A topographic study of minicolumnar core width by lamina comparison between autistic subjects and controls: possible minicolumnar disruption due to an anatomical element in-common to multiple laminae

Abstract

Radial cell minicolumns are basic cytoarchitectonic motifs of the mammalian neocortex. Recent studies reveal that autism is associated with a "minicolumnopathy" defined by decreased columnar width and both a diminished and disrupted peripheral neuropil compartment. This study further characterizes this cortical deficit by comparing minicolumnar widths across layers. Brains from seven autistic patients and an equal number of age-matched controls were celloidin embedded, serially sectioned at 200 microm and Nissl stained with gallocyanin. Photomicrograph mosaics of the cortex were analyzed with computerized imaging methods to determine minicolumnar width at nine separate neocortical areas: Brodmann Area's (BA) 3b, 4, 9, 10, 11, 17, 24, 43 and 44. Each area was assessed at supragranular, granular and infragranular levels. Autistic subjects had smaller minicolumns whose dimensions varied according to neocortical area. The greatest difference between autistic and control groups was observed in area 44. The interaction of diagnosis x cortical area x lamina (F(16,316) = 1.33; P = 0.175) was not significant. Diminished minicolumnar width across deep and superficial neocortical layers most probably reflects involvement of shared constituents among the different layers. In this article we discuss the possible role of double bouquet and pyramidal cells in the translaminar minicolumnar width narrowing observed in autistic subjects.

Figure 1

The Gibbs energy function of a sample neighborhood (outlined gray boxes) is illustrated, left. Also shown is the neighborhood system, in the form of the weight b assigned to each pixel adjacent to the central pixel X. The range of the function is determined by the lowest and highest intensities of the pixels in the neighborhood, N. The energy is calculated for every integer value in the range, and the intensity corresponding to the minimum is chosen as the new gray level. An image is shown before (center) and after (right) five iterations of the noise reduction algorithm. No contrast enhancement was done, in order to illustrate the effects of smoothing alone.

Figure 2

The low contrast of a raw image (top) is evident in the narrow range of its gray level histogram. Although two peaks (modes) are present, each corresponding to one region, the boundary between them is indistinct. A threshold placed at the local minimum separating the peaks would cause many pixels to be mislabeled. After five iterations of noise reduction and contrast enhancement (bottom), there is a clear separation between modes associated with each region of the image.

Figure 3

Boundaries of cells from the segmented, contrast‐enhanced image are superposed on the original image (left). Objects are classified by size such that anything smaller than 10 µm in diameter (black outlines) is considered not a pyramidal cell and excluded from measurement. Cell centerpoints (middle) are determined as the point inside each pyramidal cell profile with relative maximum Euclidean distance from the boundary. In the case of multiple relative maxima, the centerpoint is the centroid of the set of maximizers. The line tracing method finds two minimal distance paths (right), from centerpoint to centerpoint, spanning the lamina, for each seed point on the laminar boundary. Core width is the average distance between the boundary lines and their median (not shown), or simply half the average distance between boundary lines.

Figure 4

Mean minicolumnar core width by lamina and cortical area. Errorbars are +1 standard deviation.