1. Top of page
  2. Abstract

Autism is a brain disorder characterized by abnormalities in how a person relates and communicates to others. Both post-mortem and neuroimaging studies indicate the presence of increased brain volume and, in some cases, an altered gray/white matter ratio. Contrary to established gross findings there is no recognized microscopic pathology to autism. Early studies provided multiple leads none of which have been validated. Clinicopathological associations have been difficult to sustain when considering possible variables such as use of medications, seizures, mental retardation and agonal/pre-agonal conditions. Research findings suggest widespread cortical abnormalities, lack of a vascular component and an intact blood–brain barrier. Many of the previously mentioned findings can be explained in terms of a mini-columnopathy. The significance of future controlled studies should be judged based on their explanatory powers; that is, how well do they relate to brain growth abnormalities and/or provide useful clinicopathological correlates.


  1. Top of page
  2. Abstract

Autism is a neurodevelopmental condition defined by operational criteria targeting several behavioral domains: social interaction, communication skills and patterns of activities (3). Symptoms are described as severe, pervasive and manifested during the first 3 years of life. Diagnosis is often made as early as 18 months of age but most patients are not formally diagnosed until 5 years (43). While not being part of an official classification scheme the term autism spectrum disorders (ASD) is used to encompass three conditions that share core symptoms: autism, Asperger disorder and Pervasive Developmental Disorders-not otherwise specified. Although once considered a rare disorder, a recent survey by the Centers for Disease Control and Prevention found that 1 in 150 children had an ASD within examined communities (23). Genetic epidemiological studies, including twin studies done in separate or conserved environments, strongly support the presence of heritable factors. Autism in monozygotic twins is stated to be 12 times higher than in the normal population. By way of comparison the rate for dizygotic twins is only four times higher than in the general population (51). However, the fact that many affected identical twins exhibit differences in severity and types of disabilities also suggests prenatal and postnatal influences (135).

Autism often occurs in the presence of other medical conditions such as seizures, mental retardation and chromosomal anomalies. Specific genotypes identified within the autistic spectrum include the tuberous sclerosis complex (TSC), mitochondrial disorders, fragile X, Down, Williams–Beuren, Angleman/Prader–Willi, velocardiofacial and Möbius syndromes (42). The presence or absence of any of these comorbidities does not provide an exclusionary criterion for diagnosis, rather, it bespeaks of etiologic heterogeneity and the possibility that autism originates early in the first trimester of gestation. Similarly, the term “spectrum” within ASDs recognizes a broad range of severity and the potential for many underlying etiologies. Given this scenario it is difficult to envision a single cause or treatment for the condition.

Optimal interventions should provide for the acquisition of skills that are generalized across settings and maintained over time. However, most claims of treatment effectiveness in autism are based on testimonials, small series of patients and clinical trials of short-term duration (78, 87). Thus far, Applied Behavioral Analysis (ABA), the application of learning theory based on operant conditioning, is the only intervention recommended by the Surgeon General (1). Pharmacological interventions serve as adjunctive therapy but do not target core symptoms of the condition. Recently, the FDA approved the first drug, risperdal, for the treatment of maladaptive behaviors (eg, aggressions, tantrum and self-injury) associated with autism (39).

The dearth of beneficial therapeutic interventions in autism has been accompanied by an efflorescence of classifications and screening scales that lack in construct validity. Although the accuracy of some scales has been studied, there is no empirical way of validating them against a recognized biomarker. In other words, scales have been primarily generated keeping in mind the creation of a homogenous group of patients from which to compare clinical data across investigators and research centers. The results, although providing for reliability and internal consistency, lack an empirical framework to provide meaningful inferences. Neuropathological studies are urgently needed to facilitate the analysis of clinicopathological associations, confirm diagnosis, assess risk factors and provide improved characterization of environmental influences and comorbid conditions.

The Children's Health Act of 2000 (24) mandated the establishment of an Interagency Coordinating Committee to manage autism research within the Department of Health and Human Services. This authority was delegated to the National Institutes of Health who created a committee to facilitate the exchange of information among member agencies and to coordinate autism-related research initiatives. A scientific panel then prioritized goals to promote autism research and placed them into a matrix (37). It was indicated that a possible roadblock toward advancing our understanding of this condition was the shortage of brains for a national registry (Autism Tissue Program) used in post-mortem studies. Neuropathological studies were characterized as medium risk research to be undertaken in the 4–6 years following the matrix implementation. The panel promoted within a shorter timeframe development of technology and infrastructure “for multi-site in vivo imaging studies, to identify the neuropathology of autism”(37).


  1. Top of page
  2. Abstract

Genome-wide linkage screens have thoroughly refuted a monogenetic mode of inheritance for autism. It may be that the case series studied thus far have been limited by including autistic individuals that vary widely in terms of neurological and social skill impairments. Finding susceptibility genes may therefore depend on creating dimensional measures that span a broader phenotype of autism (32). Identified domains of this broader phenotype include: face processing, sensitivity to social reward, imitation of body actions, memory, executive function and phonology (32). Linkage studies using more homogenous populations based on selected biological and behavioral endophenotypes are presently being pursued. The idea behind these studies is that groups exhibiting homogenous social subskills may be endophenotypes of the autism phenotype. In the meanwhile, current clinical consensus regards autism as a multifactorial trait as opposed to an oligogenic or polygenic disorder. Studies suggest the presence of multiple susceptibility and protective genes that modify the risk of developing the condition. As in other multifactorial conditions autism appears to involve primarily one organ, the brain. Not surprisingly, the increased risk among first degree relatives falls into the 3%–5% range proposed for other multifactorial conditions (133). This recurrence risk is lower than for single gene inheritance.

Multifactorial disorders offer a threshold phenomenon where manifestations supervene when three factors impinge to various degrees upon a particular infant (a triple hit hypothesis): (i) a critical period of brain development, (ii) an underlying vulnerability, and (iii) exogenous stressor(s). Environmental influences play a major role in multifactorial conditions when symptoms manifest themselves during adult life, for example cancer. However, they can still play a significant role for similar disorders manifesting at birth. Thus, it is well known that exposure to cigarette smoke, alcohol, or illicit drugs early in pregnancy increase the risk for developing a cleft lip. The following paragraphs detail different aspects of this triple hit hypothesis for autism. Although the exact mechanisms of interaction remain unknown we feel confident there is enough evidence to discuss salient aspects related to comorbidities, exogenous factors and a time window of vulnerability in autism.

Recent research has paid a great deal of attention at the possible link between TSC and autism. Estimated rates of autism in patients with TSC range from 17% to 68% (115). Contrariwise, the number of autistic patients having TSC has been estimated between 0.4% and 4% (42). The intracranial pathology of TSC is characterized by malformative, hamartomatous and neoplastic lesions that include cortical/subcortical tubers and subependymal nodules. Tubers are characterized by architectural disarray indicative of cortical dysplasia and a cellular profile suggesting a failure of commitment in neuroglial differentiation. When tubers are numerous and located in the frontotemporal regions of the brain the risk of developing an ASD increases (47).

It has been noted that in utero exposure to valproic acid and other anticonvulsants increase the risk for manifesting autism or autistic-like traits postnatally (79, 134). Four of the 57 patients reported by Moore et al were diagnosed with autism; two with valproic acid alone and one with both valproic acid and phenytoin (79). Some authors have suggested that the risk for fetal valproate syndrome is dosage dependent (109). This fact has been used to create an animal model (rat) of autism. Prenatal exposure to valproic acid on the 12.5th day of gestation reproduces some of the clinical (eg, decreased number of social behaviors) and pathological features (eg, diminished number of Purkinje cells) putatively observed in autism (62, 113).

Another strong association to neuroembryological dysfunction comes from observations that approximately 5% of individuals exposed in utero to thalidomide develop autism (119, 120). What appears interesting of this observation is that thalidomide patients who manifest autism also exhibit external ear abnormalities and an uncommon form of strabismus (Duane syndrome) but no malformations of their arms or legs (phocomelia). Timing of these “minor” malformations and the supposition that autism may have arisen during the same stage of development suggests a time window of vulnerability early in gestation (20–24 days) (107).

The first trimester is the time of development for multiple congenital anomalies that include autism as one of their manifestations. Among these congenital disorders is Möbius syndrome (facial diplegia). The genesis of Möbius syndrome is disputed. Theories on causation include genetics, vascular deficits, maternal trauma and the use of certain drugs during pregnancy, for example ergotamine, misoprostol (a drug used for self-induced and planned abortions) and thalidomide. Case series reveal aplasia/hypoplasia of various cranial nerve nuclei, focal necrosis and calcification of brainstem tissue (123). The heterogeneous histopathology and varied clinical expression suggests that, in terms of the “behavioral” phenotype, timing of the underlying insult may be of equal, if not greater importance, than the anatomical loci affected. In other words when considering the developing brain and its orchestrated sequential maturation of neurons, synapses and cortical maps “when is as important as what”(11).

Sparse clinical data support a vulnerability period for autism during the second and third trimester of gestation (28). Bleeding and maternal infections during the second trimester have been associated with increased risk for developing autism (26, 63, 122). Case reports of maternal infections (eg, cytomegalovirus) during the perinatal period offer less convincing proof in favor of the third trimester of gestation as a critical period of development (137). The disparity of epidemiological profiles regarding a time frame of environmental susceptibility indicates our incomplete understanding of the role of environmental co-variates within the cascade of events leading to autism.


  1. Top of page
  2. Abstract

Although autism is a heterogenous disorder with respect to etiology, there are likely to be shared pathophysiological mechanisms among individuals with autism. Throughout the literature a number of abnormalities of brain structure have been reported; however, increased brain size appears to be the most consistent morphometric observations reported in autism. In general, increased brain size occurs without obliteration of the subarachnoid space, flattened gyri, or reduced ventricular space. The brains of autistic individuals show preservation of the demarcation between gray and white matter. Microscopically, there is no cytoplasmic vacuolation or enlargement of perivascular spaces. Many of the reported cases of increased brain volume stem from series of patient within the neuroimaging literature. Increased brain size is therefore not the result of edema or post-mortem artifacts, and should be considered a core deficit and integral part of the neuropathology of autism.

Increased brain size has been reported in first degree relatives of autistic patients and in ASDs (41, 136). Brain enlargement appears generalized and may or may not include the cerebellum (31, 116). The initial findings of hypoplasia and hyperplasia of cerebellar vermian lobules in different subgroups of autistic patients remains controversial (29, 30). They have not replicated by many groups and are not specific; being found in fragile X syndrome (110). No differences in the vermian lobules were found in a study by Piven et al when adjustment was made for midsagittal brain area and IQ (95).

Morphometric evidence from neuroimaging studies suggests that the white matter is disproportionably larger to both absolute and proportional brain volume (57). Although not presently corroborated, parcellating the white matter into deeper and more superficial tracts showed that volume increase is confined to the radiate zone (arcuate fibers) (58). In this study the frontal lobes showed the greatest enlargement as compared with controls. This finding has also been reported by Carper et al using an independent sample (14). Casanova (15) suggested that the additional white matter is the result of short range association fibers required by an increased number of cortical mini-columns.

On a post-mortem series of 19 cases reported by Kemper and Bauman (65), eight of 11 subjects under 12 years of age had increased brain weight as compared with control. In the same series, six of eight brains from individuals aged 18 years and over showed reduced brain weights. Cross-sectional MRI studies in autism suggest that brain volume increases more than normal during early childhood. This rate of growth decelerates later on so that by late childhood or adolescence brain volumes of autistic and control samples are similar (6, 31, 53).

The finding of macrocephaly in autism persists after controlling for height, gender, the presence of epilepsy and other medical disorders (44, 96). The relationship between megalencephaly and performance IQ (high or low IQ autistics) is less clear (69, 96, 118). Goldberg et al has reviewed the neuroimaging literature (49). According to these researchers, most findings lack from replication and/or control for confounding variables. They concluded that enlarged brain size (megaloencephaly), particularly in the temporoparietal brain region, and decreased size of the posterior corpus callosum are the only independently replicated findings (49).

At gross inspection the course and pattern of gyri in the brains of autistic patients appears normal. There is one report of increased gyrification index in the frontal lobes of autistic patients (54). This study matched for anatomical location a single slice through the frontal lobes. Previous comparative anatomical studies have required measuring a minimum of 40 hemispheric slices before a measure of gyrification could be considered valid (5). Conclusions stemming from this study should be regarded as exploratory in nature. Alternatively, Levitt et al has used registered images to study cortical displacement maps of sulcal patterns in autism (70). The study corrected for total brain volume and gender. The results indicated subtle anterior and/or superior displacement of corresponding vectors for the superior and inferior frontal sulci and the superior temporal sulcus. Recently, our group found a reduction in the gyral window of patients with autism (22). The gyral window is the aperture of passage for projection fibers to and from the cortex (97). Measurements of the gyral window were directly correlated with the size (area of midsagittal section) of the corpus callosum. The results suggest that in autism a reduction in the gyral window constrains the possible size of projection fibers and offer a bias in connectivity emphasizing shorter association fibers at the expense of longer corticortical connections.

Several studies have reported the presence of nonspecific antenatal disruptive lesions in autism. These have provided for interrelated pathologies, for example schizencephaly, polymicrogyria, heterotopias (7, 45, 65, 75, 94, 106, 117). Schizencephaly is presently regarded as a destructive lesion occurring before 28 weeks as areas of polymicrogyric cortex are often found in the walls of the cleft and elsewhere in the brain. An even earlier onset is suggested when it is associated with neuronal heterotopia, indicating disruption of neuronal migration. Rather than disturbed morphogenesis the comorbid pathology indicates the influence of early onset encephaloclastic lesions. The overall significance of these lesions remains uncertain. The large majority of autistic patients do not have smaller brains, a concomitant to many encephaloclastic lesions.


  1. Top of page
  2. Abstract

Bauman and Kemper surveyed whole brain celloidin embedded serial sections that had been stained with Nissl. Sections were screened with a two headed stereomicroscope where corresponding anatomical levels were examined side by side in the same field of view at the same magnification. There were no abnormalities in neuronal morphometry, lamination or cellular density within examined areas of the isocortex. Alterations, when present, were described for the limbic system (ie, hippocampus, subiculum, amygdala, entorhinal cortex, mammillary bodies and septal nuclei) and the cerebellum (9, 10). In comparison to controls, autistic individuals showed reduced neuronal size and increased cell-packing density in these areas. Purkinje and granule cells were reduced in numbers throughout the cerebellar hemispheres without evidence of reactive gliosis. Furthermore, the olivary nuclei failed to manifest any atrophy as expected with Purkinje cell loss. Neurons within the emboliform, fastigeal and globose nuclei as well as those in the inferior olivary nuclei exhibited abnormalities that varied according to the age of the patients. In the older autistic patients cells were small and pale while in the younger patients they were enlarged and present in adequate numbers. Four of six patients reported (10) suffered from seizures but the reported pathology was similar regardless of comorbidity. Bauman and Kemper believed that these features were characteristic of a curtailment of normal development. The same group later examined Golgi impregnated hippocampal sections of two autistic subjects (a 7-year-old girl and a 9-year-old boy, both with mental retardation, but without epilepsy) and two control subjects (8 and 13 years old) (103). Only one autistic case showed adequate impregnation (and sufficient lack of artifacts) for analysis. This patient showed smaller neurons in the CA4 field. These early reports on cell size and numbers were based on qualitative observations and relied on biased (non stereological) assumptions.

Small neurons may represent a “developmental phenotype” but also a stage within a metabolic and morphological sequence leading to cell death, aposklesis (cell withering associated with neurodegeneration), or a type of non-apoptotic dark degenerating cell. The neurodevelopmental nature of the underlying cell changes has not been further pursued. There are no immunocytochemical studies with immature (MAP 1B, MAP 2C) MAP antibodies. Similarly, there are no studies staining for the embryonal form of the cell adhesion molecule N-CAM and developmental neurofilament nestin and internexin. Positive results from these studies would support the hypothesis of maturational failure in autism.

In autism cerebellar folial pattern appears normal. Loss of Purkinje and granule cells suggest a putative role for the cerebellum in autism (9, 10, 103). There is no evidence of disorganization (heterotaxia) of remaining cellular elements. This is the case even in the patches where cell loss has been claimed. A hypoxic etiology may account for these findings, that is, Purkinje cell loss may be the result of seizures (recognized or unrecognized), anticonvulsant medications (eg, Dilantin), or pre-agonal conditions. Marked reactivity of Bergmann's glia cells along with prominent microglial reaction in areas of Purkinje cell loss underlines the fact that these changes are acquired rather than neurodevelopmental in nature (7, 89, 125). Furthermore, if cerebellar damage occurs early during development it is accompanied by secondary changes of the inferior olivary nuclei (67). However, in autism the inferior olives may be thickened (7) but not poorly folded or fragmented. The olivary nuclei are found in their ordinary anatomical location and bear the crenated configuration produced by the normal migration of neurons from the rhombic lip into the olive.

Bailey et al investigated six autistic cases (all mentally handicapped and three with epilepsy) and seven age- and sex-matched controls (7). Four of the six cases were megalencephalic. In one cases there was increased cell packing density in all CA subfields of the hippocampus. Purkinje cell loss was occasionally accompanied by gliosis. Four cases showed areas of cortical abnormalities and the authors concluded that the findings pointed toward possible involvement of the cerebral cortex in autism. The cortical abnormalities were different depending on case but tended to involve primarily the frontal lobes. These abnormalities (suggestive of cortical dysgenesis to the authors) included a slightly irregular laminar pattern, thickened cortex and increased neuronal density. Occasional patients (not necessarily with the “cortical dysgenesis”) showed patches of ectopic gray matter and/or increased numbers of single neurons within the white matter.

Scattered post-mortem and radiological data in autism support the presence of heterotopias and their Magnetic Resonance Imaging correlates: unidentified bright objects (7, 86). It is difficult to surmise their relevance as they have not been related to nodular formations or other cerebral malformations. Heterotopias in modest numbers and just beneath the cortex are a normal occurrence (55). However, the presence of heterotopias in autism may provide a link to both seizures (microdysgenesis) and tuberous sclerosis. Heterotopias can result from an early fault in the migratory process or a lesion within the germinal zone (55). Of some importance is the fact that brainstem development transpires in parallel to the patterning of the neocortical protomap within the germinal zone (25). Generalized lesions that affect brainstem formation can similarly alter the interwoven cascade of intrinsic and extrinsic cues that define cortical development. This anatomical coincidence may help explain why some patients with Möbius syndrome exhibit symptoms characteristic of autism (see above).


  1. Top of page
  2. Abstract

Neurologists traditionally consider autism as a disease of the cortex. The tentative localization is supported by evidence of seizures in a significant proportion of cases and the absence of either spasticity or vision loss. There is little credible evidence that other organs besides the brain are involved. Similarly, there is no indication of peripheral nervous system involvement. Clinically, the dysfunction of higher cognitive functions further pinpoints the putative deficit to the isocortex. More specifically, autism appears to arise from a defect in the modular organization of the cortex that provides for the emergence of cognitive properties.

Modular organization of the cortex. Spatiotemporal regularities in the environment are reflected in the geometry of neural connections in which output at any position must be integrated into larger patterns of activity throughout the cortical network. Neurons are connected via widely convergent inputs and divergent outputs. A cortical pyramidal cell may typically maintain on the order of 1000 synaptic connections. In the dorsal cortex of reptiles, these connections are ordered within a relatively undifferentiated lattice in which subpallial projections pass tangentially through fields of radially oriented dendrites of several poorly differentiated layers (83, 84). With phylogenetic expansion of the pallial sheet, the number and length of processes must necessarily grow geometrically relative to growth in cell number in order to maintain a constant degree of connectedness among neurons. In living brains, expansion is constrained both by space available for white matter projections within the skull as well as nonlinear increases in metabolic demands as the length of these projections increases. Networks in which each neuron maintains synaptic links only with its immediate neighbors require shorter and fewer projections for each neuron. Yet total path length among all connected nodes is not reduced. Moreover, the number of synapses traversed between any two neurons in the network increases with attendant time delay, signal degradation and metabolic cost imposed at each synaptic relay. Selective pressure gives rise to a “small world” network (27, 130) in which neurons maintain short connection lengths within clusters which in turn are linked by longer-range projections. Emergence of this topology optimizes connectedness while minimizing wiring costs within the network (17).

This hierarchical clustering of cell connections is a defining feature of the mammalian brain (8). In the developing isocortex, the most prominent of these cell motifs is the ontogenetic cell column (68, 101), a radially oriented linear array of pyramidal neurons which extends through multiple layers of the cortical plate. Proliferative cells of the germinal zones surrounding the ventricular lumen give rise to radial glial cells (RGCs) whose processes span the width of the cortical mantle (112). RGCs serve as neural progenitors which generate a succession of postmitotic neurons (56, 73, 74, 81, 82, 121, 132). These daughter neurons migrate in tandem along the radial process of their parent RGC outward toward the pial surface. Each neuron in the sequence migrates past earlier-generated neurons to detach from the process at successively more superficial positions. Proliferative cells are thus programmed with the potential to give rise to parent RGCs whose processes guide their daughter neurons in the formation of radially oriented cell columns in the cortical plate (77). This mechanism was likely key in the rapid phylogenetic expansion of primate cortex, whose surface area is up to three orders of magnitude larger than and two to three times as thick as that of rodents. In primates, early established RGC-dedicated lineages maintain glial fibrillary acidic protein (GFAP)-positive glial processes through an extended period of neurogenesis/migration and preserve the columnar arrangement of neurons migrating across the thicker and more convoluted primate cortical sheet (100, 112, 138). The radial apposition of neurons in the ontogenetic cell column hypothetically should allow for formation of compact, economically organized circuits in radial arrays according to the “small worlds” scheme and more efficient Hebbian reciprocal circuit formation among its neurons.

After 30 weeks gestational age in humans, the pervasive columnar organization of the cortical plate is obscured to varying degrees by the migration of glia and interneurons and the growth of dendritic and axonal branches, in particular horizontal collateral connections. However, the underlying radial organization of these pyramidal cell columns remains intact (Figure 1). Imaging studies of a temporal sequence of post-mortem cortical tissue have demonstrated continuity of columnar morphometry during fetal and postnatal development and throughout the lifespan (22). The maturation of radial circuits within the pyramidal cell column core and their synaptic connections with inhibitory interneurons in the column's peripheral neuropil provide the basis for emergence of the mini-column, a fundamental microstructural motif of isocortex. In the mini-column, as first conceived by Lorente de Nó(33), representative cellular elements combine to form stereotypical microcircuits which subserve canonical operations on thalamocortical and corticocortical inputs. The basic minicolumnar plan of microcircuits organized around a core pyramidal cell column has been observed in representative cortical areas across a wide range of mammalian species (12, 38). Varying with area and species, each mini-column contains on the order of 80–100 pyramidal neurons radially aligned through layers II–VI. Center-to-center spacing between mini-columns ranges from 30 and 80 µm (34). Associated interneurons comprising 15% to 25% of the total neuron compliment modulate activity throughout the mini-column with radially oriented interneurons situated in the peripheral neuropil providing a “curtain” of inhibition around the mini-column core (34, 36).


Figure 1. Radial organization of the human cerebral cortex as seen in 35-µm sections, stained with cresyl violet. The linear appearance of the ontogenetic cell column is obscured late in gestation (left), although some vertical, striped texture is still evident. Visible minicolumnar structure re-emerges as neuropil in the periphery of columns expands, and persists throughout life (right). Micrographs are of human temporal lobe at 32 weeks of gestation and 50 years of age, respectively. Scale bars measure 100 µm (left) and 300 µm (right), respectively.

Download figure to PowerPoint

Four principal cellular features have been studied to assess minicolumnar morphometry and morphology: the core pyramidal cell column, the apical dendritic and vertical myelinated axon bundles arising from the pyramidal cells, and radially oriented translaminar axon bundles of double-bouquet inhibitory cells situated in peripheral neuropil (35, 36, 40, 88, 90–92, 114, 126, 128). Measures of horizontal spacing between apical dendritic bundles (52.6 µm) and between myelinated axon bundles (50.1 µm) in rat visual cortex are comparable (71). Similar horizontal spacing correspondences have been documented between apical dendritic bundles and layer III and V pyramidal cells of monkey visual cortex (91) and between double-bouquet axon bundles and pyramidal cell columns in monkey visual and human temporal cortex (36, 92). Morphometric linkages between these four elements suggest that they provide complimentary information from which global parameters of minicolumnar morphometry can be derived (22).

Minicolumnar pathology in autism. Several neuropathological features have been identified in the cortex of autistic individuals. These include neuroinflammatory and cellular changes (60, 93), and regional differences in size, morphology, number, density, and distribution of neurons and glia. Some such findings are drawn from small-scale studies or case reports, and lack unbiased methodology or confirmatory data. To date no pathological entity has been conclusively and systematically identified with autism at the cellular level. Recent post-mortem studies have shown area-specific changes in microstructural motifs in the isocortex of autistic individuals (18–21). These motifs, termed mini-columns, consist of radially oriented pyramidal cell arrays spanning the cortical width, their aligned myelinated axon and apical dendritic bundles, and several species of inhibitory interneurons, most prominently situated in the peripheral neuropil surrounding the column's core. Mini-columns are hypothesized to be an elemental functional unit of neocortex (12, 80). Changes to the mini-column may relate disparate genetic, molecular and anatomical lesions of autism within a common functional framework. Altered minicolumnar circuits may also be related to changes in transcortical and callosal white matter pathways linking together regional networks of mini-columns.

Casanova et al applied a semi-automated method (13) to analyze changes of minicolumnar morphometry in a tissue section series from nine patients with autistism and nine controls (18). Photomicrographs were taken of layer III in area 9 in prefrontal cortex and area 21 and posterior area 22 in superior temporal cortex. The images were decomposed into segments of Nissl-stained cell clusters according to a Gaussian distribution. A line was fitted through the segment centers by the least squares method to determine the minicolumnar radial axis from which minicolumnar and core width, cell dispersal and intercellular distance values were derived. In brains of autistic individuals, minicolumnar width was found to be significantly narrower (P = 0.034) with most of that decrease attributable to reduction of peripheral neuropil space. Concomitantly, the number of mini-columns per image area was increased. In addition, a dimensionless index of cell dispersal was increased suggesting that pyramidal cell alignment along the cell column radial axis was disturbed. The same series was also analyzed (19) according to the Gray Level Index (GLI) algorithm (Figure 2), modified from a method developed by Schleicher et al (111). The GLI represents the ratio of image area covered by Nissl-stained cell bodies to the area covered by unstained elements. The modified GLI method assesses differences in gray level amplitude and can therefore be used to determine column width as well as overall “cellularity” within each column. The original finding of decreased minicolumnar width was validated by the GLI method. Additionally, GLI amplitude across the interval between column centers was increased in autistic brains indicating a more highly defined circumferential boundary between pyramidal cell core and peripheral neuropil. Arguably, this finding may reflect decreases in collateral branching of core pyramidal cells. No significant difference in GLI was evident in brains of patients with autism compared with controls.


Figure 2. The method of Schleicher et al (111) estimates the local Gray Level Index (GLI), or proportion of Nissl-stained area. A micrograph (bottom, in false color) is segmented into two classes such that stained area is labeled 1 and unstained area (ie, background or neuropil) is labeled 0. The GLI (top) is computed by smoothing the segmented image using a kernel with oblong shape and a long axis oriented parallel to the mini-columns. The end result is a set of summary statistics including the distance between the ridges (light blue) seen in the GLI profile, their width and their height.

Download figure to PowerPoint

We conducted a recent study (16), using limited modifications of the cell fragmentation and GLI methods in an independent sample population of autistic patients and matched controls. We analyzed representative primary sensory, motor and prefrontal association cortices (areas 17, S1, 4, 9). The results validated our previous findings obtained with the same methods (18, 19). Differences in tissue preparation and section thickness from our earlier studies precluded a direct comparison of parameter values. Minicolumnar width, measured as tangential distance between cell column axes, was significantly narrower in our test samples compared with controls (P = 0.0234), a finding corroborated by GLI measures of distance between amplitude peaks. We also thresholded nucleolar area from cytoplasmic background and found nucleolar size to be smaller in our test sample. In this study, we applied a Boolean model, finding a greater density of Nissl-stained elements in each image frame (Figure 3). A Delaunay triangulation method (Figure 4) was applied to determine the distribution of distances between elements (20). By this method, a bimodal distribution of intercellular distance measures was generated, with each mode representing distribution of values for intercellular distances within and between core radial cell arrays (Figure 5). Intercolumnar distances were decreased with no significant difference noted in intracolumnar distances. These results indicated that minicolumnar width was decreased with increased numbers of mini-columns per cortical area, validating our previous findings. Comparable values for intracolumnar cell spacing and core width between groups suggested no difference in pyramidal cell number per mini-column. By implication most reduction in column width resulted from decreased volume of peripheral neuropil space. Taken together, these findings indicate that the size of cells within each column was reduced (Figure 6), a conclusion consistent with observed decreases in nucleolar size. These findings exhibited an area by diagnosis interaction with the greatest width reduction among sampled areas found in area 9 of the prefrontal cortex. A subsequent study systematically examined these parameters throughout primary sensory, paralimbic, and unimodal and heteromodal association areas, revealing a posteroanterior trend across neocortex toward decreased mini-column width. Significant diagnosis by area interactions were found in frontopolar (area 10) and the anterior cingulate gyrus (area 24) of paralimbic cortex (21).


Figure 3. The Boolean germ-grain model supposes that a spatial pattern, in this case the Nissl-stained area in a segmented micrograph, is the union of random closed sets (grains) drawn from some distribution and positioned at random points in space (the germ process). This model is a simplification in that it does not account for clustering of the grains, eg the minicolumnar structure. Under these assumptions, the mean grain area (Ā) and perimeter (Ū), along with the intensity (λ) of the germ process, completely determine the spatial pattern, providing for a model-based estimate of neuronal cross-section and density. These quantities are found to be reduced and increased, respectively, in autism. Left: cortical area 9, right hemisphere, lamina III from a 25-year-old man without autism. Ā = 123 µm2; λ = 0.0052 µm−2. Right: the same region from a 24-year-old autistic man. Ā = 89.1 µm2; λ = 0.0069 µm−2. Scale bars measure 200 µm.

Download figure to PowerPoint


Figure 4. The Delaunay triangulation is a graph with vertices belonging to a given point set. Three points are mutually joined by edges of the graph when the circle through those points contains no other point in its interior. This figure illustrates the construction for a randomly generated set of points, clustered into vertical columns. Some edges of the triangulation have been omitted because of boundary effects. For the meaning of the coloring, see Figure 5.

Download figure to PowerPoint


Figure 5. When a point set exhibits clustering (A), edges of the Delaunay triangulation comprise those between two points in the same cluster, and those between two points in different clusters. If very long edges between points near the boundary of the region of interest are excluded (ie, those omitted in Figure 4), the distribution of edge lengths can be modeled as a mixture of intracluster and intercluster length distributions with means mnear and mfar, respectively. Edges likely to join two points within the same cluster, as determined by thresholding the distribution of edge lengths, are shown in green, while those likely to be intercluster edges are shown in yellow. In contrast, the Delaunay triangulation of a completely random arrangement of points (B) has a unimodal distribution. The horizontal scale of the edge length histograms is in units of w, the horizontal distance between clustering centers in (A).

Download figure to PowerPoint


Figure 6. Pairwise differences (normal autistic) in microanatomical parameters from four cortical areas in a sample of six autistic patients and six matched comparison subjects (16). For illustration, differences in each quantity have been normalized to unit standard deviation. The first two principal components C1 and C2 have been plotted. Arrows indicate the relative magnitude and direction of each normalized parameter, projected onto the (C1, C2) plane. w: minicolumnar width; s: mean distance between neurons within a mini-column; Ā, Ū, λ: Boolean model parameters (Figure 3); mnear, mfar: mean within-cluster and between-cluster edge lengths (Figure 5).

Download figure to PowerPoint

What are the functional implications of increased numbers of narrower mini-columns containing smaller projection neurons? These modular microcircuit assemblies are interconnected by thousands of collateral projections within larger networks. Each mini-column is linked to local networks through myelinated bundles in superficial, or radiate, white matter, and to more distant cortical areas via deeper white matter tracts. The neocortex is presumed to have expanded in evolution through addition of mini-columns by proliferation of their clonal progenitors across an embryonic germinal layer (101, 102). Additive increase in mini-column numbers would entail a geometric increase in short- and long-distance projection fibers in order to maintain a constant degree of transcortical connectivity among modules (61). Longer white matter fibers occupy more space, require disproportionately larger soma to support increased metabolic costs and result in signal processing delays. Selection pressure would therefore be expected to have given rise to modules internally linked by radially oriented processes and integrated into local networks by short collaterals. Proportionately less white matter would be devoted to longer-range connections, encouraging regional functional specialization. Neuropathological descriptions of decreased cell size and narrow mini-columns (20), studies revealing increased superficial white matter (58), and functional imaging studies revealing decreases in activity linking prefrontal and posterior areas (64) support this view. This reorganization might require an increase in frequency of multiple rare polymorphisms at many loci. Combinations of such polymorphisms in turn could support a bias toward a developmental extreme of this evolutionary trend. In primates, more complex prefrontal connections require an extended period of maturation. Developmental insults or epigenetic interactions during this period would unmask underlying genetic vulnerabilities. Autistic syndromes may represent various manifestations of this phenomenon. A putative increase in local interconnectivity and reduced prefrontal transcortical connections in areas subserving cognitive flexibility and prioritizing (area 10) and emotional and social cognition (area 25) is consistent with the clinical picture of stereotypy, rigidity and interpersonal deficits characterizing autism.


  1. Top of page
  2. Abstract

Classical descriptions of inflammatory processes involve a vascular component leading to the accumulation of cells and fluid within the extravascular space. In this regard autism lacks a vascular component and a classical inflammatory response. Levels of quinolinic acid and neopterin in cerebrospinal fluid are either normal or paradoxically reduced in comparison to controls (139). Samples of cerebrospinal fluid obtained from live patients show normal cell counts and protein electrophoresis. There is no evidence of recruitment of leukocytes into the central nervous system. Reported cellular changes do not support their participation in tissue repair and recovery. Furthermore, the clinical symptoms in autism denote lack of concomitant fever, myalgia, arthralgia, anorexia or somnolescence that would sustain a systemic effect for any given inflammatory agent. Rather, the tissue response and their effects appear confined to the brain and indicate an intact blood–brain barrier.

There is no evidence of an adaptive or specific immune response in the brains of autistic patients (89, 125). Immunocytochemical studies have failed to identify T- and B-lymphocyte infiltration and deposition of immunoglobulin/complement in brain tissue (125). Perivascular CD3 and CD20 lymphocyte cell subsets have been rare and reported in both patients and controls (125). The immune system appears activated exclusively in terms of its innate components, astrocytes and microglial cells and the influx of macrophages (7, 52, 125).

Cellular response. Astrocytosis and microglial activation appears most prominent in the cerebellum where the cell reaction accompanies patchy Purkinje cell loss. In the report by Vargas et al, (125) the patient with the most severe Purkinje and granule cell loss was a 25-year-old man with epilepsy. There was no comment as to whether the patient had concomitant endfolium sclerosis. The pattern in the cerebellum suggests hypoxic brain damage but leaves unexplained the sparing (if present) of the hippocampus. Unfortunately, no post-mortem autism study has examined their patients for the presence of chronic fibrillary gliosis in the hippocampus or screened the boundary zones of the cerebral hemispheres (4). In the brain hemispheres the astroglial reaction is most prominent in the subcortical white matter (125). This was assessed qualitatively through immunocytochemistry and quantitatively by Western blotting of GFAP expression in protein homogenates.

Kemper and Bauman's (65) comprehensive screening of full brain sections did not describe astrocytosis in their patients. However, it is well known that protoplasmic astrocytes, the type predominantly found in the gray matter, sparsely stain for GFAP. In the Vargas et al study the presence of GFAP staining astrocytes in the cortex of autistic patients is therefore highly suggestive of reactive gliosis (125). According to the latter authors astrocytes in the brains of autistic patients undergo both proliferative and morphological changes, for example, cells and their processes are increased in numbers (125). There is astrocytic hyperplasia without evidence of migration. Glial fibrillary expression is up-regulated but not overly abundant in the cytoplasm. Gemistocytic astrocytes (swollen with eccentric nuclei) are missing. Similarly, missing are astrocytes arranged in pairs or small groups. The process is generalized and occurs in the presence of preserved brain structures and cytoarchitecture. Preservation of the blood–brain barrier is suggested by the fact that astrocytosis within the brain hemispheres is generalized and not reactive to an identifiable lesion (contrary to Purkinje cell loss in the cerebellum). Thus far, there have not been studies for chronic fibrillary gliosis using either Holzer or phosphotungstic acid-hematoxylin stain. Long post-mortem delays have also prevented assessment of regressive changes (atrophy, pyknosis and clasmatodendrosis).

In the brain hemispheres microglial activation has been reported as most prominent at the junction of the gray and white matter (125). The anatomical location suggests involvement of arcuate fibers in this region (16). The presence of an increased number of microglia does not indicate a disruption of the blood–brain barrier. Circulatory monocytes and activated T lymphocytes are able to cross an intact blood–brain barrier in response to inflammatory or non pathological conditions (59). Post-mortem brain specimens of autistic patients reveal resting and activated non-phagocytotic cells. Brain macrophages (phagocytic microglia) are not present in autism nor is there evidence of tissue damage which would necessitate the same. Similarly, absent is a central focus of injury such as virally infected neurons which would provide for the aggregation of astrocytes and microglia in the form of nodules. The presence of such nodules has seldom been reported in autism (125). However, brain specimens of autistic patients that exhibit activation of microglia also exhibit concurrent activation of astrocytes. In these cases the possibility exists that microglial activation preceded the astrocytic reaction. Microglia may have released early stage mediators, for example IL01, which precipitated astrocytic activation.

The cellular response observed in the white matter of the brain hemispheres of autistic patients is not unexpected. A significant percentage of autistic cases within available brain banks have died from drowning or ischemia/reperfusion injuries.1 Evidence from the literature implicates ischemia/reperfusion in periventricular leukomalacia, reactive gliosis of the white matter and cerebral palsy (4, 66, 127). Reoxygenation of damaged tissue generates free radicals and increases the production of cytokines. Cause of death and differences in pre-agonal and agonal conditions among patient and control series may explain some of the neuroinflammatory findings previously discussed. Similarly, lack of control for pre-agonal and agonal factors may account for findings of increased oxidative loads (without evidence of concomitant lipofuscin accumulation) in post-mortem tissue. In the future, it may be of interest to screen available tissue with different techniques including beta-amyloid precursor protein. Immunocytochemical staining for beta-amyloid precursor protein has been validated as a marker of axonal injury in patients surviving different forms of white matter damage including those of ischemia/reperfusion observed in near-drowning victims (104, 105). Preliminary results reported by Weigel et al indicate a high level of beta-amyloid precursor protein staining in some autistic patients (131).


  1. Top of page
  2. Abstract

A revision of reported neuropathological findings in autism indicates little or no evidence of acute cellular changes. The evidence fails to sustain that the cell's capacity to maintain homeostasis has been exceeded by environmental exigencies. Similarly, lacking is proof of an agent that at a critical stage of development affects several organs. Rather, autism appears to result from a monotopic effect where a basic abnormality early in brain development results in a cascade of pathological events that are significantly influenced by environmental factors. These effects and interrelationships appear to be confined to the brain. Characteristic of the involved pathology are an intact blood–brain barrier, absence of a vascular component and a relationship to increased brain growth during early childhood.

When discussing monotopic effects many studies fail to differentiate core pathology from secondary findings, for example, the apoptotic loss of neurons and oligodendrocytes that occurs for an extended period of time after developmental insults (46, 85, 98, 124, 129). In autism secondary effects may thus account for different aspects of comorbidity. Thus, despite being a developmental disorder with manifestations antedating 3 years of age, onset of epilepsy is delayed and not present in every case. This suggests that additional mechanisms may be responsible for seizure activity. In effect, animal models show delays in development of distribution patters of subset of inhibitory cells within and adjacent to malformed cortex (108). Similarly, there is evidence of progressive reorganization and acquisition of dysplastic changes (eg, giant neurons, disturbed lamination, clustering of neurons) surrounding areas of cortex that have been injured during development (72, 76). Progressive reorganization of the cortex may provide secondary aspects of pathology responsible for some of the neurological sequela observed in autism, for example seizures and cognitive impairment.

The possibility of secondary effects bears significance for autism when discussing genetic changes. Occasional monogenetic defects underlie certain epileptic conditions. Usually, these genes encode for voltage or ligand gated ion channels. However, a majority of epileptic patients lack single gene defects suggesting the possibility that genetic profiles rather than individual genes are selectively induced in ictal foci irrespective of underlying etiology (99). The results question whether the induced genes are a cause or effect of the ictal neurons. In autism, gene profiles may be “remodeled” by comorbidities and represent downstream events to core pathological processes.

In summary, patients with autism are born full term; they are not asphyxiated and do not require resuscitation. There is no evidence of premature rupture of membranes or of antepartum/postpartum factors predictive of encephalopathy (2). Despite sporadic case reports of maternal infections and the insistence of modeling the condition with viral agents there is little evidence suggesting a primary role for maternal infections or fevers during gestation. Several studies suggest that birth complications in autism are the result of preexisting prenatal abnormalities (48, 50). The literature indicates that underlying brain alterations in autism occur well before symptom expression. Most evidence favors the first trimester of gestation as providing a time window of susceptibility. The resultant pathology appears to be widespread adding to the meaning of autism as a “pervasive” developmental disorder of childhood. The cortex seems to be preferentially affected with a special predilection for the frontotemporal lobes. Trying to distinguish primary from secondary pathological changes has proven problematic. Future studies should try to discriminate between the pathology of core processes in autism and those findings related to cause of death or comorbidities.


  1. Top of page
  2. Abstract

This article is based upon work supported by the National Alliance for Autism Research, and NIMH grants MH62654 and MH69991.


  1. Top of page
  2. Abstract
  • 1
    Satcher D (1999) U.S. Surgeon General's Report on Mental Health, ch. 3, section 6.
  • 2
    Adamson SJ, Alessandri LM, Badawi N, Burton PR, Pemberton PJ, Stanley F (1995) Predictors of neonatal encephalopathy in full term infants. BMJ 311:598602.
  • 3
    APA (2000) Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR. APA: Arlington, Virginia.
  • 4
    Archer D, Walz W (2000) Astrocytic subtypes and the gliotic response. In: Brain Hypoxia and Ischemia. MOehmichen (ed.), pp. 101113. Schmidt-Romhild: Lubeck.
  • 5
    Armstrong E, Curtis M, Buxhoeveden DP, Fregoe C, Zilles K, Casanova MF, McCarthy WF (1991) Cortical gyrification in the rhesus monkey: a test of the mechanical folding hypothesis. Cereb Cortex 1:426432.
  • 6
    Aylward EH, Minshew NJ, Field K, Sparks BF, Singh N (2002) Effects of age on brain volume and head circumference in autism. Neurology 59:175183.
  • 7
    Bailey A, Luthert P, Dean A, Harding B, Janota I, Montgomery M, Rutter M, Lantos P (1998) A clinicopathological study of autism. Brain 121:889905.
  • 8
    Bassett DS, Bullmore E (2006) Small-world brain networks. Neuroscientist 12:512523.
  • 9
    Bauman M, Kemper T (1985) Histoanatomic observations of the brain in early infantile autism. Neurology 35:866874.
  • 10
    Bauman M, Kemper T (1994) Neuroanatomic observations of the brain in autism. In: The Neurobiology of Autism. MBauman, TKemper (eds), pp. 119145. The Johns Hopkins University Press: Baltimore, MD.
  • 11
    Ben-Ari Y (2006) Basic developmental rules and their implications for epilepsy in the immature brain. Epileptic Disord 8:91102.
  • 12
    Buxhoeveden DP, Casanova MF (2005) The Cell Column in Comparative Anatomy. Nova Science Publishers, Inc: New York.
  • 13
    Buxhoeveden DP, Switala AE, Roy E, Casanova MF (2000) Quantitative analysis of cell columns in the cerebral cortex. J Neurosci Methods 97:7.
  • 14
    Carper RA, Moses P, Tigue ZD, Courchesne E (2002) Cerebral lobes in autism: early hyperplasia and abnormal age effects. NeuroImage 16:1038.
  • 15
    Casanova M (2004) White matter volume increase and minicolumns in autism. Ann Neurol 56:453.
  • 16
    Casanova MF (2006) Neuropathological and genetic findings in autism: the significance of a putative minicolumnopathy. Neuroscientist 12:435441.
  • 17
    Casanova MF, Tillquist C (2007) Encephalization, emergent properties, and psychiatry: a minicolumnar perspective. Neuroscientist (in press).
  • 18
    Casanova MF, Buxhoeveden DP, Switala AE, Roy E (2002) Minicolumnar pathology in autism. Neurology 58:428432.
  • 19
    Casanova MF, Buxhoeveden DP, Switala AE, Roy E (2002) Neuronal density and architecture (Gray Level Index) in the brains of autistic patients. J Child Neurol 17:515521.
  • 20
    Casanova MF, Van Kooten IA, Switala AE, Van Engeland H, Heinsen H, Steinbusch H, Hof P, Trippe J, Stone J, Schmitz C (2006) Minicolumnar abnormalities in autism. Acta Neuropathol (Berl) 112:287303.
  • 21
    Casanova MF, Van Kooten IA, Switala AE, Van Engeland H, Heinsen H, Steinbusch HWM, Hof PR, Schmitz C (2006) Abnormalities of cortical minicolumnar organization in the prefrontal lobes of autistic patients. Clin Neurosci Res 6:127133.
  • 22
    Casanova MF. Autism as a Minicolumnopathy: Theory and Evidence. In: Autism Current Theories and Evidence, Current Clinical Neurology. AZimmeiman (ed.). The Human Press Inc. In press.
  • 23
    CDC (2007) CDC releases new data on autism spectrum disorders (ASDs) from multiple communities in the United States. 18 February 2007.
  • 24
    CHA (2000) Public Law (P.L.) 106–310, Title I, Section 104.
  • 25
    Chambers D, Fishell G (2006) Functional genomics of early cortex patterning. Genome Biol 7:202.
  • 26
    Chess S (1977) Follow-up report on autism in congenital rubella. J Autism Child Schizophr 7:6981.
  • 27
    Chklovskii DB, Schikorski T, Stevens CF (2002) Wiring optimization in cortical circuits. Neuron 34:341.
  • 28
    Coleman M, Betancur C (2005) Introduction. In: The Neurology of Autism. MColeman (ed.), pp. 339. Oxford University Press: New York.
  • 29
    Courchesne E, Yeung-Courchesne R, Press G, Hesselink J, Jernigan T (1988) Hypoplasia of cerebellar vermal lobules VI and VII in autism. N Engl J Med 318:13491354.
  • 30
    Courchesne E, Saitoh O, Yeung-Courchesne R, Press G, Lincoln A, Haas R, Schreibman L (1994) Abnormality of cerebellar vermian lobules VI and VII in patients with infantile autism: identification of hypoplastic and hyperplastic subgroups with MR imaging. AJR Am J Roentgenol 162:123130.
  • 31
    Courchesne E, Karns CM, Davis HR, Ziccardi R, Carper RA, Tigue ZD, Chisum HJ, Moses P, Pierce K, Lord C, Lincoln AJ, Pizzo S, Schreibman L, Haas RH, Akshoomoff NA, Courchesne RY (2001) Unusual brain growth patterns in early life in patients with autistic disorder: an MRI study. Neurology 57:245254.
  • 32
    Dawson G, Webb S, Schellenberg G, Dager S, Friedman S, Aylward E, Richards T (2002) Defining the broader phenotype of autism: genetic, brain, and behavioral perspectives. Dev Psychopathol 14:581611.
  • 33
    De No L (1938) The cerebral cortex: architecture, intracortical connections, and motor projections. In: Physiology of the Nervous System. JFFulton (ed.), pp. 291339. Oxford University Press: London.
  • 34
    DeFelipe J (2005) Reflections on the structure of the cortical minicolumn. In: Neocortical Modularity and the Cell Minicolumn. MFCasanova (ed.), pp. 5792. Nova Biomedical Books: New York.
  • 35
    DeFelipe J, Hendry S, Hashikawa T, Molinari M, Jones E (1990) A microcolumnar structure of monkey cerebral cortex revealed by immunocytochemical studies of double bouquet cell axons. Neuroscience 37:655673.
  • 36
    Del Rio MR, DeFelipe J (1997) Double bouquet cell axons in the human temporal neocortex: relationship to bundles of myelinated axons and colocalization of calretinin and calbindin D-28k immunoreactivities. J Chem Neuroanat 13:243.
  • 37
    Department of Health and Human Services NIoMH (2004) Congressional appropriations committee reports on the state of autism research.
  • 38
    Douglas RJ, Martin KAC (2004) Neuronal circuits of the neocortex. Annu Rev Neurosci 27:419451.
  • 39
    FDA (2006) FDA approves the first drug to treat irritability associated with autism, risperdal. FDA News. .
  • 40
    Ferrer I, Tunon T, Soriano E, Del Rio A, Iraizoz I, Fonseca M, Guionnet N (1992) Calbindin immunoreactivity in normal human temporal neocortex. Brain Res 572:3341.
  • 41
    Fidler DJ, Bailey JN, Smalley SL (2000)Macrocephaly in autism and other pervasive developmental disorders. Dev Med Child Neurol 42:737740.
  • 42
    Filipek PA (2005) Medical aspects of autism. In. Handbook of Autism and Pervasive Developmental Disorders. FRVolkmar, RPaul, AKlin, DCohen (eds), pp. 534578. John Wiley and Sons: Hoboken.
  • 43
    Filipek PA, Accardo PJ, Baranek GT, Cook EH, Dawson G, Gordon B, Gravel JS, Johnson CP, Kallen RJ, Levy SE, Minshew NJ, Ozonoff S, Prizant BM, Rapin I, Rogers SJ, Stone WL, Teplin S, Tuchman RF, Volkmar FR (1999) The screening and diagnosis of autistic spectrum disorders. J Autism Dev Disord 29:439484.
  • 44
    Fombonne E, Rogé B, Claverie J, Courty S, Frémolle J (1999) Microcephaly and macrocephaly in autism. J Autism Dev Disord 29:113119.
  • 45
    Gaffney GR, Tsai LY (1987) Magnetic resonance imaging of high level autism. J Autism Dev Disord 17:433438.
  • 46
    Gill R, Soriano M, Blomgren K, Hagberg H, Wybrecht R, Miss M-T, Hoefer S, Adam G, Niederhauser O, Kemp JA, Loetscher H (2002) Role of caspase-3 activation in cerebral ischemia-induced neurodegeneration in adult and neonatal brain. J Cereb Blood Flow Metab 22:420.
  • 47
    Gillberg C (2005) The epidemiology of autism. In: The Neurology of Autism. MColeman (ed.), pp. 119135. Oxford University Press: New York.
  • 48
    Gillberg C, Gillberg IC (1983) Infantile autism: a total population study of reduced optimality in the pre-, peri-, and neonatal period. J Autism Dev Disord 13:153166.
  • 49
    Goldberg J, Szatmari P, Nahmias C (1999) Imaging of autism: lessons from the past to guide studies in the future. Can J Psychiatry 44:793801.
  • 50
    Goodman R (1990) Technical note: are perinatal complications causes or consequences of autism? J Child Psychol Psychiatry 31:809812.
  • 51
    Greenberg DA, Hodge SE, Sowinski J, Nicoll D (2001) Excess of twins among affected sibling pairs with autism: implications for the etiology of autism. Am J Hum Genet 69:10621067.
  • 52
    Guerin P, Lyon G, Barthelemy C, Sostak E, Chevrollier V, Garreau B, Lelord G (1996) Neuropathological study of a case of autistic syndrome with severe mental retardation. Dev Med Child Neurol 38:203211.
  • 53
    Hardan AY, Minshew NJ, Mallikarjuhn M, Keshavan MS (2001) Brain volume in autism. J Child Neurol 16:421424.
  • 54
    Hardan AY, Jou RJ, Keshavan MS, Varma R, Minshew NJ (2004) Increased frontal cortical folding in autism: a preliminary MRI study. Psychiatr Res Neuroimaging 131:263.
  • 55
    Harding BN, Copp AJ (2002) Malformations. In: Greenfield's Neuropathlogy. DIGraham, PLLantos (eds), pp. 357483. Arnold: London.
  • 56
    Hartfuss E, Galli R, Heins N, Gotz M (2001) Characterization of CNS precursor subtypes and radial glia. Dev Biol 229:15.
  • 57
    Herbert MR, Ziegler DA, Deutsch CK, O'Brien LM, Lange N, Bakardjiev A, Hodgson J, Adrien KT, Steele S, Makris N, Kennedy D, Harris GJ, Caviness VS (2003) Dissociations of cerebral cortex, subcortical and cerebral white matter volumes in autistic boys. Brain 126:11821192.
  • 58
    Herbert MR, Ziegler DA, Makris N, Filipek PA, Kemper TL, Normandin JJ, Sanders HA, Kennedy DN, Caviness VS (2004) Localization of white matter volume increase in autism and developmental language disorder. Ann Neurol 55:530540.
  • 59
    Hickey WF (1999) Leukocyte traffic in the central nervous system: the participants and their roles. Semin Immunol 11:125.
  • 60
    Hof PR, Knabe R, Bovier P, Bouras C (1991) Neuropathological observations in a case of autism presenting with self-injury behavior. Acta Neuropathol (Berl) 82:321326.
  • 61
    Hofman MA (2001) Brain evolution in hominids:are we at the end of the road? In: Evolutionary Anatomy of the Primate Cerebral Cortex. DFalk, KRGibson (eds), pp. 113127. Cambridge University Press: Cambridge.
  • 62
    Ingram JL, Peckham SM, Tisdale B, Rodier PM (2000) Prenatal exposure of rats to valproic acid reproduces the cerebellar anomalies associated with autism. Neurotoxicol Teratol 22:319.
  • 63
    Ivarsson SA, Bjerre I, Vegfors P, Ahlfors K (1990) Autism as one of several disabilities in two children with congenital cytomegalovirus infection. Neuropediatrics 21:102103.
  • 64
    Just MA, Cherkassky VL, Keller TA, Minshew N (2004) Functional connectivity in an fMRI working memory task in high-functioning autism. NeuroImage 24:810821.
  • 65
    Kemper TL (1988) Neuroanatomic studies of dyslexia and autism. In: Disorders of the Developing Nervous System: Changing Views On Their Origins, Diagnosis and Treatments. JSwann, AMesser (eds), pp. 125154. Alan R. Liss: New York.
  • 66
    Kinney HC (2002) Perinatal neurpathology. In: Greenfield's Neuropathology. DIGraham, PLLantos (eds), pp. 519606. Arnold: New York.
  • 67
    Krageloh-Mann I, Toft P, Lunding J, Andresen J, Pryds O, Lou HC (1999) Brain lesions in preterms: origin, consequences and compensation. Acta Paediatr 88:897908.
  • 68
    Kriegstein AR, Noctor SC (2004) Patterns of neuronal migration in the embryonic cortex. Trends Neurosci 27:392.
  • 69
    Lainhart JE, Piven J, Wzorek M, Landa R, Santangelo SL, Coon H, Folstein SE (1997) Macrocephaly in children and adults with autism. J Am Acad Child Adolesc Psychiatry 36:282290.
  • 70
    Levitt JG, Blanton RE, Smalley S, Thompson PM, Guthrie D, McCracken JT, Sadoun T, Heinichen L, Toga AW (2003) Cortical sulcal maps in autism. Cereb Cortex 13:728735.
  • 71
    Lohmann H, Koppen H (1995) Postnatal development of pyramidal dendritic and axonal bundles in the visual cortex of the rat. J Hirnforsch 36:101111.
  • 72
    Lombroso CT (2000) Can early postnatal closed head injury induce cortical dysplasia. Epilepsia 41:245253.
  • 73
    Malatesta P, Hartfuss E, Gotz M (2000) Isolation of radial glial cells by fluorescent-activated cell sorting reveals a neuronal lineage. Development 127:52535263.
  • 74
    Malatesta P, Hack MA, Hartfuss E, Kettenmann H, Klinkert W, Kirchhoff F, Gotz M (2003) Neuronal or glial progeny: regional differences in radial glia fate. Neuron 37:751.
  • 75
    Mancini J, Lethel V, Hugonenq C, Chabrol B (2001) Brain injuries in early foetal life: consequences for brain development. Dev Med Child Neurol 43:5255.
  • 76
    Marín-Padilla M (1999) Developmental neuropathology and impact of perinatal brain damage. III: gray matter lesions of the neocortex. J Neuropathol Exp Neurol 58:407429.
  • 77
    Meyer G (2007) Genetic control of neuronal migrations in human cortical development. Adv Anat Embryol Cell Biol 189(1 p preceding 1):1111.
  • 78
    Mintz M, Alessandri M, Curatolo P (2006) Treatment approaches for the autism spectrum disorders. In: Autism: Neurological Disorder of Early Brain Development. RFTuchman, IRapin (eds), pp. 281307. Mac Keith Press for the International Child Neurology Association: London.
  • 79
    Moore SJ, Turnpenny P, Quinn A, Glover S, Lloyd DJ, Montgomery T, Dean JCS (2000) A clinical study of 57 children with fetal anticonvulsant syndromes. J Med Genet 37:489497.
  • 80
    Mountcastle VB (2003) Introduction. Computation in cortical columns. Cereb Cortex 13:24.
  • 81
    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.
  • 82
    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 22:31613173.
  • 83
    Northcutt RG (1981) Evolution of the telencephalon in nonmammals. Annu Rev Neurosci 4:301350.
  • 84
    Northcutt RG, Kaas JH (1995) The emergence and evolution of mammalian neocortex. Trends Neurosci 18:373.
  • 85
    Northington FJ, Ferriero DM, Flock DL, Martin LJ (2001) Delayed neurodegeneration in neonatal rat thalamus after hypoxia-ischemia is apoptosis. J Neurosci 21:19311938.
  • 86
    Nowell MA, Hackney DB, Muraki AS, Coleman M (1990) Varied MR appearance of autism: fifty-three pediatric patients having the full autistic syndrome. Magn Reson Imaging 8:811816.
  • 87
    NRC (2001) Educating Children with Autism. National Academic Press: Washington, DC.
  • 88
    Ong W, Garey L (1990) Neuronal architecture of the human temporal cortex. Anat Embryol (Berl) 181:351364.
  • 89
    Pardo CA, Vargas DL, Zimmerman AW (2005) Immunity, neuroglia and neuroinflammation in autism. Int Rev Psychiatry 17:485495.
  • 90
    Peters A, Sethares C (1991) Organization of pyramidal neurons in area 17 of monkey visual cortex. J Comp Neurol 306:123.
  • 91
    Peters A, Sethares C (1996) Myelinated axons and the pyramidal cell modules in monkey primary visual cortex. J Comp Neurol 365:232255.
  • 92
    Peters A, Sethares C (1997) The organization of double bouquet cells in monkey striate cortex. J Neurocytol 26:779.
  • 93
    Pickett J, London E (2005) The neuropathology of autism: a review. J Neuropathol Exp Neurol 64:925935.
  • 94
    Piven J, Berthier ML, Starkstein SE, Nehme E, Pearlson G, Folstein S (1990) Magnetic resonance imaging evidence for a defect of cerebral cortical development in autism. Am J Psychiatry 147:734739.
  • 95
    Piven J, Nehme E, Simon J, Barta P, Pearlson G, Folstein SE (1992) Magnetic resonance imaging in autism: measurement of the cerebellum, pons, and fourth ventricle. Biol Psychiatry 31:491.
  • 96
    Piven J, Arndt S, Bailey J, Havercamp S, Andreasen NC, Palmer P (1995) An MRI study of brain size in autism. Am J Psychiatry 152:11451149.
  • 97
    Prothero JW, Sundsten JW (1984) Folding of the cerebral cortex in mammals. A scaling model. Brain Behav Evol 24:152167.
  • 98
    Puka-Sundvall M, Wallin C, Gilland E, Hallin U, Wang X, Sandberg M, Karlsson J-O, Blomgren K, Hagberg H (2000) Impairment of mitochondrial respiration after cerebral hypoxia-ischemia in immature rats: relationship to activation of caspase-3 and neuronal injury. Dev Brain Res 125:43.
  • 99
    Rakhade SN, Yao B, Ahmed S, Asano E, Beaumont TL, Shah AK, Draghici S, Krauss R, Chugani HT, Sood S, Loeb JA (2005) A common pattern of persistent gene activation in human neocortical epileptic foci. Ann Neurol 58:736747.
  • 100
    Rakic P (1972) Mode of cell migration to the superficial layers of fetal monkey neocortex. J Comp Neurol 145:6183.
  • 101
    Rakic P (1988) Intrinsic and extrinsic determinants of neocortical parcellation. A radial unit model. In: Neuorbiology of the Neocortex. PRakic, WSinger (eds), pp. 527. Wiley and Sons: New York.
  • 102
    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.
  • 103
    Raymond GV, Bauman ML, Kemper TL (1995) Hippocampus in autism: a Golgi analysis. Acta Neuropathol (Berl) 91:117.
  • 104
    Reichard RR, White CL, Hladik CL, Dolinak D (2003) Beta-amyloid precursor protein staining in nonhomicidal pediatric medicolegal autopsies. J Neuropathol Exp Neurol 62:237241.
  • 105
    Reichard RR, White CL, Hladik CL, Dolinak D (2003) Beta-amyloid precursor protein staining of nonaccidental central nervous system injury in pediatric autopsies. J Neurotrauma 20:347355.
  • 106
    Ritvo ER, Freeman BJ, Scheibel A, Duong T, Robinson H, Guthrie D, Ritvo A (1986) Lower Purkinje cell counts in the cerebella of four autistic subjects: initial findings of the UCLA-NSAC Autopsy Research Report. Am J Psychiatry 143:862866.
  • 107
    Rodier PM (2002) Converging evidence for brain stem injury in autism. Dev Psychopathol 14:537557.
  • 108
    Rosen GD, Jacobs KM, Prince DA (1998) Effects of neonatal freeze lesions on expression of parvalbumin in rat neocortex. Cereb Cortex 8:753761.
  • 109
    Samrén EB, Van Duijn CM, Christiaens GC, Hofman A, Lindhout D (1999) Antiepileptic drug regimens and major congenital abnormalities in the offspring. Ann Neurol 46:739746.
  • 110
    Schaefer GB, Thompson JN, Bodensteiner JB, McConnell JM, Kimberling WJ, Gay CT, Dutton WD, Hutchings DC, Gray SB (1996) Hypoplasia of the cerebellar vermis in neurogenetic syndromes. Ann Neurol 39:382385.
  • 111
    Schlaug G, Schleicher A, Zilles K (1995) Quantitative analysis of the columnar arrangement of neurons in the human cingulate cortex. J Comp Neurol 351:441452.
  • 112
    Schmechel D, Rakic P (1979) Arrested proliferation of radial glial cells during midgestation in rhesus monkey. Nature 277:303305.
  • 113
    Schneider T, Przewlocki R (2005) Behavioral alterations in rats prenatally exposed to valproic acid: animal model of autism. Neuropsychopharmacology 30:8089.
  • 114
    Seldon HL (1981) Structure of human auditory cortex. I. Cytoarchitectonics and dendritic distributions. Brain Res 229:277.
  • 115
    Smalley SL, Petri HE (2005) Autism and tuberous sclerosis complex. In: The Neurobiology of Autism. MBauman, TLKemper (eds), pp. 265275. Johns Hopkins University Press: Baltimore.
  • 116
    Sparks BF, Friedman SD, Shaw DW, Aylward EH, Echelard D, Artru AA, Maravilla KR, Giedd JN, Munson J, Dawson G, Dager SR (2002) Brain structural abnormalities in young children with autism spectrum disorder. Neurology 59:184192.
  • 117
    Squier M, Chamberlain P, Zaiwalla Z, Anslow P, Oxbury J, Gould S, McShane MA (2000) Five cases of brain injury following amniocentesis in mid-term pregnancy. Dev Med Child Neurol 42:554560.
  • 118
    Stevenson RE, Schroer RJ, Skinner C, Fender D, Simensen RJ (1997) Autism and macrocephaly. Lancet 349:17441745.
  • 119
    Strömland K, Miller MT (1993) Thalidomide embryopathy: revisited 27 years later. Acta Ophthalmol (Copenh) 71:238245.
  • 120
    Strömland K, Nordin V, Miller M, Akerström B, Gillberg C (1994) Autism in thalidomide embryopathy: a population study. Dev Med Child Neurol 36:351356.
  • 121
    Tamamaki N, Nakamura K, Okamoto K, Kaneko T (2001) Radial glia is a progenitor of neocortical neurons in the developing cerebral cortex. Neurosci Res 41:51.
  • 122
    Torrey EF, Hersh SP, McCabe KD (1975) Early childhood psychosis and bleeding during pregnancy. A prospective study of gravid women and their offspring. J Autism Child Schizophr 5:287297.
  • 123
    Towfighi J, Marks K, Palmer E, Vannucci R (1979) Möbius syndrome. Neuropathologic observations. Acta Neuropathol (Berl) 48:1117.
  • 124
    Van de Berg WDJ, Schmitz C, Steinbusch HWM, Blanco CE (2002) Perinatal asphyxia induced neuronal loss by apoptosis in the neonatal rat striatum: a combined TUNEL and stereological study. Exp Neurol 174:29.
  • 125
    Vargas DL, Nascimbene C, Krishnan C, Zimmerman AW, Pardo CA (2005) Neuroglial activation and neuroinflammation in the brain of patients with autism. Ann Neurol 57:6781.
  • 126
    Viebahn C (1990) Correlation between differences in the structure of dendrite bundles and cytoarchitectonic patterns in the cerebral cortex of the rabbit. J Hirnforsch 31:645652.
  • 127
    Volpe JJ (1995) Neurology of the Newborn, 3rd edn. WB Saunders: Philadelphia.
  • 128
    Von Bonin G, Mehler W (1971) On columnar arrangement of nerve cells in cerebral cortex. Brain Res 27:19.
  • 129
    Wang X, Karlsson JO, Zhu C, Bahr BA, Hagberg H, Blomgren K (2001) Caspase-3 activation after neonatal rat cerebral hypoxia-ischemia. Biol Neonate 79:172179.
  • 130
    Watts DJ, Strogatz SH (1998) Collective dynamics of /“small-world/” networks. Nature 393:440.
  • 131
    Weigel J, Nowicki K, Kuchna I, Weigel J, Wiezba-Bobrowicz T, Rabe A, Cohen IL, Brown TW (2007) Abnormal striatal circuits in autism-links between structure and function. 6th International Meeting for Autism Research: Seattle, Washington.
  • 132
    Weissman T, Noctor SC, Clinton BK, Honig LS, Kriegstein AR (2003) Neurogenic radial glial cells in reptile, rodent and human: from mitosis to migration. Cereb Cortex 13:550559.
  • 133
    Westman JA (2006) Medical Genetics for the Modern Clinician. Lippincoat, Williams, and Wilkins: Baltimore, MD.
  • 134
    Williams G, King J, Cunningham M, Stephan M, Kerr B, Hersh JH (2001) Fetal valproate syndrome and autism: additional evidence of an association. Dev Med Child Neurol 43:202206.
  • 135
    Wing L (2001) The Autism Spectrum. Ulysses Press: Berkley, CA.
  • 136
    Woodhouse W, Bailey A, Rutter M, Bolton P, Baird G, Le Couteur A (1996) Head circumference in autism and other pervasive developmental disorders. J Child Psychol Psychiatry 37:665671.
  • 137
    Yamashita Y, Fujimoto C, Nakajima E, Isagai T, Matsuishi T (2003) Possible association between congenital cytomegalovirus infection and autistic disorder. J Autism Dev Disord 33:455.
  • 138
    Zecevic N (2004) Specific characteristic of radial glia in the human fetal telencephalon. Glia 48:2735.
  • 139
    Zimmerman AW, Jyonouchi H, Comi AM, Connors SL, Milstien S, Varsou A, Heyes MP (2005) Cerebrospinal fluid and serum markers of inflammation in autism. Pediatr Neurol 33:195.
  • 1

    A recent survey (7/20/07) of material collected by the Autism Tissue Program (ATP) shows 35 autistic patients. Eleven of these patients drowned (31.4% of total, PMI 21.0 ± 12 h), 1 stated no cause of death, and 23 died from diverse causes (PMI 23.7 ± 16.7 hours) including seizures, circulatory failure, sepsis, anoxic encephalopathy, and acute respiratory distress. Of the eleven drowning victims 3 were missing medical history and autopsy findings, 2 received CPR and survived for an indeterminate amount of time.