient group
 
Total sites
 
Male/Female
 
Left/Right
 
Age (years)
 
Seizure onset
 
Seizure duration
 
Autopsy (n = 13) 18 6/7 6/7 10.7 ± 3.3 (range 0–33) NA NA 
Non-CD (n = 14) 17 6/8 5/9 10.2 ± 2.5 (range 3–38) 4.0 ± 1.1 (range 0.5–15) 4.0 ± 1.1 (range 0.5–13) 
CD (n = 25) 31 10/15 14/11 6.4 ± 1.4 (range 0.2–30) 0.73 ± 0.2 (range 0–3) 3.9 ± 0.7 (range 0.2–12) 
P-value
 

 
0.94
 
0.47
 
0.30
 
0.0004
 
0.91
 
Patient group
 
Total sites
 
Male/Female
 
Left/Right
 
Age (years)
 
Seizure onset
 
Seizure duration
 
Autopsy (n = 13) 18 6/7 6/7 10.7 ± 3.3 (range 0–33) NA NA 
Non-CD (n = 14) 17 6/8 5/9 10.2 ± 2.5 (range 3–38) 4.0 ± 1.1 (range 0.5–15) 4.0 ± 1.1 (range 0.5–13) 
CD (n = 25) 31 10/15 14/11 6.4 ± 1.4 (range 0.2–30) 0.73 ± 0.2 (range 0–3) 3.9 ± 0.7 (range 0.2–12) 
P-value
 

 
0.94
 
0.47
 
0.30
 
0.0004
 
0.91
 

Data reported at mean ± SEM in years. NA, not applicable.

The clinical characteristics of the autopsy, non-CD, and CD patient groups are shown in Table 1. There were no statistically significant differences in gender (M/F), side resected (L/R), age at collection, or seizure duration between patient groups (P > 0.30). Age at seizure onset was less in CD patients compared with non-CD patients (P = 0.0004), similar to previous reports from this center (Mathern et al., 1999; Cook et al., 2004). A history of infantile spasms was noted in 13 (33%) surgical patients, and was not different between CD and non-CD groups (P = 0.06). Post-surgery (1 year or more) seizure outcomes were available on 35 patients; 83% were seizure free, supporting the notion that the seizure substrate was removed. This is a relatively young cohort, with 54% of cases aged 5 years or less at collection.

Significant postnatal changes occur in cerebral volumes, neuronal densities, and cell sizes as a logarithmic function of age in normal human neocortex. To determine if the same postnatal developmental processes take place in CD and non-CD cases we performed an analysis of covariance (ANCOVA), and the statistical results (F/P-values) are shown for the MRI assessed cerebral volumes (Table 2) and NeuN-defined neuronal densities and cell sizes (Table 3). The ANCOVA incorporated the pathology groups and log of age at surgery or collection as the independent statistical variables. Statistically significant interactions would indicate different changes with age between pathology groups. No significant interactions were found meaning that we found age-related changes for all groups and/or differences between pathology groups (Tables 2 and 3). The remainder of the Results section will sequentially discuss the statistically significant findings that occurred with age and/or pathology groups (Figs 2–11).

Figure 2.

Coronal MRI sections from patients with the least and greatest differences in hemispheric cerebral volumes for non-cortical dysplasia (non-CD; A and B), and cortical dysplasia patients (CD; C and D). Hemisphere side indicated at the bottom of (A). (A) Severe volume loss in the affected right hemisphere of a non-CD case. This 8.5-year-old had prolonged focal left body motor status epilepticus at age 1 year and subsequent MRI revealed severe atrophy of the right cerebral hemisphere. On the non-affected left hemisphere, the total cerebral volume measured 493 cc, the white matter was 175 cc, and the gray matter and basal ganglia (GM and BG) was 317 cc. On the affected right hemisphere, the cerebral volume was 324 cc (−34%), the white matter was 94 cc (−46%), and the GM and BG was 230 cc (−27.5%), indicating diffuse loss of gray and white matter. (B) Minimal volume changes in the affected right hemisphere of a non-CD case. This 8-year-old male had a 3 year history of limbic epilepsy from a right temporal low-grade brain tumor (DNET). On the non-affected left hemisphere, the cerebral volume measured 607 cc, the white matter was 280 cc, and the GM and BG was 327 cc. On the affected right hemisphere, the cerebral volume was 586 cc (−3.5%), the white matter was 230 cc (−18%), and the GM and BG was 356 cc (+9%) indicating minimal or no change of the affected hemisphere. (C) Minimal changes in the affected left hemisphere of a CD case. This 14-month-old presented with seizures at age 2 months apparently from cortical malformation involving the left temporal and peri-insular region (white arrows) that extended into the occipital and parietal lobes. On the non-affected right hemisphere, the cerebral volume measured 426 cc, the white matter was 161 cc, and the GM and BG was 265 cc. On the affected left hemisphere, the cerebral volume was 426 cc (0%), the white matter was 127 cc (−21%), and the GM and GB was 298 cc (+12.5%), indicating slightly greater gray matter and basal ganglia volume in the affected hemisphere. (D) An example of an enlarged affected left hemisphere of a CD case. This 15-month-old began to have seizures shortly after birth arising from the left hemisphere (by EEG) that involved the entire left hemisphere by MRI, and the brain showed pachygyri (arrow). On the non-affected right hemisphere, the cerebral volume measured 379 cc, the white matter was 120 cc, and the GM and BG was 258 cc. On the affected left hemisphere, the cerebral volume was 432 cc (+14%), the white matter was 116 cc (−3%), and the GM and BG was 316 cc (+22.5%), indicating an enlarged affected hemisphere with more MRI identified gray matter.

Figure 2.

Coronal MRI sections from patients with the least and greatest differences in hemispheric cerebral volumes for non-cortical dysplasia (non-CD; A and B), and cortical dysplasia patients (CD; C and D). Hemisphere side indicated at the bottom of (A). (A) Severe volume loss in the affected right hemisphere of a non-CD case. This 8.5-year-old had prolonged focal left body motor status epilepticus at age 1 year and subsequent MRI revealed severe atrophy of the right cerebral hemisphere. On the non-affected left hemisphere, the total cerebral volume measured 493 cc, the white matter was 175 cc, and the gray matter and basal ganglia (GM and BG) was 317 cc. On the affected right hemisphere, the cerebral volume was 324 cc (−34%), the white matter was 94 cc (−46%), and the GM and BG was 230 cc (−27.5%), indicating diffuse loss of gray and white matter. (B) Minimal volume changes in the affected right hemisphere of a non-CD case. This 8-year-old male had a 3 year history of limbic epilepsy from a right temporal low-grade brain tumor (DNET). On the non-affected left hemisphere, the cerebral volume measured 607 cc, the white matter was 280 cc, and the GM and BG was 327 cc. On the affected right hemisphere, the cerebral volume was 586 cc (−3.5%), the white matter was 230 cc (−18%), and the GM and BG was 356 cc (+9%) indicating minimal or no change of the affected hemisphere. (C) Minimal changes in the affected left hemisphere of a CD case. This 14-month-old presented with seizures at age 2 months apparently from cortical malformation involving the left temporal and peri-insular region (white arrows) that extended into the occipital and parietal lobes. On the non-affected right hemisphere, the cerebral volume measured 426 cc, the white matter was 161 cc, and the GM and BG was 265 cc. On the affected left hemisphere, the cerebral volume was 426 cc (0%), the white matter was 127 cc (−21%), and the GM and GB was 298 cc (+12.5%), indicating slightly greater gray matter and basal ganglia volume in the affected hemisphere. (D) An example of an enlarged affected left hemisphere of a CD case. This 15-month-old began to have seizures shortly after birth arising from the left hemisphere (by EEG) that involved the entire left hemisphere by MRI, and the brain showed pachygyri (arrow). On the non-affected right hemisphere, the cerebral volume measured 379 cc, the white matter was 120 cc, and the GM and BG was 258 cc. On the affected left hemisphere, the cerebral volume was 432 cc (+14%), the white matter was 116 cc (−3%), and the GM and BG was 316 cc (+22.5%), indicating an enlarged affected hemisphere with more MRI identified gray matter.

Figure 3.

Scatter plots showing the changes in MRI assessed cerebral volumes by age for non-CD and CD patients. The P-values from the ANCOVA (Table 2) are indicated along with the r-values. Upper left and right graph: affected and non-affected total cerebral and white matter volumes logarithmically increased with age (P < 0.0044). Lower left graph: by comparison, affected and non-affected gray matter and basal ganglia volumes did not correlate with age (P = 0.73).

Figure 3.

Scatter plots showing the changes in MRI assessed cerebral volumes by age for non-CD and CD patients. The P-values from the ANCOVA (Table 2) are indicated along with the r-values. Upper left and right graph: affected and non-affected total cerebral and white matter volumes logarithmically increased with age (P < 0.0044). Lower left graph: by comparison, affected and non-affected gray matter and basal ganglia volumes did not correlate with age (P = 0.73).

Figure 4.

Bar graphs showing the mean (± SEM) MRI assessed volumes (upper graph) and differences between affected and non-affected cerebral hemispheres (lower graph) for non-CD and CD cases. The P-values from the ANCOVA (Table 2) are indicated. Upper graph: statistical analyses indicated that controlling for changes associated with age (ANCOVA; Table 2) there were no significant differences between non-CD and CD cases in cerebral, white matter, and gray matter and basal ganglia volumes between the affected and non-affected cerebral hemispheres (P > 0.13). Lower graph: comparison of the differences in MRI assessed volumes of the affected versus non-affected hemisphere were statistically different between non-CD and CD cases. For total cerebral volume, non-CD cases showed a mean 14.5% (−72 cc) decrease of the affected hemisphere compared with the non-affected side, and CD cases showed a mean 3.2% (−12 cc) increase (lower left; P = 0.017). For gray matter and basal ganglia volumes, non-CD cases showed a mean 12.5% (−40 cc) decrease in the affected hemisphere while the CD cases showed a mean 8% (+21 cc) increase (lower middle; P = 0.014). For white matter volumes, the non-CD cases showed a mean 18% (−36 cc) decrease and the CD cases showed a mean 4% (−8 cc) decrease (lower right; P = 0.049).

Figure 4.

Bar graphs showing the mean (± SEM) MRI assessed volumes (upper graph) and differences between affected and non-affected cerebral hemispheres (lower graph) for non-CD and CD cases. The P-values from the ANCOVA (Table 2) are indicated. Upper graph: statistical analyses indicated that controlling for changes associated with age (ANCOVA; Table 2) there were no significant differences between non-CD and CD cases in cerebral, white matter, and gray matter and basal ganglia volumes between the affected and non-affected cerebral hemispheres (P > 0.13). Lower graph: comparison of the differences in MRI assessed volumes of the affected versus non-affected hemisphere were statistically different between non-CD and CD cases. For total cerebral volume, non-CD cases showed a mean 14.5% (−72 cc) decrease of the affected hemisphere compared with the non-affected side, and CD cases showed a mean 3.2% (−12 cc) increase (lower left; P = 0.017). For gray matter and basal ganglia volumes, non-CD cases showed a mean 12.5% (−40 cc) decrease in the affected hemisphere while the CD cases showed a mean 8% (+21 cc) increase (lower middle; P = 0.014). For white matter volumes, the non-CD cases showed a mean 18% (−36 cc) decrease and the CD cases showed a mean 4% (−8 cc) decrease (lower right; P = 0.049).

Figure 5.

NeuN-stained sections of middle temporal gyral neocortex of autopsy cases to illustrate neuronal density changes as a function of age. (A–C) With increased age there were visible decreases in neurons per unit area (i.e. density) within the cortical gray matter. Averaged level 1–6 NeuN densities were: (A, 0.75 years) 1290 neurons/μm2, (B, 2 years) 947 neurons/μm2 and (C, 4 years) 818 neurons/μm2. By comparison, cortical thickness did not change as a function of age over the range of this study. (D–F) Similar changes in neuronal densities were observed in the white matter. Averaged superficial and deep white matter NeuN densities were (D) 111 neurons/μm2, (E) 42 neurons/μm2 and (F) 33 neurons/μm2. All micrographs at identical magnification.

Figure 5.

NeuN-stained sections of middle temporal gyral neocortex of autopsy cases to illustrate neuronal density changes as a function of age. (A–C) With increased age there were visible decreases in neurons per unit area (i.e. density) within the cortical gray matter. Averaged level 1–6 NeuN densities were: (A, 0.75 years) 1290 neurons/μm2, (B, 2 years) 947 neurons/μm2 and (C, 4 years) 818 neurons/μm2. By comparison, cortical thickness did not change as a function of age over the range of this study. (D–F) Similar changes in neuronal densities were observed in the white matter. Averaged superficial and deep white matter NeuN densities were (D) 111 neurons/μm2, (E) 42 neurons/μm2 and (F) 33 neurons/μm2. All micrographs at identical magnification.

Figure 6.

Scatter plots showing changes in neocortical and white matter NeuN densities or cortical thickness by age for autopsy, non-CD and CD cases. The P-values from the ANCOVA (Table 3) are indicated along with the r-values. Upper row: by age, there were no statistically significant changes in cortical thickness (upper left; P = 0.45) or layer 1 neuronal densities (P = 0.86). Lower row: by comparison, NeuN densities for levels 1–6 of the neocortex (lower left; P = 0.0015) and superficial and deep white matter (lower right; P = 0.0038) logarithmically decreased with age.

Figure 6.

Scatter plots showing changes in neocortical and white matter NeuN densities or cortical thickness by age for autopsy, non-CD and CD cases. The P-values from the ANCOVA (Table 3) are indicated along with the r-values. Upper row: by age, there were no statistically significant changes in cortical thickness (upper left; P = 0.45) or layer 1 neuronal densities (P = 0.86). Lower row: by comparison, NeuN densities for levels 1–6 of the neocortex (lower left; P = 0.0015) and superficial and deep white matter (lower right; P = 0.0038) logarithmically decreased with age.

Figure 7.

NeuN-stained sections illustrating neuronal density changes for layer 1 and upper cortex (levels 1–3; A–C), lower cortex (levels 4–6; D–F), and white matter (G–I) of middle temporal gyrus from an autopsy case (A, D, G), middle frontal gyrus of a non-CD patient (B, E, H), and middle temporal gyrus from a CD case (C, F, I). All patients were of similar age (3.75–4.3 years). Compared with the autopsy and non-CD case, the CD case shows an increased density of NeuN-stained neurons in layer 1 and upper levels of the gray matter (compare C with A and B). NeuN densities did not appear very different between sections of the lower gray matter (panels D–F). However, NeuN densities in the white matter were increased in the CD case compared with autopsy and non-CD examples (compare I with G and H). All micrographs at identical magnification.

Figure 7.

NeuN-stained sections illustrating neuronal density changes for layer 1 and upper cortex (levels 1–3; A–C), lower cortex (levels 4–6; D–F), and white matter (G–I) of middle temporal gyrus from an autopsy case (A, D, G), middle frontal gyrus of a non-CD patient (B, E, H), and middle temporal gyrus from a CD case (C, F, I). All patients were of similar age (3.75–4.3 years). Compared with the autopsy and non-CD case, the CD case shows an increased density of NeuN-stained neurons in layer 1 and upper levels of the gray matter (compare C with A and B). NeuN densities did not appear very different between sections of the lower gray matter (panels D–F). However, NeuN densities in the white matter were increased in the CD case compared with autopsy and non-CD examples (compare I with G and H). All micrographs at identical magnification.

Figure 8.

Bar graphs showing mean (± SEM) NeuN densities or cortical thickness for autopsy, non-CD, and CD patients. The P-values from the ANCOVA (Table 3) are indicated, and significant post-hoc differences (P < 0.05) compared with autopsy cases indicated by an asterisk. ANCOVA found differences in NeuN densities by pathology group in layer 1, level 2 and 3 of the neocortex, and the superficial and deep white matter (P < 0.021). Post-hoc analysis found that for: (i) layer 1, CD was greater than the other two groups (P < 0.038); (ii) level 2, CD was greater than autopsy (P = 0.005); (iii) level 3, CD was greater than autopsy (P = 0.005); (iv) superficial white matter (Sup WM), CD was greater than the other two groups (P < 0.0024); and (v) deep white matter (Deep WM), CD was greater than the other two groups (P < 0.0001).

Figure 8.

Bar graphs showing mean (± SEM) NeuN densities or cortical thickness for autopsy, non-CD, and CD patients. The P-values from the ANCOVA (Table 3) are indicated, and significant post-hoc differences (P < 0.05) compared with autopsy cases indicated by an asterisk. ANCOVA found differences in NeuN densities by pathology group in layer 1, level 2 and 3 of the neocortex, and the superficial and deep white matter (P < 0.021). Post-hoc analysis found that for: (i) layer 1, CD was greater than the other two groups (P < 0.038); (ii) level 2, CD was greater than autopsy (P = 0.005); (iii) level 3, CD was greater than autopsy (P = 0.005); (iv) superficial white matter (Sup WM), CD was greater than the other two groups (P < 0.0024); and (v) deep white matter (Deep WM), CD was greater than the other two groups (P < 0.0001).

Figure 9.

High power micrographs of NeuN sections illustrating changes in neuronal size by cortical region. The left column is from a 3.9-year-old autopsy case (A, D, G, J, M), the center column a 3.5-year-old non-CD patient (B, E, H, K, N), and the right column is a 3.2-year-old CD case (C, F, I, L, O). The top row shows layer 1 (A–C), followed by upper gray matter (D–F; levels 2–3), lower gray matter (G–I; levels 5–6), upper white matter (J–L) and lower white matter (M–O). Note that most NeuN cells in the CD and non-CD case are larger than the autopsy case for all cortical and white matter areas (compare right and center with left column). In addition to being larger, the cells in the CD case show different morphologies, especially in layer 1 (C), lower gray matter (I) and upper white matter (L). All micrographs at equal magnification.

Figure 9.

High power micrographs of NeuN sections illustrating changes in neuronal size by cortical region. The left column is from a 3.9-year-old autopsy case (A, D, G, J, M), the center column a 3.5-year-old non-CD patient (B, E, H, K, N), and the right column is a 3.2-year-old CD case (C, F, I, L, O). The top row shows layer 1 (A–C), followed by upper gray matter (D–F; levels 2–3), lower gray matter (G–I; levels 5–6), upper white matter (J–L) and lower white matter (M–O). Note that most NeuN cells in the CD and non-CD case are larger than the autopsy case for all cortical and white matter areas (compare right and center with left column). In addition to being larger, the cells in the CD case show different morphologies, especially in layer 1 (C), lower gray matter (I) and upper white matter (L). All micrographs at equal magnification.

Figure 10.

Scatter plots showing changes in neocortical and white matter NeuN assessed neuronal size by age for autopsy, non-CD, and CD cases. The P-values from the ANCOVA (Table 3) are indicated along with the r-values. Upper row: by age, there were no statistically significant changes in neuronal size in layer 1 or the upper neocortical gray matter (P > 0.58). Middle row: by comparison, neuronal size logarithmically increased with age for neurons in the lower gray matter and superficial white matter (P < 0.019). Lower graph: neuronal size for deep white matter NeuN-stained cells did not change as a function of age (P = 0.10).

Figure 10.

Scatter plots showing changes in neocortical and white matter NeuN assessed neuronal size by age for autopsy, non-CD, and CD cases. The P-values from the ANCOVA (Table 3) are indicated along with the r-values. Upper row: by age, there were no statistically significant changes in neuronal size in layer 1 or the upper neocortical gray matter (P > 0.58). Middle row: by comparison, neuronal size logarithmically increased with age for neurons in the lower gray matter and superficial white matter (P < 0.019). Lower graph: neuronal size for deep white matter NeuN-stained cells did not change as a function of age (P = 0.10).

Figure 11.

Bar graphs showing mean (± SEM) NeuN neuronal size for autopsy, non-CD and CD patients (upper graph), and neuronal size in surgical patients with and without a history of infantile spasms (lower graph). The P-values from the ANCOVA (Table 3) are indicated for the upper graph, and significant post-hoc differences (P < 0.05) compared with autopsy cases indicated by asterisks (*). Upper graph: ANCOVA found differences in neuronal size by pathology group in all gray and white matter areas studied (P < 0.0323). Post-hoc analysis found the following. In layer 1, CD and non-CD had greater neuronal size than autopsy cases (P < 0.0025). Upper gray matter (levels 2–3) neuronal size for non-CD and CD cases was greater than autopsy cases (P < 0.0036). Lower gray matter (levels 5–6) neuronal size for non-CD and CD cases was greater than autopsy cases (P < 0.011). Superficial white matter neuronal size for non-CD and CD cases was greater than autopsy cases (P < 0.0176). Deep white matter neuronal size for CD cases was greater than autopsy cases (P = 0.0032). Lower graph: CD and non-CD cases with a history of infantile spasms had neuronal hypertrophy in layer 1 NeuN cells compared with patients without a history of spasms (P = 0.0015). This finding was still statistically significant in an ANCOVA controlling for patient pathology group and age.

Figure 11.

Bar graphs showing mean (± SEM) NeuN neuronal size for autopsy, non-CD and CD patients (upper graph), and neuronal size in surgical patients with and without a history of infantile spasms (lower graph). The P-values from the ANCOVA (Table 3) are indicated for the upper graph, and significant post-hoc differences (P < 0.05) compared with autopsy cases indicated by asterisks (*). Upper graph: ANCOVA found differences in neuronal size by pathology group in all gray and white matter areas studied (P < 0.0323). Post-hoc analysis found the following. In layer 1, CD and non-CD had greater neuronal size than autopsy cases (P < 0.0025). Upper gray matter (levels 2–3) neuronal size for non-CD and CD cases was greater than autopsy cases (P < 0.0036). Lower gray matter (levels 5–6) neuronal size for non-CD and CD cases was greater than autopsy cases (P < 0.011). Superficial white matter neuronal size for non-CD and CD cases was greater than autopsy cases (P < 0.0176). Deep white matter neuronal size for CD cases was greater than autopsy cases (P = 0.0032). Lower graph: CD and non-CD cases with a history of infantile spasms had neuronal hypertrophy in layer 1 NeuN cells compared with patients without a history of spasms (P = 0.0015). This finding was still statistically significant in an ANCOVA controlling for patient pathology group and age.

Table 2

ANCOVA statistical results (F/P-values) of MRI assessed cerebral, white matter and gray matter/basal ganglia volumes for cortical dysplasia and non-cortical dysplasia cases

MRI assessment
 
CD versus non-CD
 
Age
 
Interaction
 
Figures
 
Total cerebral vol. 0.695/0.413 10.1/0.0044 0.574/0.456 2–4 
Total GM and BG vol. 0.286/0.599 0.123/0.730 0.429/0.520 2–4 
Total WM vol. 0.612/0.444 12.8/0.0020 0.558/0.464 2–4 
Affected cerebral vol. 2.49/0.129 9.4/0.0057 1.29/0.267 2–4 
Affected GM and BG vol. 2.03/0.170 0.667/0.424 1.29/0.271 2–4 
Affected WM vol. 1.54/0.230 10.5/0.0044 1.01/0.327 2–4 
Non-aff. cerebral vol. 0.550/0.817 9.7/0.0051 0.066/0.799 2–4 
Non-aff. GM and BG vol. 0.434/0.518 0.005/0.942 0.001/0.976 2–4 
Non-aff. WM vol 0.003/0.959 14.0/0.0014 0.057/0.814 2–4 
Subtracted variables     
    Aff. – non-aff. cerebral vol. 6.6/0.0174 2.19/0.153 2.05/0.166 2–4 
    Aff. – non-aff. GM and BG 7.4/0.0136 1.58/0.224 2.60/0.123 2–4 
    Aff. – non-aff. WM
 
4.42/0.0492
 
1.82/0.193
 
1.58/0.176
 
2–4
 
MRI assessment
 
CD versus non-CD
 
Age
 
Interaction
 
Figures
 
Total cerebral vol. 0.695/0.413 10.1/0.0044 0.574/0.456 2–4 
Total GM and BG vol. 0.286/0.599 0.123/0.730 0.429/0.520 2–4 
Total WM vol. 0.612/0.444 12.8/0.0020 0.558/0.464 2–4 
Affected cerebral vol. 2.49/0.129 9.4/0.0057 1.29/0.267 2–4 
Affected GM and BG vol. 2.03/0.170 0.667/0.424 1.29/0.271 2–4 
Affected WM vol. 1.54/0.230 10.5/0.0044 1.01/0.327 2–4 
Non-aff. cerebral vol. 0.550/0.817 9.7/0.0051 0.066/0.799 2–4 
Non-aff. GM and BG vol. 0.434/0.518 0.005/0.942 0.001/0.976 2–4 
Non-aff. WM vol 0.003/0.959 14.0/0.0014 0.057/0.814 2–4 
Subtracted variables     
    Aff. – non-aff. cerebral vol. 6.6/0.0174 2.19/0.153 2.05/0.166 2–4 
    Aff. – non-aff. GM and BG 7.4/0.0136 1.58/0.224 2.60/0.123 2–4 
    Aff. – non-aff. WM
 
4.42/0.0492
 
1.82/0.193
 
1.58/0.176
 
2–4
 

Affected, affected cerebral hemisphere; BG, basal ganglial GM, gray matter; non-aff., non-affected cerebral hemisphere; WM, white matter. Significant values are indicated in bold type.

Table 3

ANCOVA statistical results (F/P-values) of NeuN assessed cell densities and neuronal sizes for autopsy, cortical dysplasia and non-cortical dysplasia cases

Cortical measure
 
Pathology category
 
Age
 
Interaction
 
Figures
 
Cortical thickness 0.138/0.871 0.578/0.450 0.221/0.802 5–8 
Neuronal densites     
    Layer 1 NeuN 6.42/0.0023 0.033/0.856 0.450/0.640 5–8 
    Level 1 NeuN 1.58/0.214 4.94/0.030 2.20/0.120 5–8 
    Level 2 NeuN 3.86/0.021 4.25/0.044 1.31/0.276 5–8 
    Level 3 NeuN 4.44/0.0162 0.313/0.578 2.00/0.144 5–8 
    Level 4 NeuN 1.11/0.336 3.72/0.059 2.41/0.168 5–8 
    Level 5 NeuN 0.186/0.830 11.4/0.0013 0.506/0.605 5–8 
    Level 6 NeuN 0.255/0.776 4.67/0.035 0.524/0.595 5–8 
    Mean level 1–6 0.140/0.255 11.1/0.0015 1.51/0.193 5–8 
    Superficial WM NeuN 6.83/0.0022 4.86/0.031 0.591/0.557 5–8 
    Deep WM NeuN 10.5/<0.0001 2.71/0.105 1.43/0.247 5–8 
    Mean WM 3.88/0.026 9.05/0.0038 1.66/0.188 5–8 
White/gray ratios 4.16/0.021 1.78/0.187 0.161/0.851  
Neuronal cell size     
    Layer 1 NeuN 5.11/0.0091 0.258/0.613 0.124/0.883 9–11 
    Upper GM NeuN 4.24/0.0192 0.312/0.579 0.130/0.878 9–11 
    Lower GM NeuN 3.64/0.0323 4.25/0.019 0.117/0.890 9–11 
    Superficial WM NeuN 4.05/0.0224 5.93/0.0179 0.896/0.414 9–11 
    Deep WM NeuN
 
4.15/0.0207
 
2.80/0.100
 
0.992/0.377
 
9–11
 
Cortical measure
 
Pathology category
 
Age
 
Interaction
 
Figures
 
Cortical thickness 0.138/0.871 0.578/0.450 0.221/0.802 5–8 
Neuronal densites     
    Layer 1 NeuN 6.42/0.0023 0.033/0.856 0.450/0.640 5–8 
    Level 1 NeuN 1.58/0.214 4.94/0.030 2.20/0.120 5–8 
    Level 2 NeuN 3.86/0.021 4.25/0.044 1.31/0.276 5–8 
    Level 3 NeuN 4.44/0.0162 0.313/0.578 2.00/0.144 5–8 
    Level 4 NeuN 1.11/0.336 3.72/0.059 2.41/0.168 5–8 
    Level 5 NeuN 0.186/0.830 11.4/0.0013 0.506/0.605 5–8 
    Level 6 NeuN 0.255/0.776 4.67/0.035 0.524/0.595 5–8 
    Mean level 1–6 0.140/0.255 11.1/0.0015 1.51/0.193 5–8 
    Superficial WM NeuN 6.83/0.0022 4.86/0.031 0.591/0.557 5–8 
    Deep WM NeuN 10.5/<0.0001 2.71/0.105 1.43/0.247 5–8 
    Mean WM 3.88/0.026 9.05/0.0038 1.66/0.188 5–8 
White/gray ratios 4.16/0.021 1.78/0.187 0.161/0.851  
Neuronal cell size     
    Layer 1 NeuN 5.11/0.0091 0.258/0.613 0.124/0.883 9–11 
    Upper GM NeuN 4.24/0.0192 0.312/0.579 0.130/0.878 9–11 
    Lower GM NeuN 3.64/0.0323 4.25/0.019 0.117/0.890 9–11 
    Superficial WM NeuN 4.05/0.0224 5.93/0.0179 0.896/0.414 9–11 
    Deep WM NeuN
 
4.15/0.0207
 
2.80/0.100
 
0.992/0.377
 
9–11
 

Significant values indicated in bold type.

MRI Assessed Cerebral Volumes

Pre-surgery MRI studies performed at UCLA were available on 26 (67%) cases, with 11 in the non-CD and 15 in the CD group. Pre-surgery MRI scans for the other surgical cases were performed at outside hospitals, or there was too much movement artifact on the UCLA scan to perform the volumetric analysis. None of the four hemimegalencephaly cases were used in the MRI analysis. Results of the ANCOVA for the MRI measurements found six variables that logarithmically increased with age and three factors that were different between CD and non-CD cases (Table 2; Figs 2–4). By age, total cerebral and white matter volumes of the affected and non-affected cerebral hemispheres logarithmically increased in size (P < 0.0051). MRI gray matter/basal ganglia volumes showed no statistically significant changes with age (P > 0.42; Table 1 and Fig. 3).

There were differences in cerebral volumes comparing the affected with the non-affected hemisphere between CD and non-CD cases. Qualitatively, in non-CD cases the affected cerebral hemisphere was often smaller compared with the non-affected side, with loss of gray and white matter volumes (Fig. 2; upper row). By comparison, in CD cases the affected hemisphere was the same or slightly larger than the non-affected side with more visible gray matter by T1 or T2-weighted MRI (Fig. 2; lower row). Quantitatively, ANCOVA found no statistically significant differences by patient group for total cerebral, white matter, or gray matter volumes of the affected or non-affected hemispheres with age as a co-variable (Table 2; Fig. 4, upper graph). However, when the volumes of the affected hemisphere were subtracted from the non-affected sides, differences were noted between CD and non-CD cases (Table 2; Fig. 4, lower graph; P < 0.049). For non-CD cases, the total cerebral volumes, along with gray matter and white matter volumes were decreased on the affected compared with the non-effected hemisphere. For CD cases, the total cerebral volumes and gray matter volumes were slightly increased (3 and 8% respectively) in the affected hemisphere. The difference between CD and non-CD cases comparing the difference in volume between the two hemispheres was statistically significant for total cerebral volume, gray and white matter volumes (P = 0.02, P = 0.01 and P = 0.05, respectively).

Analysis of MRI assessed volumes of surgical cases showed no statistically significant results compared with other clinical factors. In this surgical cohort, gender (P > 0.67), side resected (P > 0.06), seizure duration (P > 0.07) and history of infantile spasms (P > 0.16) did not correlate with cerebral volumes, or white and gray matter volumes.

NeuN Neuronal Densities and Cortical Thickness

Data were available on all 52 patients in this cohort. ANCOVA found seven neuronal density variables that logarithmically decreased with age, while six factors were different between autopsy, non-CD and CD cases (Table 3; Figs 5–8). Qualitatively and quantitatively, neuronal densities of the cortical gray and white matter logarithmically decreased with age (P < 0.044; Table 3; Figs 5 and 6). By comparison, cortical thickness and layer 1 NeuN cell densities did not change with age.

Statistically accounting for the changes that occurred with age (ANCOVA), there were differences in neuronal densities between pathology groups, with CD cases showing increased neuronal densities in layer 1 and white matter regions compared with non-CD and autopsy cases (Table 3; Figs 7 and 8). In addition, there were apparent differences in somal morphology based on NeuN staining between patient groups. In non-CD and autopsy cases, most of the NeuN-labeled white matter neurons had fusiform and pyramidal morphologies, similar to descriptions from previous publications describing Nissl or Golgi stained interstitial neurons in the white matter of primates and humans (Figs 7G,H and 9J,K,M,N) (Kostovic and Rakic, 1980). In CD cases, by comparison, intermixed among the normal-appearing NeuN-labeled neurons in the lower gray and superficial white matter were other neurons of varied and complex somal morphologies. These included inverted pyramidal neurons and large cytomegalic neurons, which had morphologic similarities to large multipolar or polymorphous cells previously described in the human subplate using Nissl or Golgi techniques (Figs 7F,I and 9I,L) (Mrzljak et al., 1988, 1992; Kostovic and Rakic, 1990). Likewise, some NeuN-labeled neurons in layer 1 of autopsy and non-CD patients had morphologies consistent with Cajal-Retzius cells (data not shown), while CD cases demonstrated more diverse cell types with stellate and piriform shapes similar in appearance to previous descriptions of molecular layer cells in prenatal primates and humans (Fig. 9AC) (Marin-Padilla and Marin-Padilla, 1982; Zecevic et al., 1999; Zecevic and Rakic, 2001). Furthermore, review of the upper cortex, especially levels 2 and 3, showed slightly more neurons per unit area in CD patients compared with autopsy cases, and most of the cells were normal pyramidal shaped with the primary dendrite directed toward the pial surface (Figs 7C, 8 and 9D,F). Finally, the ratio of white to gray matter neurons was statistically different by pathology group (Table 3; P = 0.021). Post hoc analyses showed that CD cases (mean ± SEM = 16.8 ± 1.3) had lower ratios (i.e. more white matter neurons) compared with non-CD (27.9 ± 4.1) and autopsy (27.8 ± 2.8) cases (P < 0.003).

In this surgical cohort, NeuN neuronal densities did not correlate with gender (P > 0.22), side resected (P > 0.15), seizure duration (P > 0.07) or history of infantile spasms (P > 0.12). For CD and non-CD cases, however, decreased neuronal densities in levels 5 and 6 of the gray matter correlated with greater MRI assessed cerebral volumes of the affected hemisphere (level 5, r = −0.502, P = 0.012: level 6, r = −0.466, P = 0.022). In other words, there was a statistically significant correlation between postnatal increased cerebral volumes and decreased neuronal densities in the lower gray matter. Other changes of NeuN neuronal densities in layer 1, levels 1–4, superficial and deep white matter did not correlate with MRI assessed cerebral volumes, or white or gray matter volumes (P > 0.14).

Despite the changes in neuronal densities, cortical thickness was not different between autopsy, non-CD and CD cases (Figs 7 and 8: lower left). In addition, cortical thickness did not correlate with gender, side, seizure duration, history of infantile spasms, or changes in MRI assessed cerebral volumes of gray and white matter (P > 0.37).

Estimate of Total Neurons in the Affected Hemisphere

For CD and non-CD cases, we multiplied the averaged cortical ribbon NeuN densities (levels 1–6) by MRI assessed gray matter volumes, and added the white matter densities (superficial and deep) multiplied by MRI white matter volumes to estimate the total neurons per affected hemisphere. ANCOVA found a difference between CD and non-CD cases (P = 0.003) without corresponding changes as a logarithmic function of age (P = 0.11; interaction, P = 0.10). CD cases had an estimated 3.52 × 1017 neurons per affected hemisphere compared with 3.03 × 1017 neurons for non-CD cases, a 16% increase in estimated cell number.

Neuronal Size

NeuN somal size data were available on all 52 patients. ANCOVA found two size variables that logarithmically increased with age and five factors that were different between autopsy, non-CD and CD cases (Table 3; Figs 9–11). Visually and by image analysis, NeuN-labeled cell sizes increased logarithmically with age in neurons of the lower gray matter and superficial white sample sites (Fig. 10; P < 0.019). By comparison, no changes in neuronal size by age were observed in layer 1, the upper gray matter, and deep white matter regions (P > 0.10).

In addition to the changes with age, non-CD and CD cases showed neuronal hypertrophy compared with autopsy cases (Table 3; Figs 9 and 11). NeuN assessed neuronal hypertrophy was observed in layer 1, upper and lower gray matter, and superficial and deep white matter. Finally, an intriguing result was the finding of neuronal hypertrophy in layer 1 cells for patients with a history of infantile spasms compared with those cases without a history of spasms (Fig. 11; lower graph; P = 0.0015). Repeating the statistical analysis controlling for pathology category (ANCOVA) confirmed the difference between those with or without a history of infantile spasms (spasms, P = 0.0001; pathology, P = 0.036; interaction, P = 0.065), as did an ANCOVA controlling for age (spasms, P = 0.0006; age, P = 0.31; interaction, P = 0.33). In other cell layers, neuronal size did not correlate with a history of spasms (P > 0.11). Neuronal size did not correlate with gender (P > 0.11), side (P > 0.52), seizure duration (P > 0.09), changes in neuronal densities (P > 0.07) or MRI assessed cerebral volumes (P > 0.11).

Discussion

In mostly pediatric epilepsy surgery patients, this study found changes in MRI-assessed cerebral gray and white matter volumes, NeuN-defined neuronal densities, and size that related to postnatal cerebral growth, pathology group and seizure history. The main findings from our study are summarized in Table 4, and provide potential clues to the timing of CD pathogenesis during corticoneurogenesis.

Table 4

Major findings from this study


Findings related to pathology group 

 
    For CD cases 1. Increased MRI cerebral (3%) and gray matter (8%) volumes of the affected hemisphere 
 2. Increased layer 1 (131%), upper cortical (9–23%) and white matter (28–77%) NeuN neuronal densities compared with autopsy cases 
 3. More complex NeuN morphologies in layer 1, lower gray, and superficial white matter compared with non-CD and autopsy cases 
 4. Increased white to gray matter NeuN cell ratios compared with non-CD and autopsy cases 
 5. Increased estimated total neurons (16%) of the affected hemisphere compared with non-CD cases (MRI volumes×NeuN densities) 
    For non-CD cases 1. Decreased cerebral, white, and gray matter (14–18%) volumes of the affected hemisphere 
 2. Similar neuronal densities as autopsy cases 
Changes related to seizures 1. Increased (34–63%) neuronal size in all gray and white matter layers compared with autopsy cases 
 2. Increased (40%) neuronal size of layer 1 cells in patients with a history of infantile spasms 
Changes related to age 1. Logarithmic increase in cerebral and white matter volumes of both hemispheres 
 2. Logarithmic decrease in gray and white matter neuronal densities 

 
3. Logarithmic increase in neuronal size in the lower gray matter and superficial white matter
 

Findings related to pathology group 

 
    For CD cases 1. Increased MRI cerebral (3%) and gray matter (8%) volumes of the affected hemisphere 
 2. Increased layer 1 (131%), upper cortical (9–23%) and white matter (28–77%) NeuN neuronal densities compared with autopsy cases 
 3. More complex NeuN morphologies in layer 1, lower gray, and superficial white matter compared with non-CD and autopsy cases 
 4. Increased white to gray matter NeuN cell ratios compared with non-CD and autopsy cases 
 5. Increased estimated total neurons (16%) of the affected hemisphere compared with non-CD cases (MRI volumes×NeuN densities) 
    For non-CD cases 1. Decreased cerebral, white, and gray matter (14–18%) volumes of the affected hemisphere 
 2. Similar neuronal densities as autopsy cases 
Changes related to seizures 1. Increased (34–63%) neuronal size in all gray and white matter layers compared with autopsy cases 
 2. Increased (40%) neuronal size of layer 1 cells in patients with a history of infantile spasms 
Changes related to age 1. Logarithmic increase in cerebral and white matter volumes of both hemispheres 
 2. Logarithmic decrease in gray and white matter neuronal densities 

 
3. Logarithmic increase in neuronal size in the lower gray matter and superficial white matter
 

A Hypothesis on the Timing of CD Pathogenesis during Corticogenesis

Based on an understanding of normal neocortical development in humans and the number, location, and appearance of abnormal neurons, we propose that severe CD associated with epilepsy could be the consequence of excess neurogenesis and postnatal retention of some preplate neurons, and this process probably occurs in late corticoneurogenesis. We did not find the severely malformed four-layer cortex, like that observed in patients with lissencephaly, nor decreased gray matter neuronal densities associated with microcephaly, similar to that observed in irradiated rodent models of cortical dysplasia. Thus, in patients with epilepsy, the abnormal cortical development did not fit the pattern that one would expect if severe CD, as originally described by Taylor (Taylor et al., 1971), were the consequence of developmental processes that significantly altered early to mid-corticoneurogenesis. Instead, in CD cases we found normal or slightly enlarged cerebral hemispheric volumes compared with the non-affected side, and more neurons than expected in the cortical ribbon and remnants of the molecular (layer 1) and subplate (superficial white matter) regions. In addition, the somal appearance of some NeuN-labeled neurons in layer 1 and lower gray and upper white matter were similar to descriptions of prenatal molecular layer and subplate neurons found between 17 and 32 weeks gestation in human studies using different staining techniques (Marin-Padilla and Marin-Padilla, 1982; Mrzljak et al., 1988, 1992; Marin-Padilla, 1998). In the subplate, these previously described neurons include large multipolar cells with spherical or ovoid shapes, polymorphous cells with bizarre somal shapes and thick primary dendrites, fusiform cells, and normal and inverted pyramidal neurons. The polymorphous and multipolar cells previously described using Golgi techniques show a number of similarities to biocytin filled cytomegalic cells previously recorded from surgically treated CD cases from our laboratory (Kostovic and Rakic, 1990; Mathern et al., 2000; Cepeda et al., 2003). In our prior human in vitro experiments, we found that cytomegalic neurons from CD tissue were electrophysiologically active, were located most often at the gray–white matter junction, were obtained from gyri that at surgery were broad and enlarged as if the secondary folding process was interrupted or incomplete, and had numerous dendritic processes with an identifiable axon (see Cepeda et al., 2003, figs 2 and 3B2). Hence, in pediatric CD patients with epilepsy we propose that some of the large subplate cells did not undergo normal degeneration prior to birth. As a consequence some of the ‘bizarre’ abnormal appearing cytomegalic neurons in CD tissue, previously thought to be derived from abnormal neurogenesis at the germinal matrix, could be postnatally retained subplate cells (Chun and Shatz, 1989; Marin-Padilla, 1998; Kuzniecky and Barkovich, 2001). If correct, our hypothesis would predict and possibly explain why cytomegalic neurons in CD tissue seem to be specific for primates and humans, and are difficult to find in rodent CD models with smaller less morphologically complex prenatal subplate zones.

Our hypothesis would also postulate that balloon cells could be postnatally retained transitional radial glia cells that failed to disappear or completely transform into astrocytes upon completion of human cortical development. Radial glia have neurogenetic capacity, their transformation to glia are regulated by Cajal-Retzius cells, and they immunostain for both glial and neuronal markers, much like balloon cells in CD brain tissue (Super et al., 2000; Noctor et al., 2001, 2002; Fishell and Kriegstein, 2003). In our prior in vitro studies, balloon cells in CD patients showed mature dendritic morphologies but lack significant dendritic spines, and electrophysiologically are more glial than neuronal in character similar to radial glia recorded in prenatal animal studies (Lo Turco and Kriegstein, 1991) (see Cepeda et al., 2003, figs 2 and 6A–B). Thus, balloon cells appeared well differentiated and ‘mature’ even in surgical specimens operated upon in the first year of life, and have morphologies similar to transitional radial glia. Finally, our data also found increased neuronal densities in the upper gray matter, layer 1, and white matter, which would be the expected location of excess pyramidal neurons if there were increased neurogenesis or changes in progenitor cell cycle dynamics at subventricular or subcortical sites late in corticoneurogenesis (Caviness et al., 1995; Rakic, 1995; Takahashi et al., 1999; Chenn and Walsh, 2002).

Our MRI and morphometric data also support the concept that other features of postnatal cerebral development were not affected in CD and non-CD cases with epilepsy. For example, we found the expected logarithmic increased postnatal white matter volumes, decreased gray and white matter neuronal densities, and increased neuronal sizes that were previously reported using MRI and/or histological techniques for normal human postnatal cortical development (Table 4, Fig. 12) (Conel, 1939, 1941, 1947, 1951, 1955, 1959, 1963, 1967; Dobbing and Sands, 1973; Huttenlocher, 1990; Hayakawa et al., 1991; Pfefferbaum et al., 1994; Rabinowicz et al., 1996; Giedd et al., 1999; Paus et al., 2001). In epilepsy surgery patients, these findings indicate that cortical plate and neuronal maturation continues postnatally, and supports the notion that only components of corticoneurogenesis were incomplete or augmented in CD pathogenesis (Sankar et al., 1995). Such concepts also suggest hypotheses concerning epileptogenesis in CD tissue. Seizure generation in CD brain regions could be the consequence of retained preplate cells interacting with excess generated pyramidal neurons. In animal studies, prenatal layer 1 and subplate cells are known to be physiologically active, and the ontogeny of cortical signaling has pro-epileptic components, such as depolarizing GABAergic currents, robust NMDA currents, and spontaneously fast-spiking Cajal-Retzius cells (Friauf et al., 1990; Dammerman et al., 2000; Owens and Kriegstein, 2002; Andre et al., 2004). While speculative, these concepts are supported by integrating our data with the existing basic and clinical literature, and provide direction for future hypothesis driven studies to determine possible cellular and synaptic elements that produce seizures in CD tissue.

Figure 12.

Line graphs illustrating neocortical changes as a function of age of data obtained from J. LeRoy Conel's autopsy series (published from 1939 to 1967) for the middle and inferior frontal gyrus, and our study, in which the samples were most often from similar cerebral locations. All of the measures show similar changes between the two studies despite the inclusion of data from cortical dysplasia and non-dysplasia in the current analysis. The most significant postnatal changes occurred from birth to 3 months of age, and changes were less incremental thereafter.

Figure 12.

Line graphs illustrating neocortical changes as a function of age of data obtained from J. LeRoy Conel's autopsy series (published from 1939 to 1967) for the middle and inferior frontal gyrus, and our study, in which the samples were most often from similar cerebral locations. All of the measures show similar changes between the two studies despite the inclusion of data from cortical dysplasia and non-dysplasia in the current analysis. The most significant postnatal changes occurred from birth to 3 months of age, and changes were less incremental thereafter.

Morphological Changes Related to Seizures and Infantile Spasms

Our study found neuronal hypertrophy related to seizures and/or a history of infantile spasms (Table 4). Neuronal hypertrophy has been previously described in Brodman's area 38 of temporal lobe epilepsy patients and in less damaged neocortical regions adjacent to the injured site in children with acquired encephalopathies and seizures (Marin-Padilla, 1999; Bothwell et al., 2001). Thus, the finding of neuronal hypertrophy associated with epilepsy is not novel, and may relate to increased neuronal functional activity from seizures (Lazeyras et al., 2000; Miller et al., 2002). A unique finding of our study was that layer 1 neurons showed hypertrophy associated with infantile spasms. In primates, layer 1 cells include the classical Cajal-Retzius cells, which are generated first, smaller GABAergic neurons produced later from the olfactory primordium and ganglionic eminence, and possibly neurons derived from the subpial granular layer (Zecevic and Rakic, 2001; Rakic and Zecevic, 2003). We do not know if the layer 1 cells that display hypertrophy in patients with infantile spasms were Cajal-Retzius, GABAergic neurons, or other cell types, nor do we know if these cells are the ‘cause’ or ‘consequence’ of infantile spasms. However, our finding supports the notion that early age-dependent seizures, like infantile spasms, may be associated with layer 1 neuronal abnormalities that could affect postnatal cortical and developmental functions.

It is important to consider the potential limitations related to clinical sampling sites, experimental design, and techniques when interpreting our results. For example, we purposefully excluded cases of polymicrogyria from this study, and our findings and hypotheses are most relevant to severe CD cases associated with cytomegalic neurons and balloon cells as originally described by Taylor (Taylor et al., 1971). Furthermore, our results were obtained from CD patients with therapy-resistant epilepsy, and may or may not be similar to CD patients without seizures or whose epilepsy is controlled with anti-epileptic drugs. In addition, MRI gray and white matter volumes depended on the ability of the software or operator to discriminate these structures based on T1- or T2-weighted images. As mentioned, in younger children this may be difficult, and we often had to rely on manually drawing the border based on a best visual assessment. While manual methods may be less reliable, there are currently no other validated MRI techniques for young CD and non-CD surgical patients. Likewise, our MRI studies did not include non-seizure age-matched controls because such a group was difficult to recruit since most children undergoing diagnostic neuroimaging have structural abnormalities. Our experimental design assumes that the opposite non-affected cerebral hemispheres were relatively ‘normal’. Such an assumption is clinically reasonable given that the patients had normal motor and other neurological functions from the preserved hemisphere, but we do not know with certainty without a non-epilepsy age-matched comparison group. In addition, we measured NeuN cell densities from a limited number of tissue blocks per patient obtained from different parts of the cerebrum, and cortical anatomy is known to vary by brain region and age (Conel, 1939, 1941, 1947, 1951, 1955, 1959, 1963, 1967). However, statistical analysis indicated that we sampled similar brain regions for all patient groups (P = 0.18), and despite the expected increased variability due to the location of the sample sites and age at collection we still found statistically significant differences between patient groups for NeuN cell densities and neuronal sizes. Finally, we counted NeuN-positive neurons. By using NeuN we could accurately count neurons and avoid potential confusion with glia or other non-neuronal cells. However, we cannot discern whether the cells were excitatory or inhibitory, and immature cells that did not express NeuN would be excluded. Thus, additional studies will be necessary to determine if changes in cell densities for CD patients were from glutamatergic and/or GABAergic cells, the location of these cell types and whether CD tissue also contains more immature neurons than non-CD tissue.

Conclusions

In epilepsy surgery patients, this study found changes in MRI cerebral gray and white matter volumes, NeuN neuronal densities and size that provide clues to the timing of severe CD pathogenesis. Based on our results, we propose that CD pathogenesis involves excessive neurogenesis of late generated neurons and possible retention of radial glia and subplate neurons giving the abnormal appearance of cytomegalic neurons and balloon cells. We also propose that seizure generation in CD may be the consequence of interactions between excess pyramidal neurons and retained preplate cells. Finally, our findings related to layer 1 neuronal hypertrophy suggests that infantile spasms may be associated with neuronal abnormalities in the molecular layer that could affect postnatal cerebral functions. These interpretations will require additional experimental validation in both human and animal models of CD and epilepsy, but begin to form a framework for proposing future hypothesis-driven experiments.

This work was supported by grants NIH RO1 NS38992 and PO5 NS02808.

References

Andre VM, Flores-Hernandez J, Cepeda C, Starling AJ, Nguyen ST, Lobo MK, Vinters HV, Levine MS, Mathern GW (
2004
) NMDA receptor alterations in neurons from pediatric cortical dysplasia tissue.
Cereb Cortex
  (in press).
Baraban SC, Wenzel HJ, Hochman DW, Schwartzkroin PA (
2000
) Characterization of heterotopic cell clusters in the hippocampus of rats exposed to methylazoxymethanol in utero.
Epilepsy Res
 
39
:
87
–102.
Barkovich AJ, Kuzniecky RI, Jackson GD, Guerrini R, Dobyns WB (
2001
) Classification system for malformations of cortical development: update 2001.
Neurology
 
57
:
2168
–2178.
Battaglia G, Arcelli P, Granata T, Selvaggio M, Andermann F, Dubeau F, Olivier A, Tampieri D, Villemure JG, Avoli M, Avanzini G, Spreafico R (
1996
) Neuronal migration disorders and epilepsy: a morphological analysis of three surgically treated patients.
Epilepsy Res
 
26
:
49
–58.
Bothwell S, Meredith GE, Phillips J, Staunton H, Doherty C, Grigorenko E, Glazier S, Deadwyler SA, O'Donovan CA, Farrell M (
2001
) Neuronal hypertrophy in the neocortex of patients with temporal lobe epilepsy.
J Neurosci
 
21
:
4789
–4800.
Caviness VS, Jr., Takahashi T, Nowakowski RS (
1995
) Numbers, time and neocortical neuronogenesis: a general developmental and evolutionary model.
Trends Neurosci
 
18
:
379
–383.
Cepeda C, Hurst RS, Flores-Hernandez J, Hernandez-Echeagaray E, Klapstein GJ, Boylan MK, Calvert CR, Jocoy EL, Nguyen OK, Andre VM, Vinters HV, Ariano MA, Levine MS, Mathern GW (
2003
) Morphological and electrophysiological characterization of abnormal cell types in pediatric cortical dysplasia.
J Neurosci Res
 
72
:
472
–486.
Chenn A, Walsh CA (
2002
) Regulation of cerebral cortical size by control of cell cycle exit in neural precursors.
Science
 
297
:
365
–369.
Chun JJ, Shatz CJ (
1989
) Interstitial cells of the adult neocortical white matter are the remnant of the early generated subplate neuron population.
J Comp Neurol
 
282
:
555
–569.
Colacitti C, Sancini G, DeBiasi S, Franceschetti S, Caputi A, Frassoni C, Cattabeni F, Avanzini G, Spreafico R, Di Luca M, Battaglia G (
1999
) Prenatal methylazoxymethanol treatment in rats produces brain abnormalities with morphological similarities to human developmental brain dysgeneses.
J Neuropathol Exp Neurol
 
58
:
92
–106.
Conel JL (
1939
) The postnatal development of the human cerebral cortex. Vol. 1. The cortex of the newborn. Cambridge, MA: Harvard University Press.
Conel JL (
1941
) The postnatal development of the human cerebral cortex. Vol. II. The cortex of the one-month infant. Cambridge, MA: Harvard University Press.
Conel JL (
1947
) The postnatal development of the human cerebral cortex. Vol. III. The cortex of the three-month infant. Cambridge, MA: Harvard University Press.
Conel JL (
1951
) The postnatal development of the human cerebral cortex. Vol. IV. The cortex of the six-month infant. Cambridge, MA: Harvard University Press.
Conel JL (
1955
) The postnatal development of the human cerebral cortex. Vol. V. The cortex of the fifteen-month infant. Cambridge, MA: Harvard University Press.
Conel JL (
1959
) The postnatal development of the human cerebral cortex. Vol. VI. The cortex of the twenty-four month infant. Cambridge, MA: Harvard University Press.
Conel JL (
1963
) The postnatal development of the human cerebral cortex. Vol. VII. The cortex of the four-year child. Cambridge, MA: Harvard University Press.
Conel JL (
1967
) The postnatal development of the human cerebral cortex. Vol. VIII. The cortex of the six-year child. Cambridge, MA: Harvard University Press.
Cook SW, Nguyen ST, Hu B, Yudovin S, Shields DW, Vinters H, Van de Wiele BM, Harrison RE, Mathern GW (
2004
) Cerebral hemispherectomy for pediatric epilepsy patients: a comparison of three techniques by pathologic substrate in 115 patients.
J Neurosurg (Pediatrics 2)
 
100
:
125
–141.
Dammerman RS, Flint AC, Noctor S, Kriegstein AR (
2000
) An excitatory GABAergic plexus in developing neocortical layer 1.
J Neurophysiol
 
84
:
428
–434.
Dobbing J, Sands J (
1973
) Quantitative growth and development of human brain.
Arch Dis Child
 
48
:
757
–767.
Duong T, De Rosa MJ, Poukens V, Vinters HV, Fisher RS (
1994
) Neuronal cytoskeletal abnormalities in human cerebral cortical dysplasia.
Acta Neuropathol (Berl)
 
87
:
493
–503.
Farrell MA, DeRosa MJ, Curran JG, Secor DL, Cornford ME, Comair YG, Peacock WJ, Shields WD, Vinters HV (
1992
) Neuropathologic findings in cortical resections (including hemispherectomies) performed for the treatment of intractable childhood epilepsy.
Acta Neuropathol (Berl)
 
83
:
246
–259.
Fishell G, Kriegstein AR (
2003
) Neurons from radial glia: the consequences of asymmetric inheritance.
Curr Opin Neurobiol
 
13
:
34
–41.
Friauf E, McConnell SK, Shatz CJ (
1990
) Functional synaptic circuits in the subplate during fetal and early postnatal development of cat visual cortex.
J Neurosci
 
10
:
2601
–2613.
Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A, Paus T, Evans AC, Rapoport JL (
1999
) Brain development during childhood and adolescence: a longitudinal MRI study.
Nat Neurosci
 
2
:
861
–863.
Hayakawa K, Konishi Y, Kuriyama M, Konishi K, Matsuda T (
1991
) Normal brain maturation in MRI.
Eur J Radiol
 
12
:
208
–215.
Huttenlocher PR (
1990
) Morphometric study of human cerebral cortex development.
Neuropsychologia
 
28
:
517
–527.
Kerfoot C, Vinters HV, Mathern GW (
1999
) Cerebral cortical dysplasia: giant neurons show potential for increased excitation and axonal plasticity.
Dev Neurosci
 
21
:
260
–270.
Kostovic I, Rakic P (
1980
) Cytology and time of origin of interstitial neurons in the white matter in infant and adult human and monkey telencephalon.
J Neurocytol
 
9
:
219
–242.
Kostovic I, Rakic P (
1990
) Developmental history of the transient subplate zone in the visual and somatosensory cortex of the macaque monkey and human brain.
J Comp Neurol
 
297
:
441
–470.
Kuzniecky RI, Barkovich AJ (
2001
) Malformations of cortical development and epilepsy.
Brain Dev
 
23
:
2
–11.
Lazeyras F, Blanke O, Zimine I, Delavelle J, Perrig SH, Seeck M (
2000
) MRI, (1)H-MRS, and functional MRI during and after prolonged nonconvulsive seizure activity.
Neurology
 
55
:
1677
–1682.
Lo Turco JJ, Kriegstein AR (
1991
) Clusters of coupled neuroblasts in embryonic neocortex.
Science
 
252
:
563
–566.
Mackay MT, Becker LE, Chuang SH, Otsubo H, Chuang NA, Rutka J, Ben-Zeev B, Snead OC 3rd, Weiss SK (
2003
) Malformations of cortical development with balloon cells: clinical and radiologic correlates.
Neurology
 
60
:
580
–587.
Marin-Padilla M (
1998
) Cajal-Retzius cells and the development of the neocortex.
Trends Neurosci
 
21
:
64
–71.
Marin-Padilla M (
1999
) Developmental neuropathology and impact of perinatal brain damage. III. Gray matter lesions of the neocortex.
J Neuropathol Exp Neurol
 
58
:
407
–429.
Marin-Padilla M, Marin-Padilla TM (
1982
) Origin, prenatal development and structural organization of layer I of the human cerebral (motor) cortex. A Golgi study.
Anat Embryol (Berl)
 
164
:
161
–206.
Mathern GW, Babb TL, Pretorius JK, Leite JP (
1995
) Reactive synaptogenesis and neuron densities for neuropeptide Y, somatostatin, and glutamate decarboxylase immunoreactivity in the epileptogenic human fascia dentata.
J Neurosci
 
15
:
3990
–4004.
Mathern GW, Kuhlman PA, Mendoza D, Pretorius JK (
1997
) Human fascia dentata anatomy and hippocampal neuron densities differ depending on the epileptic syndrome and age at first seizure.
J Neuropathol Exp Neurol
 
56
:
199
–212.
Mathern GW, Giza CC, Yudovin S, Vinters HV, Peacock WJ, Shewmon DA, Shields WD (
1999
) Postoperative seizure control and antiepileptic drug use in pediatric epilepsy surgery patients: the UCLA experience, 1986–1997.
Epilepsia
 
40
:
1740
–1749.
Mathern GW, Cepeda C, Hurst RS, Flores-Hernandez J, Mendoza D, Levine MS (
2000
) Neurons recorded from pediatric epilepsy surgery patients with cortical dysplasia.
Epilepsia
 
41
:
S162
–S167.
Miller SP, Weiss J, Barnwell A, Ferriero DM, Latal-Hajnal B, Ferrer-Rogers A, Newton N, Partridge JC, Glidden DV, Vigneron DB, Barkovich AJ (
2002
) Seizure-associated brain injury in term newborns with perinatal asphyxia.
Neurology
 
58
:
542
–548.
Mischel PS, Nguyen LP, Vinters HV (
1995
) Cerebral cortical dysplasia associated with pediatric epilepsy. Review of neuropathologic features and proposal for a grading system.
J Neuropathol Exp Neurol
 
54
:
137
–153.
Miyata H, Chute DJ, Fink J, Villablanca P, Vinters HV (
2004
) Lissencephaly with agenesis of corpus callosum and rudimentary dysplastic cerebellum: a subtype of lissencephaly with cerebellar hypoplasia.
Acta Neuropathol (Berl)
 
107
:
69
–81.
Mrzljak L, Uylings HB, Kostovic I, Van Eden CG (
1988
) Prenatal development of neurons in the human prefrontal cortex. I. A qualitative Golgi study.
J Comp Neurol
 
271
:
355
–386.
Mrzljak L, Uylings HB, Kostovic I, van Eden CG (
1992
) Prenatal development of neurons in the human prefrontal cortex. II. A quantitative Golgi study.
J Comp Neurol
 
316
:
485
–496.
Najm I, Ying Z, Babb TL, Crino PB, Macdonald R, Mathern GW, Spreafico R (
2004
) Mechanisms of epileptogenicity in cortical dysplasias.
Neurology
 
62
(Suppl 3):
S9
–S13.
Noctor SC, Flint AC, Weissman TA, Dammerman RS, Kriegstein AR (
2001
) Neurons derived from radial glial cells establish radial units in neocortex.
Nature
 
409
:
714
–720.
Noctor SC, Flint AC, Weissman TA, Wong WS, Clinton BK, Kriegstein AR (
2002
) Dividing precursor cells of the embryonic cortical ventricular zone have morphological and molecular characteristics of radial glia.
J Neurosci
 
22
:
3161
–3173.
Okajima S, Takeshita K, Antoku S, Shiomi T, Russell WJ, Fujita S, Yoshinaga H, Neriishi S, Kawamoto S, Norimura T (
1978
) Radioactive fallout effects of the Nagasaki atomic bomb.
Health Phys
 
34
:
621
–633.
Owens DF, Kriegstein AR (
2002
) Is there more to GABA than synaptic inhibition?
Nat Rev Neurosci
 
3
:
715
–727.
Paus T, Collins DL, Evans AC, Leonard G, Pike B, Zijdenbos A (
2001
) Maturation of white matter in the human brain: a review of magnetic resonance studies.
Brain Res Bull
 
54
:
255
–266.
Pfefferbaum A, Mathalon DH, Sullivan EV, Rawles JM, Zipursky RB, Lim KO (
1994
) A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood.
Arch Neurol
 
51
:
874
–887.
Rabinowicz T, de Courten-Myers GM, Petetot JM, Xi G, de los Reyes E (
1996
) Human cortex development: estimates of neuronal numbers indicate major loss late during gestation.
J Neuropathol Exp Neurol
 
55
:
320
–328.
Rakic P (
1988
) Defects of neuronal migration and the pathogenesis of cortical malformations.
Prog Brain Res
 
73
:
15
–37.
Rakic P (
1995
) A small step for the cell, a giant leap for mankind: a hypothesis of neocortical expansion during evolution.
Trends Neurosci
 
18
:
383
–388.
Rakic S, Zecevic N (
2003
) Emerging complexity of layer I in human cerebral cortex.
Cereb Cortex
 
13
:
1072
–1083.
Raymond AA, Fish DR, Sisodiya SM, Alsanjari N, Stevens JM, Shorvon SD (
1995
) Abnormalities of gyration, heterotopias, tuberous sclerosis, focal cortical dysplasia, microdysgenesis, dysembryoplastic neuroepithelial tumour and dysgenesis of the archicortex in epilepsy. Clinical, EEG and neuroimaging features in 100 adult patients.
Brain
 
118
:
629
–660.
Roper SN (
1998
) In utero irradiation of rats as a model of human cerebrocortical dysgenesis: a review.
Epilepsy Res
 
32
:
63
–74.
Sankar R, Curran JG, Kevill JW, Rintahaka PJ, Shewmon DA, Vinters HV (
1995
) Microscopic cortical dysplasia in infantile spasms: evolution of white matter abnormalities.
AJNR Am J Neuroradiol
 
16
:
1265
–1272.
Super H, Del Rio JA, Martinez A, Perez-Sust P, Soriano E (
2000
) Disruption of neuronal migration and radial glia in the developing cerebral cortex following ablation of Cajal-Retzius cells.
Cereb Cortex
 
10
:
602
–613.
Takahashi T, Goto T, Miyama S, Nowakowski RS, Caviness VS, Jr. (
1999
) Sequence of neuron origin and neocortical laminar fate: relation to cell cycle of origin in the developing murine cerebral wall.
J Neurosci
 
19
:
10357
–10371.
Tassi L, Colombo N, Garbelli R, Francione S, Lo Russo G, Mai R, Cardinale F, Cossu M, Ferrario A, Galli C, Bramerio M, Citterio A, Spreafico R (
2002
) Focal cortical dysplasia: neuropathological subtypes, EEG, neuroimaging and surgical outcome.
Brain
 
125
:
1719
–1732.
Taylor DC, Falconer MA, Bruton CJ, Corsellis JA (
1971
) Focal dysplasia of the cerebral cortex in epilepsy.
J Neurol Neurosurg Psychiatry
 
34
:
369
–387.
Uher BF, Golden JA (
2000
) Neuronal migration defects of the cerebral cortex: a destination debacle.
Clin Genet
 
58
:
16
–24.
Vinters HV, Fisher RS, Cornford ME, Mah V, Secor DL, De Rosa MJ, Comair YG, Peacock WJ, Shields WD (
1992
) Morphological substrates of infantile spasms: studies based on surgically resected cerebral tissue.
Childs Nerv Syst
 
8
:
8
–17.
Volpe JJ (
2000
) Overview: normal and abnormal human brain development.
Ment Retard Dev Disabil Res Rev
 
6
:
1
–5.
Zecevic N, Rakic P (
2001
) Development of layer I neurons in the primate cerebral cortex.
J Neurosci
 
21
:
5607
–5619.
Zecevic N, Milosevic A, Rakic S, Marin-Padilla M (
1999
) Early development and composition of the human primordial plexiform layer: An immunohistochemical study.
J Comp Neurol
 
412
:
241
–254.

Author notes

1Division of Neurosurgery, University of California, Los Angeles, CA 90005, USA, 2Division of Neuroradiology, University of California, Los Angeles, CA 90005, USA, 3Division of Neuropathology, University of California, Los Angeles, CA 90005, USA, 4Department of Neurology, University of California, Los Angeles, CA 90005, USA, 5The Brain Research Institute, University of California, Los Angeles, CA 90005, USA, 6The Mental Retardation Research Center, David Geffen School of Medicine, University of California, Los Angeles, CA 90005, USA, 7Department of Pathology, Ribeirão Preto School of Medicine, University of São Paulo, Ribeirã Preto, SP, 14049-900, Brazil and 8Department of Neurology, Ribeirão Preto School of Medicine, University of São Paulo, Ribeirão Preto, SP, 14049-900, Brazil