Skip to main content
NCBI home page
Brain Pathology logoLink to Brain Pathology
. 2016 Aug 15;27(4):437–448. doi: 10.1111/bpa.12415

Primary olfactory cortex in autism and epilepsy: increased glial cells in autism

David A Menassa 1, Carolyn Sloan 1, Steven A Chance 1,✉
PMCID: PMC8029489  PMID: 27409070

Abstract

Autism Spectrum Disorder is characterized by sensory anomalies including impaired olfactory identification. Between 5 and 46 percent of individuals with autism have a clinical diagnosis of epilepsy. Primary olfactory cortex (piriform cortex) is central to olfactory identification and is an epileptogenic structure. Cytoarchitectural changes in olfactory cortex may underlie olfactory differences seen in autism. Primary olfactory cortex was sampled from 17 post‐mortem autism cases with and without epilepsy, 11 epilepsy cases without autism and 11 typically developed cases. Stereological and neuropathological methods were used to quantify glial, pyramidal and non‐pyramidal cell densities in layers of the piriform as well as identify pathological differences in this area and its neighbouring region, the olfactory tubercle. We found increased layer II glial cell densities in autism with and without epilepsy, which were negatively correlated with age and positively correlated with levels of corpora amylacea in layer I. These changes were also associated with greater symptom severity and did not extend to the olfactory tubercle. Glial cell organization may follow an altered trajectory of development with age in autism. The findings are consistent with other studies implicating increased glial cells in the autism brain. Altered cytoarchitecture may contribute to sensory deficits observed in affected individuals. This study provides evidence that autism is linked to alterations in the cytoarchitectural structure that underlies primary sensory processes and is not restricted to heteromodal (“higher”) cognitive centers.

Keywords: autism spectrum disorder, epilepsy, glia, olfaction, primary olfactory cortex

Background

Atypical sensory processing in the olfactory 1, 2, visual 3, 4 and auditory 5 domains has been reported in children 6 and adults 7 with autism spectrum disorders (ASDs). Of these sensory modalities, the olfactory system has been relatively neglected 2. Decreased olfactory identification is the most consistent finding in children and adults with autism 8, 9, 10, although abnormalities in olfactory discrimination and sensitivity have sometimes been reported 2.

The primary olfactory cortex, known as the piriform cortex (PiC), is a paleocortical structure that plays a central role in olfactory perception and identification 11. The cytoarchitecture of this area in ASD is undocumented. Neighbouring regions of the PiC are affected in autism: the amygdala has 12% fewer neurons 12, the hippocampus has increased cell‐packing density and decreased neuronal size in the cornu ammonis (CA1) region 13 and the fusiform gyrus has layer‐specific decreases in neuronal numbers and densities 14. Alterations affecting the fate of neurons migrating to another neighbouring region of the PiC, the olfactory tubercle (Tu), have been suggested to account for impairments in social behaviour 15. Interestingly, the trilamination of olfactory cortex (the PiC and the Tu) develops with input from the olfactory bulb and newly born cells continue to migrate along this olfactory axis well into adulthood to form interneurons 16, 17.

The PiC is also an epileptogenic structure 18, 19. The incidence of epilepsy in children with ASD varies between 5% and 46% 20, 21. In focal epilepsy patients, epileptic activity is recorded in the anterior PiC 18 and neuronal loss and severe gliosis are reported in all layers following death by status epilepticus 22. In the kainate and pilocarpine animal models of epilepsy, layer II GABA‐ergic neurons are severely affected 19.

It is unclear to what extent cytoarchitectonic changes of the PiC explain alterations in olfactory function. For example, patients with temporal lobe epilepsy (TLE) experience unpleasant olfactory auras often preceding complex partial seizures located to the PiC 23. Histology in the same patients shows gliosis in mesial temporal regions 23. Little is known about the cytoarchitecture of the PiC when the autism is co‐morbid with epilepsy and further investigations may help to clarify why individuals with autism have a lower threshold for epilepsy compared with the typically developing (TD) population.

The high incidence of epilepsy in autism 20, 24 and altered neurotransmission through the GABA‐ergic and glutamatergic systems 25 are in support of a possible imbalance between cortical excitation and inhibition (E/I) in autism 26. Non‐pyramidal inhibitory interneurons are implicated in this hypothesis 27, as are glial cells 28, 29: astrocytes, which play a key role in mopping up glutamate from the extracellular milieu to prevent excitotoxicity and microglia, the central nervous system macrophages. Glial cells also react to pathology, and may alter neuronal connectivity as a result of glial scarring 30. Furthermore, glial cells are found in unusually high numbers in ASDs 29, 31, 32.

In this article, pyramidal, non‐pyramidal and glial cell densities in layers of the PiC were quantified in a substantial cohort of post‐mortem cases. It was hypothesized that neuronal densities would be decreased and glial cell densities would be increased in ASD, autism co‐morbid with epilepsy (ASD + E) and epilepsy (E) cases, compared with TD control cases.

Materials and Methods

Selection of cases

Eleven TD, 11 idiopathic/chronic E, 11 ASD with no neurological co‐morbidity and six ASD + E cases were selected based on their medical histories. Autism diagnosis was confirmed for the majority of cases (13/17) using cut‐off scores of the Autism Diagnostic Interview‐Revised (ADI‐R) (see below). A further two cases had been diagnosed during life, three cases were identified as ASD from clinical notes and one case was confirmed in the USA. Common ASD comorbidities were not exclusion criteria for the selection of cases due to the limited availability of post‐mortem tissue. The Autism Tissue Program (ATP) and the Thomas Willis Oxford Brain Collection (TWOBC) ethics committees granted ethical approval for this study. Cases were matched for mean age (individual matching as triads across the TD, ASD and E groups was not possible). Patient data and diagnosis are presented in Table 1 (Supporting Information Additional files 1 and 2). Post‐mortem reports were available for all cases.

Table 1.

Demographics of cases.

Subjects ASD Chronic E ASD + E Comparison TD
7M:4F 7M:4F 5M:1F 5M:6F
Mean SD Mean SD Mean SD Mean SD
Age at death (years) 30.09 14.87 36.45 17.50 25.67 9.58 34.18 25.56
Post‐mortem interval (hours)† 65.14 38.59 49.01 28.72 72.50 29.42 35.01 19.48
Storage time in formalin (years)†† 5.27 4.11 19.09 9.61 5.17 5.96 14.59 9.28
Mean fixed brain weight (grams)* 1486.40 279.33 1439.27 129.07 1452.17 250.32 1446.96 124.09
Hemisphere 9L:2R 8L:3R 4L:2R 8L:2R:1**
Brain bank of origin
Harvard 0 0 1 1
NICHD 0 0 0 2
TWOBC 11 11 5 9
Total Number 11 11 6 11
Open in a new tab

†PMI n/k for 5 cases

††Storage not known for one case

*Brain weight not known for 5 cases

**Hemisphere not known for one case.

Summed ADI‐R scores of all autism subjects were: social interaction domain (diagnostic threshold is 10), ASD = 21 ±6, ASD + E = 24 ±2; verbal communication domain (diagnostic threshold is 8), ASD = 19 ±5, ASD + E = 15 ±5; restricted and repetitive behaviours domain (diagnostic threshold is 3), ASD = 6 ±3, ASD + E = 7 ±2 (see Supporting Information Additional file 3 for individual domain scores).

Intellectual disability (ID) was assessed in 14/17 autism cases using ICD‐10 guidelines (five with severe ID; one with moderate ID; four with mild ID; four with no ID; three n/k). ASD + E cases had more severe ID compared with ASD. Four E cases were diagnosed with idiopathic epilepsy and seven with chronic epilepsy, of which two had severe ID, and none of these cases had any psychiatric diagnosis.

Anatomy and delineation of the region of interest

Very few detailed anatomical descriptions of the human PiC exist. It is situated at the junction between the temporal and frontal lobes on the caudolateral aspect of the orbitofrontal cortex at the level of the entorhinal sulcus (Figure 1A,B). Its rostral limit is identified as it passes into the anterior olfactory nucleus where the olfactory peduncle is attached to the cerebral hemisphere 33, 34, 35. The PiC is bound laterally by the rhinal fissure (anterior and posterior), medially by the Tu and the amygdalar fissure that separates it from the amygdalar nuclei, and the claustral cleft that separates it from the claustrum 34, 36. Caudally, the PiC passes eventually into the entorhinal cortex. The PiC may be divided into the anterior (APC) and posterior piriform (PPC) portions based on anatomical, physiological and functional differences 11, 37. The PiC is trilaminar with a superficial and neuron‐sparse molecular layer I, a dense cellular layer II constituted of pyramidal neurons and a thick layer III with larger pyramidal cells 34, 36.

Figure 1.

Figure 1

Open in a new tab

Coronal slice of primary olfactory areas in the human brain. The piriform cortex in its frontal (PiF) and temporal (PiT) portions surrounded by various anatomical structures was sampled in the coronal plane (left hemisphere shown here). The photo was modified from 31. AcC: central accumbens, AcL: lateral accumbens, AcM: middle accumbens, Ent: entorhinal cortex, FPu: fundus of putamen, FuG: fusiform gyrus, Ilf: inferior longitudinal fasciculus, ITG: infernal temporal gyrus, MTG: middle temporal gyrus, NA: nucleus accumbens, PACl: preamygdalar claustrum, PPCl: prepiriform claustrum, PPo: planum polare, PRC: perirhinal cortex, Pu: putamen, SSTI: substriatal terminal island, STG: superior temporal gyrus, TCl: temporal claustrum, TI: terminal island, Tu: olfactory tubercle, TuTI: tubercle terminal island, Unc: uncinate fasciculus, VCl: ventral claustrum.

The dorsal and medial cytoarchitectonic limits of the PiC were identified by a cell‐free layer below the claustral complexes. The ventral limit was identified where layer III fused with the periamygdaloid area. Layer II was identified as a very dense cell layer in the shape of an –s. Layer I was identified as the cell‐sparse region between layer II and the pial surface (Figure 2A–C and Supporting Information Additional file 4). The PiC was sampled from the anterior side of the coronal slice containing the anterior hippocampus to obtain the characteristic s‐shape.

Figure 2.

Figure 2

Open in a new tab

Trilamination of the PiC and delineation of region of interest. The characteristic trilamination of the PiC with the boundaries (in dashed lines) of the different layers where the cell densities were measured are shown in a representative case (2A). High‐magnification photos (2B and 2C) show the cell‐dense second layer of the PiC. Top of 2A. is dorsal and left is medial. En: endopiriform nucleus; Ent: entorhinal cortex; PiC: piriform cortex; PiF: frontal pirifrom; PiT: temporal piriform; PPCl: prepiriform claustrum.

Neurohistological sampling and Nissl‐staining

Five millimeter‐thick coronal slices were made at the start of the fronto‐temporal junction through the entire length of the temporal lobe and from these 2 × 2 cm2 (x,y) blocks of the PiC were sampled. All blocks were photographed and marked to allow histological sections to be taken along the anteroposterior axis.

Tissue blocks were paraffin‐embedded because paraffin is known to preserve cell morphology in post‐mortem tissue, limit the guard zone problem, and restrict collapse in the z‐axis 38, 39. The spacing distance between sections was set at 200 μm from anterior to posterior. Sections were then stained for the Nissl‐substance using cresyl‐violet (0.1%) for 3 minutes. All sections were coded and identifiers were hidden until de‐blinding for statistical analysis.

Quantitative assessment

All cell counts were performed using the Axiovision image analysis system (version 4.8.1, 2010, Zeiss company, Bicester, England). A square counting frame was used to count cells lying within the frame, classified as pyramidal, non‐pyramidal and glial cells in layers II and III based on their morphology (the frame had two exclusion edges and two inclusion edges): pyramidal cells have a prominent darkly stained nucleolus and a well‐defined nuclear membrane, with an apical dendrite; non‐pyramidal cells are often smaller than pyramidal neurons (5–10 μm in diameter), sometimes aspiny and may be multipolar, bitufted or semilunar 37; astroglial cells are much smaller than neurons (2–5 μm in diameter) and tend to have an indistinct cell outline and can be binucleated in reactive astrocytosis 40. Oligodendrocytes have a small, round and darkly stained nucleus and are often in a close proximity to neurons (satellite cells). Microglial cells are irregularly shaped and have an undefined contour. Glial cell counts reported here reflect a combination of these cell types because the distinction between them is not always clear. In particular, many researchers consider non‐reactive astrocytes and oligodendrocytes to be often indistinguishable. A combined count was considered acceptable as previous studies indicated that alterations are not confined to a single glial subpopulation in ASD 31, 32 (Supporting Information Additional file 5).

Counts were performed at 60× magnification (numerical aperture = 1.4). Distance between the probes in layer III was 400 μm in the x and y directions, beginning from a systematically random start position and 150 × 120 μm2 for layer II. For layer II, cells were counted in 35 systematically randomly placed counting frames typically over five sections per case (seven frames/section) and for layer III, 40 counting frames were used (eight frames/section). Consistent with published guidelines 41, 42, more than 100 cells per type were counted per case. The different sampling distributions were due to the notable difference in cell density between the two layers. Unidentifiable cells constituted less than 5% of the total of counted cells. Counting frame area in the x and y directions was calibrated at 9025 μm2 for layer II and at 11 400 μm2 for layer III. Many authors have demonstrated that in paraffin sections, the final section thickness varies little compared to the nominal section thickness 39, 43. In this study, the mean real section thickness was recorded during sampling in all cases (200 sections × two layers). There was notable variability in the z‐direction collapse across every section and the mean z‐thickness was 16.31 ± 3.07 (μm), including a variable guard zone depth (mean = 3.3 μm, above and below the counting box).

Neuropathology

Qualitative pathological examination of the PiC and the Tu was conducted. Features of neuropathological interest were assessed such as the position of neuronal nuclei in the cell body, the size of the cell body, the presence of gliosis and the presence of corpora amylacea (Ca) in the subpial spaces and around layer I of the PiC and the Tu. Ca have been associated with the presence of reactive glia and are thought to originate from and accumulate in astroglial foot processes 44, 45. The Ca were identified as spherical and basophilic aggregates with a relatively homogeneous dark coloration. They can have various sizes and their diameter can be as high as 10 μm 46. Ca density was assessed using a scoring system of 0–3 (0 = no deposits, 1 = mild, 2 = moderate, 3 = severe) per section (Supporting Information Additional file 5).

Immunohistochemistry

Immunohistochemistry was performed on selected tissue sections (6 μm thick) to visualize reactive microglial cells (CD68 positive) and reactive astrocytes (GFAP positive). These experiments were performed using a peroxidase/DAB detection system (EnVision kit from DAKO; K5007) preceded by a hydrogen‐peroxide block for 30 minutes, a heat‐mediated antigen retrieval step in citrate buffer (pH = 6) for 3 minutes, and a primary incubation for one hour at a 1:50 dilution for the CD68 (DAKO; M0876) antibody and 1:1000 dilution for the GFAP (DAKO; Z0334) antibody. Sections were counterstained in Haematoxylin and mounted in DPX prior to imaging (Supporting Information Additional file 10).

Statistical Analysis

General linear model and correlations

A 2 × 3 × 4 repeated measures analysis of variance (ANOVA) general linear model was used to examine main effects of cell‐type, layer and diagnostic group on density measurements as well as the interactions between density and independent variables. The model included layer (two levels) and cell‐type (three levels: non‐pyramidal cells, pyramidal cells, glia) as within‐subject factors and diagnosis (four levels) as a between‐subject factor. Post‐hoc analyses were adjusted for multiple comparisons using a Bonferroni correction. In general, to aid intelligibility, statistical values are only reported below for positive findings and statistical “trends” (P < 0.1) or for notable negative findings. Assumptions of normality (the Shapiro‐Wilk statistic), homoscedasticity (Levene's test) and Mauchly's test for sphericity were met prior to the application of the model. Where sphericity was violated, the Greenhouse‐Geisser correction was applied (ε > 0.75).

Additional demographic variables such as brain weight and age did not differ between groups. Storage time of tissue in formalin was significantly different between groups (F = 7.69, d.f. = 3, P < 0.0001) (due to the shorter duration in the autism cases) and was therefore included in a repeated measures analysis of covariance (ANCOVA). The ANCOVA included layer and cell‐type as within subject factors and diagnosis between subjects with fixation time as a covariate. Pearson's correlations were applied to examine associations of cell density to age, brain weight and autism severity.

Accuracy and sample size

The coefficient of error (CE) for individual estimates of cell density was calculated to assess accuracy 47, 48 using “R” for statistical computing (version 3.0.1, Vienna, Austria, 2013). A group CE (GCE) was calculated as the SEM of all cell densities per group, layer and cell‐type divided by the mean density for the whole group. This gives an estimate of the accuracy of the mean for each subject group. The lower the GCE, the more precise the mean and typically GCE values should fall below 0.10.

In the final analysis, for all measurements in the study the actual GCE ranged between 0.03 and 0.09 for both layers and all groups except for layer II pyramidal cell densities in the small ASD + E group (0.13), indicating good accuracy of group averages overall (see Table 2 for GCEs).

Table 2.

Cell densities (cells.mm−3) and group coefficients of error.

ASD Chronic epilepsy ASD + E Comparison TD
(N = 11; 7M:4F) (N = 11; 7M:4F) (N=6; 5M:1F) (N = 11; 5M:6F)
Mean SD GCE Mean SD GCE Mean SD GCE Mean SD GCE
Layer II Pyramidal 94972.82 12838.94 0.04 113730.39 20358.84 0.05 92159.2 29081.41 0.13 96990.80 18639.05 0.06
Layer II Non‐Pyramidal 20572.93 3361.33 0.05 21211.64 3114.36 0.04 22973.16 5539.51 0.07 19119.37 4585.57 0.07
Layer II Glial 73062.60 18061.35 0.07 57346.38 10724.13 0.06 71144.04 15266.45 0.06 52825.26 12990.04 0.05
Layer II Neuronal 115545.79 14792.03 0.04 134878.37 19768.64 0.04 114879.63 30314.99 0.08 116110.17 19323.76 0.07
Layer III Pyramidal 24517.73 4893.34 0.06 25050.88 4380.35 0.05 23489.78 4576.34 0.08 23599.06 3870.27 0.05
Layer III Non‐Pyramidal 11494.80 1765.78 0.05 12571.84 2706.00 0.06 11612.41 2922.80 0.09 11595.04 1497.95 0.04
Layer III Glial 86460.55 8131.36 0.03 85375.74 15114.55 0.05 89626.94 16275.14 0.07 78857.92 12323.73 0.05
Layer III Neuronal 35931.99 6181.06 0.05 37698.12 6381.74 0.05 34729.20 6624.86 0.08 35194.10 5020.44 0.04
Open in a new tab

Based on means and standard deviations of cell density findings in different cortical areas reported by other researchers, a desirable sample size was estimated for the present study. Neuronal density changes of 13% in layer III in the fusiform gyrus, for which a sample size of n = 9 per group was estimated to provide 80% power (two‐sided, target P = 0.05) 14. Larger differences have been reported for neuropil space (19%) by one study 49 and for cell counts (64%) by another study 50. Therefore, based on the most conservative of these findings 14, it was estimated (power calculation based on a t‐test model) that a sample size of 11 subjects per group would provide >90% power.

Results

Effects of diagnosis, cell type and cortical layer

No main effect of diagnosis on cell density was found (F = 1.30, d.f. = 3, P = 0.29). There was a significant main effect of layer (F = 235.97, d.f. = 1, P < 0.0001) and cell‐type (F = 424.77, d.f. = 1.64, P < 0.0001). Layer‐by‐cell‐type interaction was significant (F = 320.59, d.f. = 1.39, P < 0.0001) and so was the cell‐type‐by‐diagnosis interaction (F = 2.76, d.f. = 4.91, P < 0.05). Layer‐by‐diagnosis and layer‐by‐cell‐type‐by‐diagnosis interactions were not significant.

Covariance—effect of storage time in formalin

There was no main effect of storage time of tissue in formalin on cell density (F = 0.03, d.f. = 1, P = 0.87). Other statistical results were not substantially affected by the inclusion of fixation time, except that the cell‐type x diagnosis effect was lost (F > 0.20, d.f. > 3, P > 0.71). The interpretation of this is considered in the Discussion section below.

Post hoc testing of mean cell densities between diagnostic groups

There was a significant between‐group difference in glial cell densities in layer II (F = 4.81, d.f. = 3, P = 0.007). Between‐group differences were not significant for glial cell densities in layer III (F = 1.13, d.f. = 3, P = 0.35), pyramidal cell densities in layers II (F = 0.80, d.f. = 3 P = 0.50) and III (F = 0.27, d.f. = 3, P = 0.85); and non‐pyramidal cell densities in layers II (F = 1.08, d.f. = 3, P = 0.37) and III (F = 0.56, d.f. = 3, P = 0.46).

Glial cell densities in layers II and III

There was a significant 38.3% increase in mean glial cell densities in ASD compared with TD (P = 0.01) in layer II. Although mean glial cell densities increased by 34.67% in ASD + E compared with TD also in layer II, this difference was not significant between groups (P = 0.10) (Table 2), presumably due to the small number of ASD + E cases. Both ASD + E and ASD groups had comparable means of layer II glial cell densities (3% difference). Means in layer II were also comparable between TD and E (P > 0.05). For all autism cases (ASD & ASD + E), mean glial densities were significantly increased in comparison with both TD and E (P < 0.01). There were no significant differences in mean glial cell densities in layer III across all pairwise comparisons (Figure 3, Table 2 and Supporting Information Additional files 6, 7).

Figure 3.

Figure 3

Open in a new tab

Cell densities across diagnostic groups in layers II and III. Glial cell densities in layer II (top left) of the PiC were significantly increased in ASD in comparison with TD. Mean glial cell densities in layer II were comparable between ASD and ASD + E. When all ASD were grouped together, glial densities are significantly increased in autism in comparison with E and TD. Layer III densities (top right) were not different across diagnostic groups. Neuronal cell densities in layers II (bottom left) and III (bottom right) of the PiC did not differ between ASD, ASD + E, E and TD. A scatter plot of individual densities per diagnostic group and layer is provided with error bars representing the standard deviation of values from the mean density shown by the middle bar in each group.

Neuronal cell densities in layers II and III

There were no significant differences between groups in mean pyramidal and non‐pyramidal cell densities in layer II or layer III across all pairwise comparisons (Figure 3, Table 2 and Supporting Information Additional files 6, 7).

The effect of age

Based on studies of abnormal development of the brain in autism as well as the effects of age on brain cell densities 51, 52, the relationship with age was investigated. Age was negatively correlated with glial cell density in layer II (r = −0.46; P = 0.06) in autism, whilst a positive correlation was noted within the TD group (r = 0.33; P = 0.09; Figure 4). The difference between the positive and negative slopes in TD and autism respectively was statistically significant (F = 4.44, d.f. = 24, P = 0.046). Age was negatively correlated with pyramidal cell density in layer II only in the TD group (r = −0.75, P = 0.009). No other significant correlations were observed.

Figure 4.

Figure 4

Open in a new tab

Glial cell densities and age in autism and TD. Glial cell densities in layer II were negative correlated with age in autism (r = −0.46, P = 0.06) (right). A more positive correlation with age (r = 0.33, P > 0.05) was seen in the TD group (left).

Pathological findings

Pathological observations showed no cortical lamination abnormalities except in one epilepsy case. Paleocortical lamination appeared normal in all cases and in one autism case, we saw some abnormally shaped cells with large vacuous cytoplasms (dotted circle in Figure 5A–O and Supporting Information Additional file 8). There was moderate to severe accumulation of Ca in outer layer I of the PiC, the subpial spaces and sometimes the perivascular areas in 15/17 autism cases, compared with only three TD cases and four E cases having numerous Ca. There were significant differences in Ca levels between groups (H = 14.04, d.f. = 2, P = 0.0009) (Kruskal–Wallis with Dunn's correction). Post hoc testing showed that the TD and E groups did not differ in mean levels of Ca (Mean Rank Diff = −6.00, P > 0.05). The autism group had significantly higher Ca levels compared with TD (Mean Rank Diff = 15.62; P = 0.0008) and tended to be higher compared with E (Mean Rank Diff = 9.62, P = 0.07) (Figure 6A–E). High Ca levels were also associated with higher glial cell densities in layer II (Spearman's rho = 0.61, P < 0.001) and a trend was observed for layer III glial densities (Spearman's rho = 0.39, P = 0.09). Where the Ca deposits were severe, we sometimes observed glial fibers, microglial cells and early calcification signs (Supporting Information Additional files 9 and 10). Ca deposits became less dense in the Tu and seemed not to extend beyond the PiC in autism (Supporting Information Additional file 11).

Figure 5.

Figure 5

Open in a new tab

The PiC across diagnostic groups. The characteristic trilamination of the PiC is shown here in representative TD (5A–C), ASD (5D–F) with multiple corpora amylacea in layer I (5D), E (5G–I) and two ASD + E cases, one (5J–L) showing abnormally shaped cells in layer I (dotted circle, 5J and Supporting Information Additional file 8) and the second case (M‐O and Supporting Information Additional file 9) marked by extensive gliosis. Top of every square is dorsal and left is medial in the coronal plane.

Figure 6.

Figure 6

Open in a new tab

Corpora amylacea deposits in ASD. Clear histological differences between autism and TD are shown here. Higher corpora amylacea levels were observed around layer I of the PiC in ASD in comparison with TD. 6A and 6C show no corpora amylacea in TD whilst 6B and 6D show many corpora amylacea in an ASD case around layer I. Corpora amylacea levels were significantly increased in autism (6E). Bars represent mean levels of corpora amylacea and the standard deviations per diagnostic group. Es: entorhinal sulcus. Top is dorsal and left is medial.

Associations with symptom severity

Layer II neuronal densities correlated positively with ADI‐R scores in the verbal communication domain (r = 0.71; P < 0.05). Glial cell densities in layer II showed a trend for a positive correlation with scores in the restricted and repetitive behaviour domain (r = 0.49, P = 0.08). The ADI‐R questionnaire includes an item entitled “unusual sensory interests” (item 71, code EUNSENS). The clinician rates this item according to a severity scale ranging from 0 to 2 (0 = none, 1 = mild, 2 = severe). Associations between item 71 scores and cell densities were not significant for any cell type (Figure 7).

Figure 7.

Figure 7

Open in a new tab

Correlations between cell densities and symptom severity in layer II. Neuronal cell densities correlated positively with scores from the verbal communication domain of the ADI‐R (r = 0.71, P < 0.01) (left graph), while glial cell densities in layer II showed a positive trend with the ADI‐R restricted and repetitive behavior domain scores (r = 0.49, P = 0.08) (right graph). Note that there are some missing values because it was not possible to obtain behavioral scores for all the cases.

Discussion

The PiC is a neglected area in human brain research and this is the first study to report cell measurements in autism in this region. The main finding was increased glial cell densities in autism. A comparable increase was seen in autism comorbid with epilepsy cases but no difference was observed in epilepsy without autism. This is consistent with reports on neocortical areas that suggest that increased glial presence could be a feature of the autism brain 28, 29, 53.

Glial cell densities and age were negatively correlated in autism, which was significantly different from the positive relationship seen in TD. In typical ageing, neuron‐glia ratios and synaptic densities decrease in the neocortex 54. The present findings suggest that the altered developmental trajectory of neuronal populations associated with age in autism 55 may be extended to include glial cell types in the PiC. The ratios of glia/neurons for each group were, Layer II: ASD = 0.63, E = 0.42, ASD + E = 0.62, Comparison TD = 0.45; Layer III: ASD = 2.40, E = 2.27, ASD + E = 2.55, Comparison TD = 2.24. There was a clear contrast between layers II and III in the ratio, but within that framework, the relatively increased density of glia can be seen in the autism groups in layer II and, to a lesser extent in layer III.

Increased Ca deposits in autism and their association with glial cell densities could be suggestive of neuroinflammatory processes though this may not necessarily be pathological as these aggregates can be observed in the normal aging brain 45. Neuropathological features of gliosis, signs of early calcification and the presence of microglial cells co‐occurring with these aggregates, lends support to the possible presence of a pathological process in some cases but the cellular mechanism requires elucidation. Increased Ca levels have been previously noted in the hippocampus and cerebellar granular layer in idiopathic autism 56. Ca are spherical basophilic aggregates formed of a mixture of short and long polysaccharides and polymerized proteins 57. They are immunoreactive to S100 proteins expressed by astrocytes and microglia 45 and are thought to accumulate in reactive astrocyte foot processes in response to cellular stress and neuroinflammation 44. Ca have been reported in the normal aging human brain 45, 58 and also in cases of hippocampal sclerosis associated with TLE 59. In this study, the combination of increased glial cells in the PiC and the marked accumulation of Ca make a compelling association. By contrast, the high neuronal and non‐neuronal cell turnover in the ventral striatum could be the reason why low levels of Ca were observed in the Tu 60. The Tu and the PiC have different functions with the former enabling odour discrimination and source identification and the latter processing odour valence 61.

Glial cell abnormality occurs outside the cerebral cortex in autism, including in the subcortical white matter 62, the cerebellar gray and white matter 29, and in the amygdala 53. All of these reports demonstrate increases in density or in glial “reactivity.” Although the glial fibrillary acid protein (GFAP) is often used as a marker for reactive astrocytes, it does not necessarily recognise astrocytes in all states of reactivity 31. The prefrontal cortex 63 and the cingulate 64 have increased astroglial profiles in autism. Primary sensory regions also show glial changes: for example, microglial cells are increased in the visual cortex (Areas 17 and 18), although astroglia were not assessed in this region 32. Our study suggests that primary sensory regions are not spared in autism and that glial cell abnormality could be characteristic of ASDs.

The strong association between glial densities and symptom severity is worth noting. The PiC has extra‐cortical connections to the prefrontal cortex and the amygdala and its role is not restricted to olfactory identification 11. Therefore, its role within wider networks may explain why alterations here are associated with general symptom severity. Similarly, it has been suggested that the Tu's function may be better understood within the context of the reward system in general 65. Furthermore, our findings are broadly in support of a general increase in cell density as a feature of the cerebral cortex in autism 50, 55, which may be an index of phenotype severity modified by age.

Consistent with Rubenstein's model of altered cortical E/I in autism 26, glial cells can contribute to modifying excitability in the PiC in the developing brain. Astroglia remove 80% of synaptic glutamate through glutamate transporters (excitatory amino acid transporter [EAAT1 and EAAT2]), modulated by gap junction proteins and can, therefore, protect against glutamate‐induced excitotoxicity 66. Therefore, glial cells may play a significant role in autism neuropathology and also contribute to the development of epileptogenicity.

Limitations and potential confounding factors

Although the sample size of brain tissue obtained from human donors is limited, the current study uses a substantial proportion of tissue available globally for autism research drawn from the UK and USA brain banks. Two ASD brains in this study were from donors diagnosed during life for whom the clinical notes were not sufficiently comprehensive to enable post‐mortem confirmation. Although there can be a challenge due to mis‐diagnosis in some community samples (estimated at 15%) the two cases in question did not constitute extreme values in the autism group and their statistical exclusion does not substantially change the key finding of a cell‐type × diagnosis interaction, which remains statistically significant. Regarding tissue sampling and anatomy, the PiC was not suitable for strict assumption‐free stereological sampling as, for example, tissue could not be truly randomly oriented with respect to the plane of sampling due to a need to preserve interpretable anatomical organization in the coronal orientation.

Finally, although inclusion of formalin fixation time had no main effect on the PiC, it resulted in reduced significance of the cell x diagnosis interaction. Typically, longer fixation is associated with tissue shrinkage with a concomitant increase in cell density. However, the ASD group, which had increased glial density, had the shortest fixation time. Therefore, the statistical effect is unlikely to indicate a causal relationship and is rather a consequence of the demographic accident of the ASD group also having a shorter storage time due to the recency of tissue collection from these cases.

Conclusions

The PiC is a novel area of investigation in autism. These data contribute to a growing body of literature indicating a widely distributed neuropathology in autism with glial cell changes involving primary sensory regions. Although this offers a challenge to previous hypotheses, which have emphasized some higher cognitive functions in heteromodal associative cortex 67, 68, it may support the development of tests in early life, which investigate basic sensory processing. Results from this study are compatible with disruption of the olfactory domain consistent with the high incidence of sensory anomalies in autism.

Author Contributions

DAM and SAC designed the study. DAM carried out the study under the supervision of SAC, selected the cases, collected the clinical details, performed the neurohistological sampling and unbiased systematic random cell counting, neuropathology and statistical analysis and drafted the manuscript. CS performed paraffin embedding of the tissue blocks and provided the histological slides for the study. SAC and DAM revised the manuscript. All authors read and approved the final manuscript (also see acknowledgements).

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's web‐site:

Additional file 1: Causes of death.

Click here for additional data file. (87.5KB, docx)

Additional file 2: Storage duration, PMI, and brain weight measures.

Click here for additional data file. (74.9KB, docx)

Additional file 3: Summed ADI‐R domain scores for autism cases.

Click here for additional data file. (30.5KB, doc)

Additional file 4: Anteroposterior mapping of the PiC in Nissl sections.

Click here for additional data file. (4.6MB, tiff)

Additional file 5: Differences in morphology with Nissl.

Click here for additional data file. (5.7MB, tiff)

Additional file 6: Individual neuronal and glial cell densities in layer II.

Click here for additional data file. (58.5KB, doc)

Additional file 7: Individual neuronal and glial cell densities in layer III.

Click here for additional data file. (57KB, doc)

Additional file 8: Morphology of abnormal cells in autism and epilepsy.

Click here for additional data file. (10.9MB, tiff)

Additional file 9: Glial cells co‐existing with corpora amylacea.

Click here for additional data file. (9.6MB, tiff)

Additional file 10: CD68 positive cells co‐occurring with corpora amylacea in layer I.

Click here for additional data file. (2.7MB, tiff)

Additional file 11: Corpora amylacea deposits in the olfactory tubercle in autism.

Click here for additional data file. (10.4MB, tiff)

Acknowledgments

We would like to thank Professor Angela Vincent, Professor Margaret Esiri for her assistance with the neuropathology, Dr. Thomas Berney for providing the behavioural scores and confirming the autism cases, Dr. Sven Braeutigam, Dr. Douglas C. Crockett, Professor Seth Love and Miss Lai San Hong for her assistance with the statistical analysis and Dr. Istvan Adorjan for his assistance with the immunohistochemistry. Written informed consent was obtained by the Harvard Brain Tissue Resource Centre, the NICHD University of Maryland Brain and Tissue Bank and the TWOBC for publication of the individual case details and accompanying images in this manuscript. The consent forms are held by the brain banks and in the patients' clinical notes and are available for review by the Editor‐in‐Chief. This work was supported by project grant #6026 from Autism Speaks and a SFARI grant ID: 307098 held by Dr. Steven Chance, and a Clarendon Fund and Department of Clinical Neurology scholarship at the University of Oxford to Dr. David A. Menassa.

The authors declare that they have no competing interests.

References

  • 1. Wicker B, Monfardini E, Royet JP (2016) Olfactory processing in adults with autism spectrum disorders. Mol Autism 7:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Schecklmann M, Schwenck C, Taurines R, Freitag C, Warnke A, Gerlach M et al (2013) A systematic review on olfaction in child and adolescent psychiatric disorders. J Neural Transmission 120:121–130. [DOI] [PubMed] [Google Scholar]
  • 3. Marco EJ, Hinkley LB, Hill SS, Nagarajan SS (2011) Sensory processing in autism: a review of neurophysiologic findings. Pediatr Res 69:48R–54R. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Simmons DR, Robertson AE, McKay LS, Toal E, McAleer P, Pollick FE (2009) Vision in autism spectrum disorders. Vision Res 49:2705–2739. [DOI] [PubMed] [Google Scholar]
  • 5. O'Connor K (2012) Auditory processing in autism spectrum disorder: a review. Neurosci Biobehav Rev 36:836–854. [DOI] [PubMed] [Google Scholar]
  • 6. Tomchek SD, Dunn W (2007) Sensory processing in children with and without autism: a comparative study using the short sensory profile. Am J Occup Ther 61:190–200. [DOI] [PubMed] [Google Scholar]
  • 7. Crane L, Goddard L, Pring L (2009) Sensory processing in adults with autism spectrum disorders. Autism 13:215–228. [DOI] [PubMed] [Google Scholar]
  • 8. Bennetto L, Kuschner ES, Hyman SL (2007) Olfaction and taste processing in autism. Biol Psychiatry 62:1015–1021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. May T, Brewer WJ, Rinehart NJ, Enticott PG, Brereton AV, Tonge BJ (2011) Differential olfactory identification in children with autism and Asperger's disorder: a comparative and longitudinal study. J Autism Dev Disord 41:837–847. [DOI] [PubMed] [Google Scholar]
  • 10. Suzuki Y, Critchley HD, Rowe A, Howlin P, Murphy DG (2003) Impaired olfactory identification in Asperger's syndrome. J Neuropsychiatry Clin Neurosci 15:105–107. [DOI] [PubMed] [Google Scholar]
  • 11. Gottfried JA (2010) Central mechanisms of odour object perception. Nat Rev Neurosci 11:628–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Schumann CM, Amaral DG (2006) Stereological analysis of amygdala neuron number in autism. J Neurosci 26:7674–7679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Kemper TL, Bauman ML (1993) The contribution of neuropathologic studies to the understanding of autism. Neurol Clin 11:175–187. [PubMed] [Google Scholar]
  • 14. Van Kooten IA, Palmen SJ, von Cappeln P, Steinbusch HW, Korr H, Heinsen H et al (2008) Neurons in the fusiform gyrus are fewer and smaller in autism. Brain 131:987–999. [DOI] [PubMed] [Google Scholar]
  • 15. Brang D, Ramachandran VS (2010) Olfactory bulb dysgenesis, mirror neuron system dysfunction, and autonomic dysregulation as the neural basis for autism. Med Hypotheses 74:919–921. [DOI] [PubMed] [Google Scholar]
  • 16. Talbot K, Woolf NJ, Butcher LL (1988) Feline islands of Calleja complex: I. Cytoarchitectural organization and comparative anatomy. J Comp Neurol 275:553–579. [DOI] [PubMed] [Google Scholar]
  • 17. Meyer G, Gonzalez‐Hernandez T, Carrillo‐Padilla F, Ferres‐Torres R (1989) Aggregations of granule cells in the basal forebrain (islands of Calleja): golgi and cytoarchitectonic study in different mammals, including man. J Comp Neurol 284:405–428. [DOI] [PubMed] [Google Scholar]
  • 18. Laufs H, Richardson MP, Salek‐Haddadi A, Vollmar C, Duncan JS, Gale K et al (2011) Converging PET and fMRI evidence for a common area involved in human focal epilepsies. Neurology 77:904–910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Loscher W, Ebert U (1996) The role of the piriform cortex in kindling. Prog Neurobiol 50:427–481. [DOI] [PubMed] [Google Scholar]
  • 20. Munoz‐Yunta JA, Palau‐Baduell M, Salvado‐Salvado B, Valls‐Santasusana A, Rosendo‐Moreno N, Clofent‐Torrento M et al (2008) [Autism, epilepsy and genetics]. Revista De Neurologia 46(Suppl. 1):S71–S77. [PubMed] [Google Scholar]
  • 21. Spence SJ, Schneider MT (2009) The role of epilepsy and epileptiform EEGs in autism spectrum disorders. Pediatr Res 65:599–606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Fujikawa DG, Itabashi HH, Wu A, Shinmei SS (2000) Status epilepticus‐induced neuronal loss in humans without systemic complications or epilepsy. Epilepsia 41:981–991. [DOI] [PubMed] [Google Scholar]
  • 23. Chen C, Shih YH, Yen DJ, Lirng JF, Guo YC, Yu HY et al (2003) Olfactory auras in patients with temporal lobe epilepsy. Epilepsia 44:257–260. [DOI] [PubMed] [Google Scholar]
  • 24. Lewine JD, Andrews R, Chez M, Patil AA, Devinsky O, Smith M et al (1999) Magnetoencephalographic patterns of epileptiform activity in children with regressive autism spectrum disorders. Pediatrics 104:405–418. [DOI] [PubMed] [Google Scholar]
  • 25. Rojas DC, Singel D, Steinmetz S, Hepburn S, Brown MS (2014) Decreased left perisylvian GABA concentration in children with autism and unaffected siblings. NeuroImage 86:28–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Rubenstein JL, Merzenich MM (2003) Model of autism: increased ratio of excitation/inhibition in key neural systems. Genes Brain Behav 2:255–267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Buzsaki G (2006) Rhythms of the Brain. Oxford University Press: Oxford. 448 p. [Google Scholar]
  • 28. Petrelli F, Pucci L, Bezzi P (2016) Astrocytes and microglia and their potential link with autism spectrum disorders. Front Cell Neurosci 10:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Vargas DL, Nascimbene C, Krishnan C, Zimmerman AW, Pardo CA (2005) Neuroglial activation and neuroinflammation in the brain of patients with autism. Ann Neurol 57:67–81. [DOI] [PubMed] [Google Scholar]
  • 30. Pekny M, Pekna M (2014) Astrocyte reactivity and reactive astrogliosis: costs and benefits. Physiol Rev 94:1077–1098. [DOI] [PubMed] [Google Scholar]
  • 31. Laurence JA, Fatemi SH (2005) Glial fibrillary acidic protein is elevated in superior frontal, parietal and cerebellar cortices of autistic subjects. Cerebellum 4:206–210. [DOI] [PubMed] [Google Scholar]
  • 32. Tetreault NA, Hakeem AY, Jiang S, Williams BA, Allman E, Wold BJ et al (2012) Microglia in the cerebral cortex in autism. J Autism Dev Disord 42:2569–2584. [DOI] [PubMed] [Google Scholar]
  • 33. Pigache RM, SpringerLink (Online service) (1970) The Anatomy of the Paleocortex: A Critical Review. Springer Berlin Heidelberg: Berlin, Heidelberg. Available from: 10.1007/978-3-642-65004-8. [DOI] [Google Scholar]
  • 34. Nieuwenhuys R, Huijzen C, Voogd J, SpringerLink (Online service) (2008) The Human Central Nervous System. Springer Berlin Heidelberg: Berlin, Heidelberg. Available from: 10.1007/978-3-540-34686-9. [DOI] [Google Scholar]
  • 35. Mai K Jr, Voss T, Paxinos G (2008) Atlas of the Human Brain, 3rd edn, Vol. VIII. Elsevier/Academic Press: London. 271 p. [Google Scholar]
  • 36. Paxinos G (1990) The Human Nervous System, Vol. XI. Academic Press: San Diego. 1195 p. [Google Scholar]
  • 37. Bekkers JM, Suzuki N (2013) Neurons and circuits for odor processing in the piriform cortex. Trends Neurosci 36:429–438. [DOI] [PubMed] [Google Scholar]
  • 38. von Bartheld CS (2012) Distribution of particles in the Z‐axis of tissue sections: relevance for counting methods. NeuroQuantology 10:66–75. [PMC free article] [PubMed] [Google Scholar]
  • 39. Ward TS, Rosen GD, von Bartheld CS (2008) Optical disector counting in cryosections and vibratome sections underestimates particle numbers: effects of tissue quality. Microsc Res Technique 71:60–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Garman RH (2011) Histology of the central nervous system. Toxicol Pathol 39:22–35. [DOI] [PubMed] [Google Scholar]
  • 41. Garcia‐Amado M, Prensa L (2012) Stereological analysis of neuron, glial and endothelial cell numbers in the human amygdaloid complex. PloS One 7:e38692. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. West MJ (2012) The precision of estimates in stereological analyses. Cold Spring Harbor Protoc 2012:937–949. [DOI] [PubMed] [Google Scholar]
  • 43. von Bartheld C (2002) Counting particles in tissue sections: choices of methods and importance of calibration to minimize biases. Histol Histopathol 17:639–648. [DOI] [PubMed] [Google Scholar]
  • 44. Nam IH, Kim DW, Song HY, Kim S, Lee KS, Lee YH (2012) Association of corpora Amylacea formation with astrocytes and cerebrospinal fluid in the aged human brain. Korean J Phys Anthropol 25:177–184. [Google Scholar]
  • 45. Hoyaux D, Decaestecker C, Heizmann CW, Vogl T, Schafer BW, Salmon I et al (2000) S100 proteins in Corpora amylacea from normal human brain. Brain Res 867:280–288. [DOI] [PubMed] [Google Scholar]
  • 46. Jaspar HH, Prick JJ (1968) [Morphology and histochemistry of the corpora amylacea]. Nederlands Tijdschrift Voor Geneeskunde 112:1153–1156. [PubMed] [Google Scholar]
  • 47. West MJ (2013) Optimizing the sampling scheme for a stereological study: how many individuals, sections, and probes should be used. Cold Spring Harbor Protoc 2013:521–532. [DOI] [PubMed] [Google Scholar]
  • 48. Gundersen HJ, Jensen EB (1987) The efficiency of systematic sampling in stereology and its prediction. J Microsc 147:229–263. [DOI] [PubMed] [Google Scholar]
  • 49. Casanova MF, Buxhoeveden DP, Switala AE, Roy E (2002) Minicolumnar pathology in autism. Neurology 58:428–432. [DOI] [PubMed] [Google Scholar]
  • 50. Courchesne E, Mouton PR, Calhoun ME, Semendeferi K, Ahrens‐Barbeau C, Hallet MJ et al (2011) Neuron number and size in prefrontal cortex of children with autism. JAMA 306:2001–2010. [DOI] [PubMed] [Google Scholar]
  • 51. Davies DC, Horwood N, Isaacs SL, Mann DM (1992) The effect of age and Alzheimer's disease on pyramidal neuron density in the individual fields of the hippocampal formation. Acta Neuropathologica 83:510–517. [DOI] [PubMed] [Google Scholar]
  • 52. Courchesne E (2002) Abnormal early brain development in autism. Mol Psychiatry 7(Suppl. 2):S21–S23. [DOI] [PubMed] [Google Scholar]
  • 53. Morgan JT, Chana G, Pardo CA, Achim C, Semendeferi K, Buckwalter J et al (2010) Microglial activation and increased microglial density observed in the dorsolateral prefrontal cortex in autism. Biol Psychiatry 68:368–376. [DOI] [PubMed] [Google Scholar]
  • 54. Terry RD, DeTeresa R, Hansen LA (1987) Neocortical cell counts in normal human adult aging. Ann Neurol 21:530–539. [DOI] [PubMed] [Google Scholar]
  • 55. Jacot‐Descombes S, Uppal N, Wicinski B, Santos M, Schmeidler J, Giannakopoulos P et al (2012) Decreased pyramidal neuron size in Brodmann areas 44 and 45 in patients with autism. Acta Neuropathologica 124:67–79. [DOI] [PubMed] [Google Scholar]
  • 56. Wegiel J, Schanen NC, Cook EH, Sigman M, Brown WT, Kuchna I et al (2012) Differences between the pattern of developmental abnormalities in autism associated with duplications 15q11.2‐q13 and idiopathic autism. J Neuropathol Exp Neurol 71:382–397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. Abel TJ, Hebb AO, Keene CD, Born DE, Silbergeld DL (2010) Parahippocampal corpora amylacea: case report. Neurosurgery 66:E1206–E1207. [DOI] [PubMed] [Google Scholar]
  • 58. Singhrao SK, Neal JW, Newman GR (1993) Corpora amylacea could be an indicator of neurodegeneration. Neuropathol Appl Neurobiol 19:269–276. [DOI] [PubMed] [Google Scholar]
  • 59. Kawamura T, Morioka T, Nishio S, Fukui K, Yamasaki R, Matsuo M (2002) Temporal lobe epilepsy associated with hippocampal sclerosis and a contralateral middle fossa arachnoid cyst. Seizure 11:60–62. [DOI] [PubMed] [Google Scholar]
  • 60. Ernst A, Alkass K, Bernard S, Salehpour M, Perl S, Tisdale J et al (2014) Neurogenesis in the striatum of the adult human brain. Cell 156:1072–1083. [DOI] [PubMed] [Google Scholar]
  • 61. Wesson DW, Wilson DA (2011) Sniffing out the contributions of the olfactory tubercle to the sense of smell: hedonics, sensory integration, and more? Neurosci Biobehav Rev 35:655–668. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62. Schmitz C, Rezaie P (2008) The neuropathology of autism: where do we stand? Neuropathol Appl Neurobiol 34:4–11. [DOI] [PubMed] [Google Scholar]
  • 63. Codagnone MG, Podesta MF, Uccelli NA, Reines A (2015) Differential local connectivity and neuroinflammation profiles in the medial prefrontal cortex and hippocampus in the valproic acid rat model of autism. Dev Neurosci 37:215–231. [DOI] [PubMed] [Google Scholar]
  • 64. Crawford JD, Chandley MJ, Szebeni K, Szebeni A, Waters B, Ordway GA (2015) Elevated GFAP protein in anterior cingulate cortical white matter in males with autism spectrum disorder. Autism Res 8(6):649–57. [DOI] [PubMed] [Google Scholar]
  • 65. Ikemoto S (2007) Dopamine reward circuitry: two projection systems from the ventral midbrain to the nucleus accumbens‐olfactory tubercle complex. Brain Res Rev 56:27–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66. Pannasch U, Freche D, Dallerac G, Ghezali G, Escartin C, Ezan P et al (2014) Connexin 30 sets synaptic strength by controlling astroglial synapse invasion. Nature Neuroscience (in press) [DOI] [PubMed] [Google Scholar]
  • 67. Casanova MF, Buxhoeveden DP, Brown C (2002) Clinical and macroscopic correlates of minicolumnar pathology in autism. J Child Neurol 17:692–695. [DOI] [PubMed] [Google Scholar]
  • 68. Casanova MF, Buxhoeveden D, Gomez J (2003) Disruption in the inhibitory architecture of the cell minicolumn: implications for autism. Neuroscientist 9:496–507. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Additional Supporting Information may be found in the online version of this article at the publisher's web‐site:

Additional file 1: Causes of death.

Click here for additional data file. (87.5KB, docx)

Additional file 2: Storage duration, PMI, and brain weight measures.

Click here for additional data file. (74.9KB, docx)

Additional file 3: Summed ADI‐R domain scores for autism cases.

Click here for additional data file. (30.5KB, doc)

Additional file 4: Anteroposterior mapping of the PiC in Nissl sections.

Click here for additional data file. (4.6MB, tiff)

Additional file 5: Differences in morphology with Nissl.

Click here for additional data file. (5.7MB, tiff)

Additional file 6: Individual neuronal and glial cell densities in layer II.

Click here for additional data file. (58.5KB, doc)

Additional file 7: Individual neuronal and glial cell densities in layer III.

Click here for additional data file. (57KB, doc)

Additional file 8: Morphology of abnormal cells in autism and epilepsy.

Click here for additional data file. (10.9MB, tiff)

Additional file 9: Glial cells co‐existing with corpora amylacea.

Click here for additional data file. (9.6MB, tiff)

Additional file 10: CD68 positive cells co‐occurring with corpora amylacea in layer I.

Click here for additional data file. (2.7MB, tiff)

Additional file 11: Corpora amylacea deposits in the olfactory tubercle in autism.

Click here for additional data file. (10.4MB, tiff)

Articles from Brain Pathology are provided here courtesy of Wiley

RESOURCES