White matter volume increase and minicolumns in autism

In Herbert and colleagues' recent article,1 the authors report a localized white matter volume increase in both autism and developmental language disorder. According to the researchers, only the later myelinating outer white matter compartment was affected. Otherwise, no significant changes were reported for the inner white matter compartment containing longer bridging connections, for example, corpus callosum. The finding may account for the postnatal brain volume increase observed in the brains of autistic individuals. We have recently described minicolumnar abnormalities capable of explaining gray/white matter variations in the brains of autistic individuals.2 These abnormalities consisted in minicolumns being smaller but less compact in their cellular configuration.3 A study using the gray level index (GLI) has confirmed the results.4 If minicolumns are both narrower and sparser in their internal cell distribution, as in autism, there may be no overall difference in neuron density.4 Estimates of cortical interconnectivity for the human brain suggest that each module is connected to on the order of 103 other modules.5 If connectivity were fixed as brain size increased, then the number of connections would scale as the number of modular units squared, that is, doubling the number of minicolumns quadruples the number of fibers (Fig). Thus, a large percentage of additional brain volume would go to maintaining the connectivity. Hofman5 suggests that the actual scaling follows a 3/2-power law; that is, a fourfold increase in modular units requires an eightfold increase in fibers. Encephalization has seen a disproportionate increase in white matter relative to gray matter. The proliferation of connections (lines) needed to maintain constant connectivity as the number of modules (points) increases. Each module communicates with 50% of the other modules in each instance. n = number of modules; m = number of interconnections. The additional white matter primarily takes the form of short-range associational fibers, which make up the bulk of white matter devoted to intracortical connections. Researchers postulate that the spatial layout of the brain minimizes total connection costs. Long-range connectivity incurs the penalties of increased conduction time and requiring a large volume of metabolically active tissue. As the brain enlarges, the distance between cells and columns increases so that attempts to maintain connectivity results in signal transmission delays and inefficiency. Modules with short-range interconnections reduce both conduction time and metabolic requirements. Long commissural pathways, such as the corpus callosum, may better serve as an index of macrocolumnar rather than minicolumnar abnormalities.6 In a study of the visual cortex of cats, only a minority of callosal axons perform strict point-to-point mapping of retinotopically corresponding sites. Many of the callosal terminals had widespread arbors capable of influencing several minicolumns. Development of callosal axons in cat visual cortex shows that from the earliest stages callosal fibers grow into the gray matter in macrocolumnar-like bundles. Manuel F. Casanova MD*, * Gottfried and Gisela Kolb Endowed Chair in Psychiatry, University of Louisville, Department of Psychiatry, Louisville, KY.