Minicolumnar Core Width by Lamina in Brains of Patients with Autism

Background: Minicolumns are modular elements of organization found within the neocortex of all examined mammalian species. Within a minicolumn there may be distinguished a core, comprising the cell somata, axons, and apical dendrites of pyramidal cells, and a periphery, comprising other processes and few cell somata, mostly interneurons. Postmortem studies examining the radial organization of pyramidal cell arrays in people with autism suggest the presence of smaller minicolumns, compared to normal controls.

Objectives: This study further characterizes this cortical deficit by comparing core width (w) minicolumnar width across laminae. It was our initial hypothesis that minicolumnar width differences in autism would be most prominent in supragranular laminae.

Methods: Brains from six autistic patients (diagnosed according to DSM IV-TR and ADI-R) and an equal number of age-matched controls were celloidin embedded, sectioned at 200 μm or 500 μm, and Nissl stained with gallocyanin. Photomicrographs mosaics of the cortex from nine different brain regions (cortical areas 3b, 4, 9, 10, 11, 17, 24, 43, and 44) were subjected to computerized image analysis for minicolumnar core width. Statistical analysis employed a linear mixed effects model, with case-control pair as a random effect and fixed effects diagnosis (autistic disorder or none), cortical area, and lamina (II+III, IV, or V+VI) together with all interaction effects, and also sex and cerebral hemisphere.

Results: Statistically significant differences were found by diagnosis (F_1,263 = 554.0; P < 0.0001), cortical area (F_8,263 = 2.69; P = 0.0073), lamina (F_2,263 = 630.0; P < 0.0001), diagnosis×cortical area interaction (F_8,263 = 4.95; P < 0.0001), and diagnosis×lamina interaction (F_2,263 = 45.06; P < 0.0001). The interaction of all three effects was not statistically significant (F_16,263 = 1.55; P = 0.0836). Autistic subjects had core widths measuring 69 %, 74 %, and 65 % of those in matched controls in supragranular laminae, lamina IV, and infragranular laminae, respectively. The greatest differences between autistic patients and controls were found in area 44, where w = 4.84 μm and w = 8.98  μm, respectively. Within any given region and diagnostic category, it held without exception that core width increased with depth: w_V+VI > w_IV > w_II+III.

Conclusions: Previous studies suggest that minicolumnar narrowing in autism is accounted for the greater part by reduction of peripheral neuropil space. We have now shown some diminution in the minicolumnar core. Our initial hypothesis was not borne out, and the fact that minicolumns appear to be diminished in width across laminae suggests involvement of a shared constituent among the different layers, e.g., double bouquet and pyramidal cells. Opercular area 44 is part of Broca’s area in humans. The role of this area in language processing, speech production, and understanding the intention behind observed motor acts makes the reported finding a potential clinicopathological correlate to autism.