Neuropathology and Applied Neurobiology (2016), 42, 115–134 doi: 10.1111/nan.12227

Review: Cortical construction in autism spectrum disorder: columns, connectivity and the subplate

J. J. Hutsler* and M. F. Casanova† *Department of Psychology, Program in Neuroscience, University of Nevada, Reno, and †Department of Psychiatry and Behavioral Science, University of Louisville School of Medicine, Louisville, USA

J. J. Hutsler and M. F. Casanova (2016) Neuropathology and Applied Neurobiology Cortical construction in autism spectrum disorder: columns, connectivity and the subplate

The cerebral cortex undergoes protracted maturation terize autism. These alterations to cortical circuitry likely during human development and exemplifies how biology underlie the behavioural phenotype in autism and con- and environment are inextricably intertwined in the con- tribute to the unique pattern of deficits and strengths that struction of complex neural circuits. Autism spectrum characterize cognitive functioning. Recent findings within disorders are characterized by a number of pathological the cortical subplate may indicate that alterations to changes arising from this developmental process. These cortical construction begin prenatally, before activity- include: (i) alterations to columnar structure that have dependent connections are established, and are in need of significant implications for the organization of cortical cir- further study. A better understanding of cortical develop- cuits and connectivity; (ii) alterations to synaptic spines ment in autism spectrum disorders will draw bridges on individual cortical units that may underlie specific between the microanatomical computational circuitry types of connectional changes; and (iii) alterations within and the atypical behaviours that arise when that circuitry the cortical subplate, a region that plays a role in proper is modified. In addition, it will allow us to better exploit the cortical development and in regulating interregional com- constructional plasticity within the brain to design more munication in the mature brain. Although the cerebral targeted interventions that better manage atypical corti- cortex is not the only structure affected in the disorder, it is cal construction and that can be applied very early in post- a fundamental contributor to the behaviours that charac- natal life.

Keywords: cerebral cortex, neurodevelopment, neuropathology, prenatal

and its delayed maturation is associated with a flexible Introduction neurocomputational architecture that is open to a wide The human cerebral cortex occupies a unique place variety of early environmental influences [6,7]. It is, within the organization of the nervous system. Its posi- perhaps, the structure in the brain that best demonstrates tion, development and evolution provide the base upon how biology and environment are inextricably inter- which complex cognitive functions are established. The twined in the construction of the complex neural circuits organization of this multilayered sheet of tissue reflects that underpin our success and behavioural flexibility as a both its evolutionary and developmental history [1–4]. species [8,9]. Relative to the other components of the central nervous In developmental disorders where cognitive abilities are system, its expansion in human primates is remarkable [5] significantly impaired, the origin of the problem is often readily apparent in the atypical organization of the cer- Correspondence: Jeffrey J. Hutsler, Department of Psychology, ebral cortex [10,11]. In contrast, pervasive developmental Program in Neuroscience, University of Nevada, MS 296, 1664 N. Virginia Ave., Reno, NV 89557-0296, USA. Tel: +1 775 682 8694; disorders, such as autism, often show a remarkably intact Fax: +1 775 784 1126; E-mail: [email protected] cortical organization upon casual microscopic examina-

© 2015 British Neuropathological Society 115 116 J. J. Hutsler and M. F. Casanova tion [12–14], although a number of subtle pathological afferent and efferent connections, their distribution of changes have been described [15]. Intensified efforts to neurones and their spatial expression of neuroactive pro- understand the organization of the cortex in autism spec- teins and peptides. Cortical layers can be described based trum disorder (ASD), along with numerous advances in upon their cellular content, their afferent and efferent quantitative microscopy, have revealed a litany of differ- connections, and the timing of maturation during neural ence from neurotypical (NT) subjects. While the cerebral development. Much like other neuronal structures, these cortex is not the only structure within the brain to be layers are not independent of one another, but are bound impacted by autism, its protracted development leaves it together and brought into functional registration by vulnerable to the shaping influences of earlier develop- columns of cells that traverse the layers [27,28]. mental events. Indeed, the very behaviours that are Single columns of cells share similarities in their required for a diagnosis of autism are largely dependent response properties [28,29] and are visually distinguish- on cortical functioning. able in many areas of cortex as vertical clusters of neu- The behavioural phenotype in ASD is not solely charac- rones separated from neighbouring minicolumns by a cell terized by deficits in cognitive functions. Rather, individu- sparse neuropil space. The sharing of response character- als with ASD often show better performance on many istics between spatially adjacent columns arises naturally types of cognitive tasks. These include embedded figures from the types of inputs that a cortical region receives, tasks, block design tasks and visual search [16–18]. In along with the fundamental rules that guide activity- addition to these simple tasks, 5% to 10% of the ASD dependent strengthening and weakening of connections population present with islands of ability that are remark- between neural units [6,7,30,31]. Neurones within a able in quality [19]. In the context of the present discus- minicolumn share a developmental history [32,33] and sion, these findings suggest that the cortex in autism may be the basis upon which evolutionary selection cannot be simply portrayed as abnormal or atypical. has acted to modify the computational architecture Rather, the organization of the cerebral cortex, and its [2,3,34,35]. More recently, the minicolumn as a distinct resulting computational and functional properties, repre- functional unit has been reaffirmed in primate prefrontal sents a kind of state that may reliably differ from the popu- cortex (PFC) by studying emergent higher cognitive func- lation as a whole. Certain features of this organization tions from encoded interlaminar communication within a may well be adaptive in specific environmental contexts. column [36]. Indeed, forms of this organization may exist in the general Although we have yet to fully characterize the connec- population in a fashion that is subclinical and even ben- tional and physiological organization within a single eficial at the individual level [20]. More extreme forms of column, models of internal columnar organization with this organization are clearly not adaptive. They interfere varying levels of microanatomical detail have been sug- with the day-to-day social and communicative activities gested [37–43]. One popular framework for understand- that have been so critical to our survival and success as a ing the column, and how minicolumns are assembled species. into the overall cortical sheet, is by framing its organiza- Although several reviews of changes within the brain tion during development. Prior to the embryonic estab- of ASD subjects are available [21–26], here we will focus lishment of the cortical plate (the aggregation of cells on three topics that are specifically related to the organi- that will become the cerebral cortex), there exists a layer zation of cerebral cortex: (i) the atypical organization of of cells and fibres at the surface of the neural tube: the cortical structure; (ii) alterations to connectional organi- preplate [32,44–46]. Many future cortical neurones are zation; and (iii) the potentially critical role the subplate generated at the inner wall of the developing neural plays in forming these structural and connectional tube by a group of precursor cells aggregated within the changes. ventricular and subventricular zones. Precursor cells undergo several rounds of symmetrical division in which more precursor cells are generated. This symmetrical NT cortical organization division gives way to several rounds of asymmetrical The organization of the human cerebral cortex is cur- division where each dividing precursor cell produces rently understood at multiple levels of analysis. Cortical another precursor cell and a neuroblast. The number of regions differ based upon their functional activity, their rounds of symmetrical cell division plays an important

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 117 role in determining the number of cortical columns that Structural changes to cortical organization will eventually populate the cortical sheet, while the in ASD asymmetrical cell divisions that follow partially deter- mine the final number of cells within a cortical column. Quantitative studies of post mortem cortical tissue have Such a model provides a window into how modifications revealed a number of important differences between ASD of cortical size, thickness and complexity have arisen and NT organization (for reviews, see Kemper and during the evolutionary history of the mammalian brain Bauman and other several authors [14,21,22,55–58]). In [47]. addition to direct post mortem studies, several types of Some neuroblasts have a leading edge attached to the alterations to cortical computational architecture have pia and migrate radially to the cortex by somal transloca- been suggested based upon a variety of methods, includ- tion. Other neuroblasts migrate towards the surface of the ing genetic studies, neurochemical studies, animal models cortex following radial glia fibres and terminate their and a variety of functional imaging techniques. Here we migration by splitting the preplate into upper and lower will focus on direct studies of cerebral cortical structure, halves. In addition, there are significant numbers of as well as those studies that have somewhat clear implica- neuroblasts migrating from the ganglionic eminences, tions for microanatomical organization in ASD. which are transitory structures located within the ventral portion of the developing neural tube. Importantly, these Cortical volume and thickness across the lifespan cells migrate tangentially and are the source of a large proportion of the GABAergic neurones within the adult Although volume, thickness and surface area of the cortex [48,49]. cortex are coarse measures of brain structure, these The upper half of the preplate will become layer I, or the metrics are partially determined by the underlying marginal zone, which is occupied by reelin-containing microanatomical structure of the cortical sheet and, as Cajal–Retzius cells, neurones that play a role in terminat- such, are often used to make inferences regarding the ing neuroblast migration and guiding the organization of internal organization of the cortex. Studies of cortical the cortex. The lower half of the preplate forms a diffuse volume in autism have produced a variety of results that area commonly referred to as the subplate. appear to be largely attributable to the age of the subject Newly arriving neuroblasts aggregate in the middle of groups. While some studies report at least some type of the preplate and form the developing cortical plate. The increased brain volume in ASD [59–66], others have first neuroblasts to arrive will form the lower layers of reported no differences [67] or even decreases in volume the cortex, and the last will form layer II, with newly [68,69]. These seemingly mixed results can be attributed arriving neuroblasts migrating past older cortical not only to variability within the spectrum and to the plate neurones to occupy the superficial boundary of regions examined, but also to the age of subject groups. the cortical plate. Speciation of the deeper cortical For example, two of the studies reporting decreased laminae occurs within the ventricular zone, while the volume only found reductions within the grey matter of subventricular zone generates the neurones of the upper the temporal lobes [68,70], and those studies reporting laminae [50]. This sequence underlies the ‘inside-out’ increased brain volume have done so largely in younger migration pattern that characterizes the cortex and sets samples. Work on larger samples that takes into account in motion a series of maturational events that all occur the subject’s age indicates that newborns who will go on to in an inside-out fashion within the cortical plate. carry an ASD diagnosis have an average head circumfer- Pyramidal cells form the core of the resulting cell ence equivalent to NT newborns. Longitudinal imaging columns and have been described as the ‘fundamental studies show that by 6 months of age, children who sub- unit’ of cortical organization, while GABAergic cells coa- sequently went on to be diagnosed with an ASD have lesce around this core (see Marin-Padilla and other increased head circumferences and an excess of extra- several authors [41,51,52] for reviews). The specifics of axial fluid, especially around the frontal lobes [71], and by cortical column organization likely differ from cortical 2 to 3 years have increased cerebral volumes [64,72–75]. region to cortical region as a function of an area’s com- Surprisingly, the magnitude of these effects is predictive of putational demands, and, as might be expected, between the eventual severity of the disorder [71]. Longitudinal various mammalian species [40,53,54]. studies that include older subjects indicate that there is a

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 118 J. J. Hutsler and M. F. Casanova rapid early growth of the brain at around 12 months of specific reductions in neurone density (layer III) have been age that decelerates near adolescence [75–77], and two found, along with a reduction in neurone number in mul- recent meta-analyses of this large literature confirm this tiple layers (III, V and VI) and decreased neurone volumes general pattern [78,79]. Because brain growth slows rela- in the deep layers (layers V and VI [86]). These findings are tive to NT cases near adolescence, young adults tend to especially notable, due to the involvement of this cortical have average brain volumes that do not differ from control region in facial recognition. More recently, Oblak et al. cases [72,80,81]. The group as a whole does show more [87] looked for abnormalities within the posterior variability than NTs and studies of head circumference cingulate gyrus and the fusiform gyrus in an ASD sample. suggest that ASD groups appear to have a higher rate of These regions were selected because they process informa- macrocephaly [82]. It has been suggested that this differ- tion associated with both face and emotion recognition. ential brain growth trajectory may be a hallmark of These authors found no qualitative abnormalities within the neuropathologies that characterize the brain of ASD the fusiform gyrus, but the posterior cingulate cortex had individuals. a poorly defined layer IV in half of the ASD subjects exam- Studies investigating cortical thickness in ASD have ined. Neurones in this study were also found to be irregu- produced mixed results. A handful of studies have larly distributed, with an abnormally high density within reported increases in thickness [74], while others have the white matter. reported no differences [35,64] or decreased thickness One of the most commonly observed cortical anomalies [83–85]. These different results can also partially be in ASD is the presence of supernumerary neurones attributed to the age of subjects, as investigations report- located within the white matter just below the cortical ing no differences or decreases in cortical thickness have sheet. Bailey et al.’s [15] study describes increased number been performed on adolescent to adult-aged subjects of neurones within the white matter in four of six cases, [35,83–85], while some of the studies that focused on and a poor differentiation between grey and white matter younger groups have reported increased cortical thickness in their youngest subject. These white matter abnormali- [74]. Cortical thickness is not a static measure across the ties varied in location across subjects and quantitative lifespan. Hardan et al. [60] performed a longitudinal study counts did not reveal differences between ASD and NT using magnetic resonance imaging (MRI) and reported subjects [15]. Quantification of the transition between greater decreases in cortical thickness over time in ASD grey and white matter in Brodmann’s areas 9, 22 and 7 subjects as compared with NTs, supporting Hutsler et al.’s demonstrates a less abrupt transition in comparison with [35] findings in post mortem tissue in which thickness NT controls in all three locations [88]. Some of the vari- measures decreased as age increased at a rate greater than ability between MRI studies of cortical thickness and cor- that found in NTs. tical volume may be attributed to this poorly differentiated Although thickened cortex may suggest migration grey–white matter boundary. An indistinct transition problems and cell density alterations, a quantitative zone may contribute to variability between studies, as the examination of the cortical layers in ASD has shown a lower boundary of the cortex may not be reliably placed. similar thickness for all six cortical layers [35]. However, qualitative examination in the same study identified a Cortical column structure small increase in the incidence of cell clustering as well as supernumerary neurones within layer I and the subplate. With respect to columnar organization, Casanova et al. Another qualitative investigation performed by Bailey and [89] discovered that ASD subjects possess an increased colleagues [15] found that out of six ASD brains exam- density of minicolumns within layer III of the cortex in ined, three of them possessed some kind of disturbance to both temporal and frontal regions. In addition to their their lamination pattern. Of these three cases, two pos- increased density, these minicolumns are of smaller size sessed irregularities in the frontal region and the other in and possess a greater dispersion of neurones (Figure 2). the temporal. Except in specific cases, laminar irregulari- The parameters used in this study were corrected for ties do not appear to be a defining characteristic of the minicolumnar fragmentation and modelled against cortex in ASD [35] (see Figure 1). Quantitative studies stereology [91]. have also revealed location-specific alterations of neurone Compartmentalization of the minicolumns has re- volume and distribution. Within the fusiform gyrus, layer- vealed that most of the size difference in autistic

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 119

Figure 1. Photomontages of the cortical layers within a neurotypical (NT) subject (a) and an individual with autism spectrum disorder (ASD) (b). In each image, a portion of the grey–white matter boundary is indicated by a rectangle and these regions are shown magnified in panels c (NT) and d (ASD). Although layer thickness varies in these two images, in individuals with ASD, the lower boundary of layer VI is somewhat indistinct, with supernumerary cells extending into the subcortical white matter. Scale bars = 250 μm.

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 120 J. J. Hutsler and M. F. Casanova

Figure 2. Methodology employed by Casanova et al. [90]. The micrograph shows part of Brodmann area 9 from the brain of an 8-year-old male on the autism spectrum. Lamina III has been divided into quadrants to illustrate three different approaches applied to the same tissue. Upper left: the Nissl stained tissue, after correcting for inhomogeneity in the background. Upper right: minicolumn method. The spanning trees of minicolumnar fragments identified by computer image analysis. Lower left: Boolean germ-grain model. Germ points are indicated in red, and grains (that is, cell cross-sections) are outlined. Lower right: grey level index (GLI). Contours of the grey level index map are coloured such that lighter colouring indicates greater GLI. Scale bar = 200 μm.

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 121 individuals is due to an abnormality of the peripheral tional geometry [97]. This applied method, called neuropil space [89]. The minicolumnar core space houses Delaunay triangulation, generated a Gaussian bimodal all of the pyramidal cells, while a significant portion of distribution of interneuronal distances wherein the interneurones are located within the peripheral neuropil higher value mode representing the longer edges (that is, space. This distribution of pathological changes has sug- intercolumnar distances) was decreased within the series gested the presence of an excitatory–inhibitory imbalance of autistic individuals. No significant changes were in autism capable of explaining some of the seizures and reported for the lower value mode distribution (Figure 3). sensory phenomena observed in this condition [92]. These findings imply that the increases in cell density Buxhoeveden et al. [93] provided corroborating evidence reported in autism are due to the close apposition of nar- by identifying decreased columnar width in frontal rower minicolumns, but otherwise, the total number of regions of layer III in ASD subjects as well, but did not find pyramidal cells within the core of minicolumns remains the same abnormalities within primary visual cortex, sug- the same. gesting that specific areas are differentially affected (see The minicolumnopathy of autism was further charac- also Casanova et al. [94]). terized by computerized image analysis in nine separate Photomicrographs from the aforementioned series neocortical areas (BA 3b, 4, 9, 10, 11, 17, 24, 43 and 44 [95,96] have also been thresholded using the grey level [90]). Each area was assessed at supragranular, granular index (GLI). In autistic individuals, cluster segments of and infragranular levels. Smaller minicolumns were GLI maxima, representing the intercolumnar distance noted in autistic individuals with the difference in mag- between pyramidal cells, were decreased and mean nitude varying according to brain parcellation. The number per image frame was increased. The findings indi- authors concluded that the diminution in minicolumnar cate a greater number of narrower minicolumns in the width across all laminar subdivisions reflected involve- examined image frames. In addition, GLI segment ampli- ment of shared anatomical constituents among the dif- tude was increased, while mean GLI values showed no ferent layers. significant differences, suggesting that the cell bodies of A recent post mortem study employing stereological pyramidal cells may be smaller in autistic individuals. methods found increased numbers of neurones within A multinational study attempting to reproduce the the PFC of ASD children [76], as might be predicted minicolumnar findings was conducted in an independent from studies of decreased intercolumnar distances [94]. case series with the investigators doing the morphometric Although this may logically be assumed to correlate with analysis blind to diagnosis [94]. Minicolumnar width from increased brain volume and cortical thickness, neurone Brodmann areas 17, 4, 9 and S1, determined by both a number and percent deviation from normative brain segmentation algorithm and GLI values, indicated nar- weight were highly correlated only in NT subjects, but not rower minicolumns in autistic individuals as compared in the ASD group [98]. Additional studies within the infe- with NTs. In order to directly assess pyramidal cell size and rior frontal cortex show no change in neuronal density, density, the investigators applied a Boolean morphometric but a decrease in pyramidal cell size. These areas are model to the tissue sections [97]. Multivariate analysis notable for their involvement in social communication yielded significant diagnosis dependence, suggestive of [99]. smaller pyramidal cells and increased neuronal densities Taken together, these results suggest that cortical devel- in autistic individuals. The bias towards smaller pyramidal opment, and the resulting cortical structure, is altered in cells may underlie the presence of supernumerary short specific ways in autism. The vast majority of cells migrate corticocortical projections (for example, arcuate fibres) at into their appropriate positions within the developing cor- the expense of longer ones (for example, commissural pro- tical plate, and there is only limited evidence of misplaced jections; see Cortical Connectivity in ASD [94]). neuronal cell bodies and other alterations to the orderly In order to investigate whether the increased cell laminar structure. Furthermore, the columnar structure density in ASD was due to the apposition of narrower of the cortex suggests there may be an increased number minicolumns or to an increased number of neurones of cortical columns in autism that are more closely spaced within the core of the minicolumns, investigators applied than in NTs, and that their internal construction, as an algorithm that determined the distribution of dis- reflected in the quantitative distributions of neurones, tances between image frame kernels using computa- suggests an altered intracolumnar organization. This

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 122 J. J. Hutsler and M. F. Casanova

Figure 3. Summary findings of Casanova et al. [97]. Raw measurements from matched autism spectrum disorder (ASD)/control pairs were converted to relative differences by dividing their respective difference (ASD− control) by their mean (ASD+ control)/2. Box plots of the relative differences are plotted on a percentage scale and three analytical approaches are illustrated. Using minicolumnar morphometry (top right), ASD subjects demonstrate narrower column widths and closer cell spacing while showing similar variability in column width relative to neurotypical controls. A Boolean-germ-grain model (lower left) suggests smaller grain sizes (Nissl profiles) and higher density measures (gamma). Finally, a grey level index analysis (lower right) that shows the peaks and troughs of a bell-shaped kernel with the long axis positioned parallel to the axes of the cell columns. Again, ASD subjects showed shorter distances between amplitude peaks, narrower half amplitude widths and lower amplitudes relative to neurotypical controls. increase in density of cell columns suggests early prenatal Local overconnectivity in ASD abnormalities in the ontogeny of cortical structures. Indeed, morphometric analysis of cell size in focal Several types of evidence have been used to suggest that dysplastic areas of the cerebral cortex in autism suggests a local overconnectivity may be a feature of the ASD brain. desynchronization in the maturation of radially migrat- In order to maintain the level of connectivity within ing neuroblasts (future pyramidal cells) from those that the NT brain, the increased density of minicolumns migrate tangentially (future interneurones [100]). described in ASD would require an exponential increase in the fibre number and associated white matter volume [95,103,104]. Hofman [104] suggests that the actual Cortical connectivity in ASD scaling of minicolumns and white matter follows a 3/2 The cerebral cortex is marked by protracted development power law, or an eightfold increase in the number of and numerous connectional alterations during the first connections for a fourfold increase in column number. few postnatal years [101,102]. One of the most enduring This additional white matter would consist primarily of hypotheses regarding cortical organization in the ASD short-range corticocortical projections, as long-range brain is that connectional alterations play a critical role in connectivity incurs increased metabolic demands and the behavioural phenotype. These connectional changes conduction times [103]. fall into two major categories: local overconnectivity and Increased white matter volumes just under the cortex long-range functional underconnectivity. have been used to suggest an increase in local, short-range

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 123 connections [65,105], while behaviourally, the possibility sities arise during development in ASD, one possible of local overconnectivity has been linked to hyper-arousal, hypothesis is that the environmentally driven synapse reduced selectivity and increased computational noise culling that occurs after spines typically reach their peak [106]. More recently, genetic linkage studies have demon- density is impaired. Interestingly, fragile X syndrome, a strated the involvement of a number of synapse-related disorder that is highly comorbid with autism, also shows proteins in subgroups of ASD subjects including epac-2 increased spine densities and a high proportion of imma- and members of the neuroligin/neurexin family [107– ture spines [119]. 110]. Unfortunately, these specific genetic mutations have ASD subjects also show an increase in long, thin spine been found to be rare in the general ASD population types and a greater proportion of spines with heads [111]. (Hutsler et al., in prep.; see Figure 4b). Mature, stable In the developing human brain, primary sensory spines tend to be short and compact, while long, thin regions reach a maximum synaptic spine density around spines are commonly thought to be immature and poten- 1 year postnatally, while other cortical regions can take as tially unstable [110,120]. In experimental animal studies, long as 3 to 4 years to reach their peak [112]. This peak in spine lengths have been found to shorten following pro- density is followed by a long and protracted decrease in longed stimulation [121], reflecting a stabilized synaptic spine density that has been attributed to the removal of connection. spines as cortical networks mature. The organization of cerebral cortical connectivity is highly plastic, and the Long-range underconnectivity in ASD strengthening and weakening of individual synapses are an activity-dependent process that has been demonstrated An associated connectional change within the cerebral repeatedly by studying postnatal environmental perturba- cortex of individuals with ASD that involves functional tion [113]. This extensive literature indicates that NT underconnectivity between distant cortical locations has brain development is a constructive process that relies also been hypothesized [122–127]. Several types of evi- upon information and constraints present in the postnatal dence have been cited in support of long-range connec- environment [114,115]. In the context of ASD, a tional changes within the cortex of ASD individuals. neuroconstructive viewpoint suggests that minor biologi- Functional imaging studies have shown a reduction in cal alterations can have large effects on the developmental coordinated activity between distant cortical regions complexity of the cerebral cortex. This is due not only to [126] and an absence of the top-down modulation of the cumulative effects of biological cascades, but also to early sensory processing that is found in typically develop- the alteration of the very nature of the gene/environment ing individuals [124]. Behaviourally, long-range connec- interaction required for continued cortical construction tional changes may be linked to difficulties with and maintenance [116,117]. integration of information into its wider context for Connectional alterations on individual cortical pyrami- higher level meaning, a problem characteristic of ASD dal cells in ASD are evident in alterations to spine densities [124,128]. According to Just et al. [126], functional and spine morphologies [118]. ASD subjects have underconnectivity would be measured by the time corre- increased spine densities on pyramidal neurones in corti- lation or synchrony of cortical network activity. In addi- cal regions that include the frontal, temporal and parietal tion, they argue that the core deficits of autism are those lobes (Figure 4). These spine density increases are largely that require a high degree of coordination between distant restricted to superficial layer II, the last cortical layer cortical regions. Electrophysiological recordings demon- to mature during postnatal development. In addition, strate reduced coherence across electrode sites during rest layer II pyramidal cells are predominately involved in in both children and adults with ASD [129,130]. Toddlers interconnectivity between cortical regions within a hemi- with ASD (age 12 to 46 months) also show abnormally sphere. Although spine density increases were found in low interhemispheric synchronization and, most interest- eulaminate isocortex taken from the frontal, temporal and ingly, the strength of interhemispheric synchronization parietal lobes, the largely layer-specific finding of spine correlates positively with early verbal abilities and nega- density increases argues against an explanation based tively with autism severity [131]. solely on a global change to synapse-relevant proteins. Finally, post mortem examinations of frontal lobe white Although it is currently unclear how increased spine den- matter in a small number of ASD cases have shown

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 124 J. J. Hutsler and M. F. Casanova

Figure 4. Dendritic spines from layer II of an autism spectrum disorder (ASD) (a, b) and a neurotypical (c) subject. Spine densities on apical dendrites (a) were found to be approximately 30% higher than those found in neurotypical control cases [118]. Photomontages of dendritic spines in ASD (b) indicate a greater preponderance of spines with clearly defined synaptic heads relative to neurotypicals (c). Scale bars = 10 μm.

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 125 changes in axon populations within frontal lobe regions as distribution, density and composition from areal measure- well as alterations in myelin [132]. These studies point ments alone. Changes to cross-sectional area can occur directly to reductions in the largest fibres, which likely without changes to total fibre numbers, either by altering correspond to fast-conducting, long-range connections, the diameter or spacing of individual fibres. Aboitiz et al. and also report an excessive number of thin axons just [133] argue that the presence of large and small diameter below the anterior cingulate cortex that may further fibres makes it difficult to establish a relationship between support the presence of short-range overconnectivity CC area and overall fibre numbers (also see Lamantia and within the frontal lobes. Although further studies of the Rakic [145]). microanatomy of the white matter in ASD are needed to A wealth of recent diffusion tensor imaging (DTI) evi- assess the structural basis of functional underconnectivity, dence also indicates that the fibre content within the CC is the corpus callosum (CC) may provide an ideal location to significantly altered in ASD. DTI studies have revealed examine this issue because of its predictable fibre orienta- smaller CC volumes, low anisotropy and increased radial tion and lack of axons that form local connections. diffusivity in ASD groups [146–148]. These measures In typically developing subjects, the CC contains may reflect underlying changes to myelin and the density approximately two hundred million interhemispheric and size of axons [149]. Interestingly, measures of axons that are distributed among a range of diameters and anisotropy may also be related to measures of social that can be myelinated or unmyelinated [133]. These impairment [150]. These alterations are somewhat subpopulations of axons include a small number of very regionally specific, with the most notable changes being fast fibres, which likely play an important role in bringing associated with the anterior third of the CC [149]. Devel- distant cortical regions into synchronization [134], as well opmental DTI studies in ASD also show increased as large populations of medium and slow-communicating anisotropy in several major white matter tracts and fibres that may be responsible for more detailed informa- suggest an early and abnormal maturation of the white tiontransfer[135].Inaddition,becausetheCCisorganized matter in children less than 5 years of age. In these topographically from anterior to posterior, alterations in studies, the spatial pattern of white matter alterations is specific regions of the callosum would indicate whether noticeably different from those reported in adolescent and specific cortical regions are preferentially impacted adult studies [151]. Together, these results provide struc- by alterations in interhemispheric communication tural evidence for disrupted cortical connectivity in [133,136]. autism that is evident at very young ages. Magnetization In ASD, the CC has been studied primarily with struc- transfer imaging, an MRI technique sensitive to the pres- tural imaging methods aimed at examining its size ence of myelin, has also revealed abnormal myelin devel- [136,137], shape [138,139] and measures of diffusion opment in the CC, which may contribute to altered that can localize fibre bundles and their degree of interhemispheric connectivity in autism [152]. anisotropy. In a study by Hardan et al. [137], it was noted Studies examining developmental disconnection syn- that the anterior regions of the CC, such as the genu, were dromes, such as agenesis of the corpus callosum (ACC), significantly smaller in a group of non-mentally retarded have also demonstrated consistent behavioural pheno- individuals with ASD. Evidence from functional MRI and types that appear to arise from CC impairment. Children behavioural studies has suggested that cognitive functions with ACC and children with ASD show similar social defi- associated with the anterior regions are specifically cits, such as problems maintaining conversations, devel- impaired in individuals with ASD. Some examples include oping peer relationships and understanding social deficits in working memory [140], executive functioning etiquette [153]. Based on narratives of pictures depicting [141], theory of mind abilities [142] and the inability to people in various situations, individuals with ACC also suppress inappropriate responses [143]. The posterior show deficiencies imagining and inferring the mental, regions of the CC have also been reported to show a reduc- emotional and social functioning of others [154]. Com- tion in area [136,144]. Despite the presence of these gross munication, a hallmark deficit in autism, is abnormal in morphological alterations, little is known about the types all cases of ACC [155] and a large portion of ACC subjects or quantities of fibres present within the callosum in this exhibit echolalia and the meaningless repetition of words population. Structural MRI has excellent spatial resolu- and phrases – behaviours that also frequently occur in tion, but it is impossible to make inferences about fibre autism populations [153]. Individuals with autism have

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 126 J. J. Hutsler and M. F. Casanova problems accurately and rapidly shifting attention [156], ated cortical neurones. Golgi studies of both human and a deficit also characteristic of ACC [157]. As with ASD, the non-human primates have identified two main morpho- combination of these deficits impacts all aspects of devel- logical classes of these so-called ‘interstitial’ neurones: opment for these populations. fusiform and polymorphic [159]. Polymorphic cells Despite a great deal of converging evidence regarding occupy the white matter compartment directly subjacent this long-range underconnectivity, many fundamental to cortical layer VI; however, as development proceeds, issues remain unresolved. Does functional under- fusiform cells become the dominant cell type throughout connectivity represent an actual loss of connectional the white matter. Programmed apoptosis, which removes fibres? If so, are specific axonal populations affected? Func- many of these subplate neurones during development, is tional underconnectivity could also be an epiphenomenon thought to underlie the transition to a greater representa- of local overconnectivity or the result of alterations to tion of fusiform cells. signal gating mechanisms that exist within the cortex. Thymidine autoradiography in foetal cats demonstrates that these white matter neurones are generated for only a few days around embryonic day 28 preceding the genera- Potential alterations to the cortical subplate tion of the earliest neurones that will form the cortical Both qualitative [15,35,87,158] and quantitative [88] plate (layer VI [160]). Such birth-dating studies also dem- studies have described excess cells within the subcortical onstrate that marginal zone neurones (layer I) are gener- white matter in ASD. Although this abnormality has been ated at the same time as subplate neurones and, as such, found at a number of cortical locations and the cortical belong to a single cell population that is split by the forma- subplate has been a focus of study in other neuro- tion of the cortical plate. Many of the white matter neu- pathological conditions such as schizophrenia, to date, its rones that persist into adulthood are vestiges of the role in ASD has been underexplored. The subplate in the developmental subplate zone, and the marked develop- adult arises during development from an earlier precursor mental decrease of these cells is attributable to apoptosis to the cortical plate, a structure variously called the rather than dilution [161]. preplate or the primordial plexiform layer [32,45]. This Probably, the most studied aspect of subplate neurones structure is widely thought to play a critical role in the is the critical role they play during development to estab- spatial organization of cortical neurones as they arrive in lish thalamocortical and corticothalamic connections. the preplate, as well as the establishment of both cortical For thalamocortical projections, the subplate serves a afferents and efferents. In NT subjects, the vast majority of temporary termination site early in development. Using subplate cells undergo apoptosis after the formation of the transneuronal tracing and H-thymidine birth-dating, cortical plate; however, many remain, and it has been sug- Shatz and Luskin [162] demonstrated an accumulation gested that they play an important role in gating cortical of these thalamic afferents within the subplate during a inputs as well as driving synchronous activity between time in which the laminae of the cortex are still imma- cortical locations. Excess neuronal cell bodies within the ture. It is not until approximately 2 weeks later that these subplate of ASD subjects could either be the result of the fibres finally invade their targets in both cortical layers IV overproduction of subplate neurones or reduced apoptosis and VI. of subplate neurones, or these profiles may be the result of After the discovery that subplate neurones serve as tem- radial or tangential migration problems. In primate porary targets for incoming thalamic afferents, McConnell species, a mix of excitatory and inhibitory neurones et al. [163] explored the role subplate neurones play in migrate radially from the ventricular and subventricular guiding efferent connections from the cerebral cortex. zones, while tangentially migrating neurones from the Using anterograde tract tracing, these researchers demon- ganglionic eminence are thought to be only inhibitory. strated that cortical axons project to the thalamus rela- tively late in prenatal development (around E50 in the cat) after subplate neurones have already established their The role of subplate neurones during thalamic connections. To examine the influence subplate neurodevelopment neurones have on pioneering the cortical-thalamic As noted, neurones of the cortical subplate originate from pathway from visual cortex, kainic acid injections were the preplate and, as such, are among the earliest gener- made within the subplate zone during development.

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 127

Subsequently, deep cortical layer neurones were labeled the excess white matter neurones reported in autism rep- with H-leucine to determine if cortical projections to the resent neuropathology in the early development of the thalamus (and other subcortical targets) were affected by subplate. As suggested earlier, these profiles could be subplate ablation. In approximately half of the animals neurones originally destined for the cortical plate that receiving subplate lesions, pathways from the visual failed to reach their destination. Despite this, its potential cortex to subcortical targets including the thalamus were involvement in other neurological disorders, along with disrupted. its influence on the spatial distribution of neurones The subplate also appears to play an important role in within the cortex and long-range cortical connections, establishing the organization of the neurones in the over- suggests that further studies are warranted. Of course, lying cortical plate. In visual cortex, kainic acid lesions tracking this cell population during development, well administered to the subplate after thalamic axons have ahead of any potential diagnosis of ASD, will be a diffi- invaded layer IV,but prior to the segregation of axons into cult undertaking, but several approaches could help to ocular dominance patches, disrupt their subsequent for- determine the origins and appearance of this cell group. mation [164,165]. These same studies also found that First, morphological and neurochemical characterization subplate lesions resulted in long-term changes to the of these neurones in post mortem materials can clarify cytoarchitecture of layer IV and inappropriate thalamic whether these cells are consistent with the characteris- axon terminations within layers II and III. tics of subplate neurones that have been so well described in both animal and human studies. Secondly, longitudinal pre- and postnatal imaging studies can be The function of persistent subplate neurones in used to quantify and describe the subplate compartment the adult brain and how it changes over time [170]. Retrospective use of While many studies have shown that a large majority of such materials following diagnosis in early childhood subplate neurones undergo programmed cell death after could help to identify atypical developmental trajectories executing their roles during neurodevelopment, it is clear within this region and might inform how the NT that a subpopulation of these neurones persists in the subplate population may be altered in ASD. Lastly, if adult brain and become embedded in the mature circuitry subplate ontogeny is affected in autism, it is difficult to of the cerebral cortex [166]. Relative to the rodent fully interpret the neurobiological effects of these subplate, more interstitial neurones are present in primate changes without an animal model that can inform how species and appear to occupy a larger zone beneath the these alterations might be related to changes in cortical cortex [139,167]. architecture and connectivity. In a review of the functional properties of subplate Although the consistency of certain neuropathological neurones in mammals, Luhmann et al. [168] suggest findings has greatly bolstered our understanding of corti- that subplate neurones serve as ‘cortical amplifiers’ that cal organization in the ASD brain over the past two drive synchronous activity during neurodevelopment. decades, these advances have depended upon the hard Importantly, the ability to act as amplifiers may be lost as work of brain banking post mortem materials. These the cerebrum matures, but abnormalities within the efforts have been hampered by a number of practical limi- subplate population may have a lasting impact on subse- tations. First, much of the published research on post quent cortical physiology. The view that subplate neu- mortem ASD cases has relied on a limited number of cases rones coordinate oscillatory network activity is further that have been made available for study. Coupled with the supported by Suarez-Sola et al. [167] who describe inherent variability in phenotype, inadequate and incon- species differences in subplate neurone populations and sistent contextual information regarding the history of speculate that these neurones may play a role in coordi- individual cases (especially in idiopathic cases) and prob- nating cortical network activity. In the same vein, lems with adequate and consistent tissue preservation are Kostovic et al. [169] describe subplate neurones as currently slowing progress in the field. These shortcom- network ‘gatekeepers’ between frontal and limbic areas, ings have undermined our confidence regarding the and propose that increases in the number of subplate veracity of the data thus far acquired. Future efforts to neurones are an important contributor to the patho- minimize these methodological problems must be under- physiology of schizophrenia. Currently, it is unclear if taken if we are to continue to move forward in the field.

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 128 J. J. Hutsler and M. F. Casanova

at drawing bridges between the computational circuitry of Conclusions the cerebral cortex and the atypical behaviours that can ASD has been described as a disorder of brain structure, arise when that circuitry is modified. brain growth, synaptic alterations, and altered short- and long-range cortical connectivity. Despite a recent wealth of information regarding pathological changes in ASD, Acknowledgement there is still much remaining to be learned about the The authors would like to thank Andrew E. Switala for his organization of the cerebral cortex in autism. Several con- invaluable assistance with Figures 2 and 3. clusions are, however, clear and have important relevance for our understanding of early interventions. The ASD phenotype may be associated with alterations to cortical References structure that begin in the prenatal and perinatal period. 1 Blazek V, Bruzek J, Casanova MF. Plausible mechanisms Most notable in this regard are the findings of supernu- for brain structural and size changes in human evolu- merary cell profiles in the white matter and the possibili- tion. Coll Antropol 2011; 35: 949–55 ties that cell counts and column numbers may be altered 2 Buxhoeveden DP, Casanova MF. The minicolumn and in the cerebral cortex. These characteristics of organiza- evolution of the brain. Brain Behav Evol 2002; 60: tion in the adult brain are largely determined by early 125–51 3 Marin-Padilla M. The evolution of the structure of the events, including neurogenesis, neuronal migration and neocortex in mammals: a new theory of cytoar- subplate apoptosis. Other findings of structural alterations chitecture. Rev Neurol 2001; 33: 843–53 in individuals with ASD are likely to occur later in devel- 4 Marin-Padilla M. Development of the human cerebral opment. These may be secondary to prenatal abnormali- cortex. A cytoarchitectonic theory. Rev Neurol 1999; ties or could be part of the primary pathology of the 29: 208–16 5 Finlay BL, Darlington RB. Linked regularities in the disorder. For example, because of their activity-dependent development and evolution of mammalian brains. plasticity, changes in spine density and spine morphology Science 1995; 268: 1578–84 may be the result of early alterations to cortical structure 6 Dawson G. Early behavioral intervention, brain plastic- that subsequently alter an individual’s interaction with ity,and the prevention of autism spectrum disorder. Dev environmental influences that shape cortical circuitry. Psychopathol 2008; 20: 775–803 Conversely, mutations to synapse-relevant proteins that 7 Oberman LM, Pascual-Leone A. Cortical plasticity: a proposed mechanism by which genomic factors lead to are directly involved in the strengthening and weakening the behavioral and neurological phenotype of autism of connections could suggest a primarily genetic origin for spectrum and psychotic-spectrum disorders. Behav spine alterations. Finally, the establishment of long-range Brain Sci 2008; 31: 276–7 cortical connections has been shown to depend on 8 Karmiloff-Smith A. Development itself is the key to subplate neurones. The potential links between subplate understanding developmental disorders. Trends Cogn Sci 1998; 2: 389–98 organization and alterations to cortical connectivity and 9 Karmiloff-Smith A. Nativism versus neurocon- function in ASD remain speculative, but constitute an structivism: rethinking the study of developmental dis- exciting area of future exploration in the disease. orders. Dev Psychol 2009; 45: 56–63 In general, plasticity decreases with age, and its direct 10 Benavides-Piccione R, Ballesteros-Yanez I, de Lagran impact on the still developing circuitry of the cerebral MM, Elston G, Estivill X, Fillat C, DeFelipe J, Dierssen M. cortex may underlie the success of early and intensive On dendrites in Down syndrome and DS murine models: a spiny way to learn. Prog Neurobiol 2004; 74: 111–26 behavioural interventions in ASD. The open nature of the 11 Craig AM, Kang Y. Neurexin-neuroligin signaling in developing cortical system provides great hope for addi- synapse development. Curr Opin Neurobiol 2007; 17: tional treatment methods that would take advantage of 43–52 this early plasticity. An understanding of early brain 12 Bauman M, Kempter TL. Histoanatomic observations of development can inform the creation of new early behav- the brain in early infantile autism. Neurology 1985; 35: 866–74 ioural interventions, as well as a better ability to identify 13 Guerin P, Lyon G, Barthelemy C, Sostak E, Chevrollier V, those individuals at risk for autism long before the typical Garreau B, Lelord G. Neuropathological study of a case time of a first diagnosis. This potential for improved treat- of autistic syndrome with severe mental retardation. ments is a strong impetus to push forward with our efforts Dev Med Child Neurol 1996; 38: 203–11

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 129

14 Kemper TL, Bauman M. Neuropathology of infantile 34 Casanova MF. The significance of minicolumnar size autism. J Neuropathol Exp Neurol 1998; 57: 645–52 variability in autism. In Autism: Current Theories and Evi- 15 Bailey A, Luthert P, Dean A, Harding B, Janota I, dence. Ed. W Zimmerman. New York, NY: Humana Montgomery M, Rutter M, Lantos P. A clinico- Press, 2008; 349–60 pathological study of autism. Brain 1998; 121: 889– 35 Hutsler JJ, Love T, Zhang H. Histological and magnetic 905 resonance imaging assessment of cortical layering and 16 Happe F. Autism: cognitive deficit or cognitive style? thickness in autism spectrum disorders. Biol Psychiatry Trends Cogn Sci 1999; 3: 216–22 2007; 61: 449–57 17 Happe F, Frith U. The weak coherence account: detail- 36 Opris I, Fuqua JL, Huettl PF, Gerhardt GA, Berger TW, focused cognitive style in autism spectrum disorders. J Hampson RE, Deadwyler SA. Closing the loop in Autism Dev Disord 2006; 36: 5–25 primate prefrontal cortex: inter-laminar processing. 18 Simmons DR, Robertson AE, McKay LS, Toal E, McAleer Front Neural Circuits 2012; 6:88 P, Pollick FE. Vision in autism spectrum disorders. Vision 37 DeFelipe J. Reflections on the structure of the cortical Res 2009; 49: 2705–39 minicolumn. In Neocortical Modularity and the Cell 19 Hughes JR. A review of Savant Syndrome and its possi- Minicolumn. Ed. MF Casanova. Hauppauge, NY: Nova ble relationship to epilepsy. Epilepsy Behav 2010; 17: Science, 2005; 57–92 147–52 38 DeFelipe J, Lopez-Cruz PLL, Benavides-Piccione R, Bielza 20 Baron-Cohen S. The extreme male brain theory of C, Larranaga P, Anderson S, Burkhalter A, Cauli B, autism. Trends Cogn Sci 2002; 6: 248–54 Fairen A, Feldmeyer D, Fishell G, Fitzpatrick D, Freund 21 Amaral DG, Schumann CM, Nordahl CW. Neuro- TF, Gonzalez-Burgos G, Hestrin S, Hill S, Hof PR, Huang anatomy of autism. Trends Neurosci 2008; 31: 137–45 J, Jones EG, Kawaguchi Y, Kisvarday Z, Kubota Y, Lewis 22 Casanova MF. The neuropathology of autism. Brain DA, Marin O, Markram H, McBain CJ, Meyer HS, Pathol 2007; 17: 422–33 Monyer H, Nelson SB, Rockland K, Rossier J, Rubenstein 23 Kemper TL. The developmental neuropathology of JLR, Rudy B, Scanziani M, Shepherd GM, Sherwood CC, autism. In The Neurochemical Basis of Autism. Ed. GJ Staiger JF, Tamas G, Thomson A, Wang Y, Yuste R, Blatt. Boston, MA: Springer, 2010; 69–82 Ascoli GA. New insights into the classification and 24 Lam KSL, Aman MG, Arnold LE. Neurochemical corre- nomenclature of cortical GABAergic interneurons. Nat lates of autistic disorder: a review of the literature. Res Rev Neurosci 2013; 14: 202–16 Dev Disabil 2006; 27: 254–89 39 Jones EG. Microcolumns in the cerebral cortex. Proc Natl 25 Pardo CA, Eberhart CG. The neurobiology of autism. Acad SciUSA2000; 97: 5019–21 Brain Pathol 2007; 17: 434–47 40 Jones EG, Rakic P. Radial columns in cortical architec- 26 Penn HE. Neurobiological correlates of autism: a review ture: it is the composition that counts. Cereb Cortex of recent research. Child Neuropsychol 2006; 12: 57–79 2010; 20: 2261–4 27 Buxhoeveden DP, Casanova MF. The minicolumn 41 Marin-Padilla M. The pyramidal cell and its local-circuit hypothesis in neuroscience. Brain 2002; 125: 935–51 interneurons: a hypothetical unit of the mammalian 28 Mountcastle VB. The columnar organization of the cerebral cortex. J Cogn Neurosci 1990; 2: 180–94 neocortex. Brain 1997; 120: 701–22 42 Peters A, Jones EG. Classification of cortical neurons. 29 Mountcastle VB. Modality and topographic properties of Cereb Cortex 1984; 1: 107–21 single neurons of cat’s somatic sensory cortex. J 43 Szentagothai J. The Ferrier Lecture, 1977: the neuron Neurophysiol 1957; 20: 408–34 network of the cerebral cortex: a functional interpreta- 30 Bourgeois JP. Normal, pathological and mendable tion. Proc R Soc London B Biol Sci 1978; 201: 219– synaptogenesis in the cerebral cortex. Psychiatr Sci Hum 48 Neurosci 2008; 6: 124–36 44 Bystron I, Rakic P, Molnar Z, Blakemore C. The first 31 Pascual-Leone A, Freitas C, Oberman L, Horvath JC, neurons of the human cerebral cortex. Nat Neurosci Halko M, Eldaief M, Bashir S, Vernet M, Shafi M, 2006; 9: 880–6 Westover B, Vahabzadeh-Hagh AM, Rotenberg A. Char- 45 Zecevic N, Milosevic A, Rakic S, Marin-Padilla M. Early acterizing brain cortical plasticity and network dynam- development and composition of the human primordial ics across the age-span in health and disease with TMS- plexiform layer: an immunohistochemical study. J Comp EEG and TMS-fMRI. Brain Topogr 2011; 24: 302–15 Neurol 1999; 412: 241–54 32 Bystron I, Blakemore C, Rakic P. Development of the 46 Marin-Padilla M. Cajal-Retzius cells and the develop- human cerebral cortex: Boulder Committee revisited. ment of the neocortex. Trends Neurosci 1998; 21: Nat Rev Neurosci 2008; 9: 110–22 64–71 33 Kornack DR, Rakic P. Radial and horizontal deployment 47 Caviness VS Jr, Takahashi T, Nowakowski RS. Numbers, of clonally related cells in the primate neocortex: rela- time and neocortical neuronogenesis: a general devel- tionship to distinct mitotic lineages. Neuron 1995; 15: opmental and evolutionary model. Trends Neurosci 311–21 1995; 18: 379–83

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 130 J. J. Hutsler and M. F. Casanova

48 Letinic K, Zoncu R, Rakic P. Origin of GABAergic 63 Hazlett HC, Poe MD, Gerig G, Smith RG, Piven J. Cortical neurons in the human neocortex. Nature 2002; 417: gray and white brain tissue volume in adolescents and 645–9 adults with autism. Biol Psychiatry 2006; 59: 1–6 49 Petanjek Z, Kostovic´ I, Esclapez M. Primate-specific 64 Hazlett HC, Poe MD, Gerig G, Styner M, Chappell C, origins and migration of cortical GABAergic neurons. Smith RG, Vachet C, Piven J. Early brain overgrowth in Front Neuroanat 2009; 3:26 autism associated with an increase in cortical surface 50 Tarabykin V, Stoykova A, Usman N, Gruss P. Cortical area before age 2 years. Arch Gen Psychiatry 2011; 68: upper layer neurons derive from the subventricular zone 467–76 as indicated by Svet1 gene expression. Development 65 Herbert MR, Ziegler DA, Deutsch CK, O’Brien LM, Lange 2001; 128: 1983–93 N, Bakardjiev A, Hodgson J, Adrien KT, Steele S, Makris 51 Bednar JA. Building a mechanistic model of the devel- N, Kennedy D, Harris GJ, Caviness VS. Dissociations of opment and function of the primary visual cortex. J cerebral cortex, subcortical and cerebral white matter Physiol 2012; 106: 194–211 volumes in autistic boys. Brain 2003; 126: 1182–92 52 Cuntz H. The dendritic density field of a cortical pyrami- 66 Palmen S, Pol HE, Kemner C, Schnack HG, Durston S, dal cell. Front Neuroanat 2012; 6:2 Lahuis BE, Kahn RS, van Engeland H. Increased gray- 53 Casanova MF, Trippe J, Tillquist C, Switala AE. matter volume in medication-naive high-functioning Morphometric variability of minicolumns in the striate children with autism spectrum disorder. Psychol Med cortex of Homo sapiens, Macaca mulatta, and Pan trog- 2005; 35: 561–70 lodytes. J Anat 2009; 214: 226–34 67 Hallahan B, Daly EM, McAlonan G, Loth E, Toal F, 54 Hutsler J, Galuske RAW. Hemispheric asymmetries in O’Brien F, Robertson D, Hales S, Murphy C, Murphy KC, cerebral cortical networks. Trends Neurosci 2003; 26: Murphy DGM. Brain morphometry volume in autistic 429–35 spectrum disorder: a magnetic resonance imaging study 55 Bauman ML, Kemper TL. The neuropathology of the of adults. Psychol Med 2009; 39: 337–46 autism spectrum disorders: what have we learned? In 68 Boddaert N, Chabane N, Gervais H, Good CD, Bourgeois Autism: Neural Basis and Treatment Possibilities. Eds G M, Plumet MH, Barthelemy C, Mouren MC, Artiges E, Bock, J Goode. Chichester: John Wiley & Sons, 2008; Samson Y, Brunelle F, Frackowiak RSJ, Zilbovicius M. 112–28 Superior temporal sulcus anatomical abnormalities in 56 Pickett J, London E. The neuropathology of autism: a childhood autism: a voxel-based morphometry MRI review. J Neuropathol Exp Neurol 2005; 64: 925–35 study. Neuroimage 2004; 23: 364–9 57 Schmitz C, Rezaie P. The neuropathology of autism: 69 McAlonan GM, Suckling J, Wong N, Cheung V, where do we stand? Neuropathol Appl Neurobiol 2008; Lienenkaemper N, Cheung C, Chua SE. Distinct patterns 34: 4–11 of grey matter abnormality in high-functioning autism 58 Wegiel J, Kuchna I, Nowicki K, Imaki H, Marchi E, Ma and Asperger’s syndrome. J Child Psychol Psychiatry SY, Chauhan A, Chauhan V, Bobrowicz TW, de Leon M, 2008; 49: 1287–95 Saint Louis LA, Cohen IL, London E, Brown WT, 70 Rojas DC, Bawn SD, Benkers TL, Reite ML, Rogers SJ. Wisniewski T. The neuropathology of autism: defects of Smaller left hemisphere planum temporale in adults neurogenesis and neuronal migration, and dysplastic with autistic disorder. Neurosci Lett 2002; 328: 237–40 changes. Acta Neuropathol 2010; 119: 755–70 71 Shen MD, Nordahl CW, Young GS, Wootton-Gorges SL, 59 Freitag CM, Luders E, Hulst HE, Narr KL, Thompson PM, Lee A, Liston SE, Harrington KR, Ozonoff S, Amaral DG. Toga AW, Krick C, Konrad C. Total brain volume and Early brain enlargement and elevated extra-axial fluid corpus callosum size in medication-naive adolescents in infants who develop autism spectrum disorder. Brain and young adults with autism spectrum disorder. Biol 2013; 136: 2825–35 Psychiatry 2009; 66: 316–19 72 Courchesne E, Pierce K, Schumann CM, Redcay E, 60 Hardan AY, Libove RA, Keshavan MS, Melhem NM, Buckwalter JA, Kennedy DP, Morgan J. Mapping early Minshew NJ. A preliminary longitudinal magnetic brain development in autism. Neuron 2007; 56: 399– resonance imaging study of brain volume and corti- 413 cal thickness in autism. Biol Psychiatry 2009; 66: 73 Courchesne E, Redcay E, Kennedy DP. The autistic 320–6 brain: birth through adulthood. Curr Opin Neurol 2004; 61 Hardan AY, Minshew NJ, Mallikarjuhn M, Keshavan 17: 489–96 MS. Brain volume in autism. J Child Neurol 2001; 16: 74 Hardan AY, Muddasani S, Vemulapalli M, Keshavan 421–4 MS, Minshew NJ. An MRI study of increased cortical 62 Hazlett HC, Poe M, Gerig G, Smith RG, Provenzale J, Ross thickness in autism. Am J Psychiatry 2005; 57: 10S– A, Gilmore J, Piven J. Magnetic resonance imaging 11S and head circumference study of brain size in autism – 75 Schumann CM, Bloss CS, Barnes CC, Wideman GM, birth through age 2 years. Arch Neurol 2005; 62: 1366– Carper RA, Akshoomoff N, Pierce K, Hagler D, Schork 76 N, Lord C, Courchesne E. Longitudinal magnetic reso-

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 131

nance imaging study of cortical development through spectrum disorders. Brain Res 2010; 1360: 138– early childhood in autism. J Neurosci 2010; 30: 46 4419–27 89 Casanova MF, Buxhoeveden DP, Switala AE, Roy E. 76 Courchesne E, Carper R, Akshoomoff N. Evidence of Minicolumnar pathology in autism. Neurology 2002; brain overgrowth in the first year of life in autism. JAMA 58: 428–32 2003; 290: 337–44 90 Casanova MF, El-Baz A, Vanbogaert E, Narahari P, 77 Webb SJ, Nalty T, Munson J, Brock C, Abbott R, Dawson Switala A. A topographic study of minicolumnar core G. Rate of head circumference growth as a function of width by lamina comparison between autistic subjects autism diagnosis and history of autistic regression. J and controls: possible minicolumnar disruption due to Child Neurol 2007; 22: 1182–90 an anatomical element in-common to multiple laminae. 78 Redcay E, Courchesne E. When is the brain enlarged in Brain Pathol 2010; 20: 451–8 autism? A meta-analysis of all brain size reports. Biol 91 Casanova MF, Switala AE. Minicolumnar morpho- Psychiatry 2005; 58: 1–9 metry: computerized image analysis. In Neocortical 79 Stanfield AC, McIntosh AM, Spencer MD, Philip R, Gaur Modularity and the Cell Minicolumn. Ed. MF Casanova. S, Lawrie SM. Towards a neuroanatomy of autism: a New York, NY: Nova Science, 2005; 161–80 systematic review and meta-analysis of structural mag- 92 Casanova MF, Buxhoeveden D, Gomez J. Disruption in netic resonance imaging studies. Eur Psychiatry 2008; the inhibitory architecture of the cell minicolumn: 23: 289–99 implications for autism. Neuroscientist 2003; 9: 496– 80 Courchesne E. Brain development in autism: early over- 507 growth followed by premature arrest of growth. Ment 93 Buxhoeveden DP, Semendeferi K, Buckwalter J, Retard Dev Disabil Res Rev 2004; 10: 106–11 Schenker N, Switzer R, Courchesne E. Reduced 81 Courchesne E, Karns CM, Davis HR, Ziccardi R, Carper minicolumns in the frontal cortex of patients with RA, Tigue ZD, Chism HJ, Moses P, Pierce K, Lord C, autism. Neuropathol Appl Neurobiol 2006; 32: 483– Lincoln AJ, Pizzo S, Schriebman L, Haas RH, 91 Akshoomoff NA, Courchesne RY. Unusual brain growth 94 Casanova MF, van Kooten I, Switala AE, van Engeland patterns in early life in patients with autistic disorder – H, Heinsen H, Steinbusch HWM, Hoff PR, Schmitz C. an MRI study. Neurology 2001; 57: 245–54 Abnormalities of cortical minicolumnar organization in 82 Lainhart JE, Piven J, Wzorek M, Landa R, Santangelo SL, the prefrontal lobes of autistic patients. Clin Neurosci Res Coon H, Folstein SE. Macrocephaly in children and 2006; 6: 127–33 adults with autism. J Am Acad Child Adolesc Psychiatry 95 Casanova MF, Buxhoeveden DP, Brown C. Clinical and 1997; 36: 282–90 macroscopic correlates of minicolumnar pathology in 83 Hadjikhani N, Joseph RM, Snyder J, Tager-Flusberg H. autism. J Child Neurol 2002; 17: 692–5 Anatomical differences in the mirror neuron system 96 Casanova MF, Buxhoeveden DP, Switala AE, Roy EL. and social cognition network in autism. Cereb Cortex Minicolumnar pathology in the brains of autistic and 2006; 16: 1276–82 Asperger’s patients. Neurology 2002; 58: A2–3 84 Scheel C, Rotarska-Jagiela A, Schilbach L, Lehnhardt 97 Casanova MF, Switala AE, van Engeland H, Heinsen H, FG, Krug B, Vogeley K, Tepest R. Imaging derived corti- Steinbusch HWM, Hof PR, Trippe J, Stone J, Schmitz cal thickness reduction in high-functioning autism: key C. Minicolumnar abnormalities in autism. Acta regions and temporal slope. Neuroimage 2011; 58: 391– Neuropathol 2006; 112: 287–303 400 98 Courchesne E, Mouton PR, Calhoun ME, Semendeferi 85 Wallace GL, Dankner N, Kenworthy L, Giedd JN, Martin K, Ahrens-Barbeau C, Hallet MJ, Carter-Barnes C, A. Age-related temporal and parietal cortical thinning Pierce K. Neuron number and size in prefrontal cortex in autism spectrum disorders. Brain 2010; 133: of children with autism. JAMA 2011; 306: 2001– 3745–54 10 86 Van Kooten IAJ, Palmen SJMC, von Cappeln P, 99 Jacot-Descombes S, Uppal N, Wicinski B, Santos M, Steinbusch HWM, Korr H, Heinsen H, Hof PR, van Schmeidler J, Giannakopoulos P, Heinsein H, Schmitz C, Engeland H, Schmitz C. Neurons in the fusiform gyrus Hoff PR. Decreased pyramidal neuron size in Brodmann are fewer and smaller in autism. Brain 2008; 131: areas 44 and 45 in patients with autism. Acta 987–99 Neuropathol 2012; 124: 67–79 87 Oblak AL, Rosene DL, Kemper TL, Bauman ML, Blatt GJ. 100 Casanova MF, El-Baz AS, Kamat SS, Dombroski BA, Altered posterior cingulate cortical cyctoarchitecture, Khalifa F, Elnakib A, Soliman A, Allison-McNutt A, but normal density of neurons and interneurons in the Switala A. Focal cortical dysplasias in autism spectrum posterior cingulate cortex and fusiform gyrus in autism. disorders. Acta Neuropathol 2013; 1:67 Autism Res 2011; 4: 200–11 101 Huttenlocher PR. Basic neuroscience research has 88 Avino TA, Hutsler JJ. Abnormal cell patterning at important implications for child development. Nat the cortical gray-white matter boundary in autism Neurosci 2003; 6: 541

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 132 J. J. Hutsler and M. F. Casanova

102 Tsujimoto S. The prefrontal cortex: functional neural 115 McAllister AM. Dynamic aspects of CNS synapse forma- development during early childhood. Neuroscientist tion. Annu Rev Neurosci 2007; 30: 425–50 2008; 14: 345–58 116 Sirois S, Spratling M, Thomas MSC, Westermann G, 103 Casanova MF. White matter volume increase and Mareschal D, Johnson MH. Precis of neurocon- minicolumns in autism. Ann Neurol 2004; 56: 453 structivism: how the brain constructs cognition. Behav 104 Hofman MA. Brain evolution in hominids: are we at the Brain Sci 2008; 31: 321–31 end of the road? In Evolutionary Anatomy of the Primate 117 Westermann G, Mareschal D, Johnson MH, Sirois S, Cerebral Cortex. Eds D Falk, KR Gibson. Boston, MA: Spratling MW, Thomas MS. Neuroconstructivism. Dev Cambridge University Press, 2001; 113–27 Sci 2007; 10: 75–83 105 Herbert MR, Ziegler DA, Makris N, Filipek PA, Kemper 118 Hutsler JJ, Zhang H. Increased dendritic spine densities TL, Normandin JJ, Sanders HA, Kennedy DN, Caviness on cortical projection neurons in autism spectrum dis- VS. Localization of white matter volume increase in orders. Brain Res 2010; 1309: 83–94 autism and developmental language disorder. Ann 119 Irwin SA, Patel B, Idupulapati M, Harris JB, Crisostomo Neurol 2004; 55: 530–40 RA, Larsen BP, Kooy F, Willems PJ, Cras P, Kozlowski 106 Belmonte MK, Yurgelun-Todd DA. Functional anatomy PB, Swain RA, Weiler IJ, Greenough WT. Abnormal of impaired selective attention and compensatory pro- dendritic spine characteristics in the temporal and cessing in autism. Brain Res Cogn Brain Res 2003; 17: visual cortices of patients with fragile-X syndrome: a 651–64 quantitative examination. Am J Med Genet A 2001; 98: 107 Jamain S, Quach H, Betancur C, Rastam M, Colineaux 161–7 C, Gillberg IC, Soderstrom H, Giros B, Leboyer M, 120 Holtmaat A, Trachtenberg JT, Wilbrecht L, Shepherd Gillberg C, Bourgeron T. Mutations of the X-linked GM, Zhang XQ, Knott GW, Svoboda K. Transient and genes encoding neuroligins NLGN3 and NLGN4 are persistent dendritic spines in the neocortex in vivo. associated with autism. Nat Genet 2003; 34: 27–9 Neuron 2005; 45: 279–91 108 Laumonnier F, Bonnet-Brilhault F, Gomot M, Blanc R, 121 Yuste R. Dendritic spines and distributed circuits. David A, Moizard MP, Raynaud M, Ronce N, Lemonnier Neuron 2011; 71: 772–81 E, Calvas P, Laudier B, Chelly J, Fryns JP, Ropers HH. 122 Cherkassky VL, Kana RK, Keller TA, Just MA. Func- X-linked mental retardation and autism are associated tional connectivity in a baseline resting-state network with a mutation in the NLGN4 gene, a member of the in autism. Neuroreport 2006; 17: 1687–90 neuroligin family. Am J Hum Genet 2004; 74: 552–7 123 Courchesne E, Pierce K. Why the frontal cortex in 109 Penzes P, Cahill ME, Jones KA, VanLeeuwen JE, autism might be talking only to itself: local over- Woolfrey KM. Dendritic spine pathology in neuro- connectivity but long-distance disconnection. Curr Opin psychiatric disorders. Nat Neurosci 2011; 14: 285– Neurobiol 2005; 15: 225–30 93 124 Frith C. Is autism a disconnection disorder? Lancet 110 Woolfrey KM, Srivastava DP, Photowala H, Yamashita Neurol 2004; 3: 577 M, Barbolina MV V, Cahill ME, Xie Z, Jones KA, Quillam 125 Just MA, Cherkassky VL, Keller TA, Kana RK, Minshew LA, Prakriya M, Penzes P. Epac2 induces synapse NJ. Functional and anatomical cortical under- remodeling and depression and its disease-associated connectivity in autism: evidence from an fMRI study of forms alter spines. Nat Neurosci 2009; 12: 1275– an executive function task and corpus callosum 96 morphometry. Cereb Cortex 2007; 17: 951–61 111 O’Roak BJ, Vives L, Fu WQ, Egertson JD, Stanaway IB, 126 Just MA, Cherkassky VL, Keller TA, Minshew NJ. Corti- Phelps IG, Carvill B, Kumar A, Lee C, Ankenman K, cal activation and synchronization during sentence Munson J, Hiatt JB, Turner EH, Levy R, O’Day DR, comprehension in high-functioning autism: evidence of Krumm N, Coe BP, Martin BK, Borenstein E, Nickerson underconnectivity. Brain 2004; 127: 1811–21 DA, Mefford HC, Doherty D, Akey JM, Bernier R, Eichler 127 Just MA, Keller TA, Malave VL, Kana RK, Varma S. EE, Shendure J. Multiplex targeted sequencing identifies Autism as a neural systems disorder: a theory of recurrently mutated genes in autism spectrum disor- frontal-posterior underconnectivity. Neurosci Biobehav ders. Science 2012; 338: 1619–22 Rev 2012; 36: 1292–313 112 Huttenlocher PR. Dendritic and synaptic development 128 Geschwind DH, Levitt P. Autism spectrum disorders: in human cerebral cortex: time course and critical developmental disconnection syndromes. Curr Opin periods. Dev Neuropsychol 1999; 16: 347–9 Neurobiol 2007; 17: 103–11 113 Singer W. Development and plasticity of cortical pro- 129 Coben R, Clarke AR, Hudspeth W, Barry RJ. EEG power cessing architectures. Science 1995; 270: 758–64 and coherence in autistic spectrum disorder. Clin 114 Dawson G, Ashman SB, Carver LJ. The role of early Neurophysiol 2008; 119: 1002–9 experience in shaping behavioral and brain develop- 130 Murias M, Webb SJ, Greenson J, Dawson G. Resting state ment and its implications for social policy. Dev cortical connectivity reflected in EEG coherence in indi- Psychopathol 2000; 12: 695–712 viduals with autism. Biol Psychiatry 2007; 62: 270–3

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 Cortical construction in autism 133

131 Dinstein I, Pierce K, Eyler L, Solso S, Malach R, evidence from diffusion tensor imaging. Biol Psychiatry Behrmann M, Courchesne E. Disrupted neural synchro- 2004; 55: 323–6 nization in toddlers with autism. Neuron 2011; 70: 148 Shukla DK, Keehn B, Lincoln AJ, Muller RA. White 1218–25 matter compromise of callosal and subcortical fiber 132 Zikopoulos B, Barbas H. Changes in prefrontal axons tracts in children with autism spectrum disorder: a dif- may disrupt the network in autism. J Neurosci 2010; 30: fusion tensor imaging study. J Am Acad Child Adolesc 14595–609 Psychiatry 2010; 49: 1269–78 133 Aboitiz F, Scheibel AB, Fisher RS, Zaidel E. Fiber compo- 149 Hong S, Ke XY, Tang TY, Hang Y, Chu K, Huang sition of the human corpus-callosum. Brain Res 1992; HQ, Ruan ZC, Lu ZH, Tao GT, Liu YJ. Detecting abnor- 598: 143–53 malities of corpus callosum connectivity in autism 134 Uhlhaas PJ, Roux F, Singer W, Haenschel C, Sireteanu using magnetic resonance imaging and diffu- R, Rodriguez E. The development of neural synchrony sion tensor tractography. Psychiatry Res 2011; 194: reflects late maturation and restructuring of functional 333–9 networks in humans. Proc Natl Acad SciUSA2009; 150 Noriuchi M, Kikuchi Y, Yoshiura T, Kira R, Shigeto H, 106: 9866–71 Hara T, Tobimatsu S, Kamio Y. Altered white matter 135 Singer W, Gray CM. Visual feature integration and the fractional anisotropy and social impairment in children temporal correlation hypothesis. Annu Rev Neurosci with autism spectrum disorder. Brain Res 2010; 1362: 1995; 18: 555–86 141–9 136 Piven J, Bailey J, Ranson BJ, Arndt S. An MRI study of 151 Ben Bashat D, Kronfeld-Duenias V, Zachor DA, Ekstein the corpus callosum in autism. Am J Psychiatry 1997; PM, Hendler T, Tarrasch R, Even A, Levy Y, Ben Sira L. 154: 1051–6 Accelerated maturation of white matter in young chil- 137 Hardan AY, Minshew NJ, Keshavan MS. Corpus dren with autism: a high b value DWI study. Neuroimage callosum size in autism. Neurology 2000; 55: 1033–6 2007; 37: 40–7 138 Casanova MF, El-Baz A, Elnakib A, Switala AE, Williams 152 Gozzi M, Nielson DM, Lenroot RK, Ostuni JL, EL, Williams DL, Minshew NJ, Conturo TE. Quantitative Luckenbaugh DA, Thurm AE, Giedd JN, Swedo SE. A analysis of the shape of the corpus callosum in patients magnetization transfer imaging study of corpus with autism and comparison individuals. Autism 2011; callosum myelination in young children with autism. 15: 223–38 Biol Psychiatry 2012; 72: 215–20 139 Chung MK, Dalton KM, Alexander AL, Davidson RJ. 153 Badaruddin DH, Andrews GL, Bolte S, Schilmoeller KJ, Less white matter concentration in autism: 2D voxel- Schilmoeller G, Paul LK, Brown WS. Social and based morphometry. Neuroimage 2004; 23: 242–51 behavioral problems of children with agenesis of the 140 Luna B, Doll SK, Hegedus SJ, Minshew NJ, Sweeney JA. corpus callosum. Child Psychiatry Hum Dev 2007; 38: Maturation of executive function in autism. Biol Psy- 287–302 chiatry 2007; 61: 474–81 154 Turk A, Brown WS, Symington M, Paul LK. Social 141 O’Hearn K, Asato M, Ordaz S, Luna B. Neuro- narratives in agenesis of the corpus callosum: lingui- development and executive function in autism. Dev stic analysis of the Thematic Apperception Test. Psychopathol 2008; 20: 1103–32 Neuropsychologia 2010; 48: 43–50 142 Sabbagh MA. Understanding orbitofrontal contribu- 155 O’Brien G. The behavioral and developmental conse- tions to theory-of-mind reasoning: implications for quences of corpus callosal agenesis and Aicardi syn- autism. Brain Cogn 2004; 55: 209–19 drome. In Callosal Agenesis. Eds M Lassonde, MA Jeeves. 143 Minshew NJ, Luna B, Sweeney JA. Oculomotor evidence New York: Springer, 1994; 235–46 for neocortical systems but not cerebellar dysfunction in 156 Courchesne E, Townsend J, Akshoomoff NA, Saitoh O, autism. Neurology 1999; 52: 917–22 Yeung-Couchesne R, Lincoln AJ, James HE, Haas RH, 144 Egaas B, Courchesne E, Saitoh O. Reduced-size of Schreibman L, Lau L. Impairment in shifting attention corpus-callosum in autism. Arch Neurol 1995; 52: 794– in autistic and cerebellar patients. Behav Neurosci 1994; 801 108: 848–65 145 Lamantia AS, Rakic P. Axon overproduction and elimi- 157 Reuter-Lorenz PA,Kinsbourne M, Moscovitch M. Hemi- nation in the corpus-callosum of the developing rhesus- spheric control of spatial attention. Brain Cogn 1990; monkey. J Neurosci 1990; 10: 2156–75 12: 240–66 146 Alexander AL, Lee JE, Lazar M, Boudos R, DuBray MB, 158 Simms ML, Kemper TL, Timbie CM, Bauman ML, Blatt Oakes TR, Miller JN, Lu J, Jeong EK, McMahon WM, GJ. The anterior cingulate cortex in autism: heterogene- Bigler ED, Lainhart JE. Diffusion tensor imaging of the ity of qualitative and quantitative cytoarchitectonic fea- corpus callosum in autism. Neuroimage 2007; 34: tures suggests possible subgroups. Acta Neuropathol 61–73 2009; 118: 673–84 147 Barnea-Goraly N, Kwon H, Menon V, Eliez S, Lotspeich L, 159 Kostovic I, Rakic P. Cytology and time of origin of inter- Reiss AL. White matter structure in autism: preliminary stitial neurons in the white matter in infant and adult

© 2015 British Neuropathological Society NAN 2016; 42: 115–134 134 J. J. Hutsler and M. F. Casanova

human and monkey telencephalon. J Neurocytol 1980; 166 Torres-Reveron J, Friedlander MJ. Properties of persis- 9: 219–42 tent postnatal cortical subplate neurons. J Neurosci 160 Luskin MB, Shatz CJ. Studies of the earliest generated 2007; 27: 9962–74 cells of the cat’s visual cortex: cogeneration of 167 Suarez-Sola ML, Gonzalez-Delgado FJ, Pueyo-Morlans subplate and marginal zones. J Neurosci 1985; 5: M, Medina-Bolivar OC, Hernandez-Acosta NC, 1062–75 Gonzalez-Gomez M, Meyer G. Neurons in the white 161 Chun JJ, Shatz CJ. Interstitial cells of the adult matter of the adult human neocortex. Front Neuroanat neocortical white matter are the remnant of the early 2009; 3:7 generated subplate neuron population. J Comp Neurol 168 Luhmann HJ, Kilb W, Hanganu-Opatz IL. Subplate cells: 1989; 282: 555–69 amplifiers of neuronal activity in the developing cer- 162 Shatz CJ, Luskin MB. The relationship between the ebral cortex. Front Neuroanat 2009; 3:19 geniculocortical afferents and their cortical target cells 169 Kostovic I, Judas M, Sedmak G. Developmental history of during development of the cat’s primary visual cortex. J the subplate zone, subplate neurons and interstitial Neurosci 1986; 6: 3655–68 white matter neurons: relevance for schizophrenia. Int J 163 McConnell SK, Ghosh A, Shatz CJ. Subplate pioneers Dev Neurosci 2011; 29: 193–205 and the formation of descending connections from cer- 170 Pugash D, Hendson G, Dunham CP, Dewar K, Money ebral cortex. J Neurosci 1994; 14: 1892–907 DM, Prayer D. Sonographic assessment of normal and 164 Ghosh A, Shatz CJ. Involvement of subplate neurons in abnormal patterns of fetal cerebral lamination. Ultra- the formation of ocular dominance columns. Science sound Obstet Gynecol 2012; 40: 642–51 1992; 255: 1441–3 165 Ghosh A, Shatz CJ. Segregation of geniculocortical Received 8 September 2014 afferents during the critical period – a role for subplate Accepted after revision 16 January 2015 neurons. J Neurosci 1994; 14: 3862–80 Published online Article Accepted on 29 January 2015

© 2015 British Neuropathological Society NAN 2016; 42: 115–134