Factors shaping the corpus callosum

Perhaps because of its prominence in midline sections of the human brain, the corpus callosum, one of the few brain structures identifiable by even the slowest student, has attracted wide interest throughout history. Cajal described how the visual field representation was first inverted by the lens of the eye and split into two halves, represented in the two hemispheres, and then knit together by the corpus callosum into a single representation of the visual world. This general idea has fared well since that time (Sperry, 1962; Berlucchi, 1981). In many species, including the cat and lower primates, the representation of the visual field is not cleanly split between the two hemispheres but includes a region of overlap, represented in both (Illing and Wässle, 1981). The callosum was envisioned as providing corticocortical connections to this region of overlap. As the nature of corticocortical connections within each hemisphere was elucidated, it became possible to entertain the hypothesis that the corpus callosum consisted merely of particularly long fibers that played the same role between the hemispheres as was played by the local connections within each hemisphere (Innocenti, 1986). Insofar as local connections within a hemisphere exhibit various types of specificity according to functional cortical modules, such as orientation and ocular dominance columns, it was not surprising that investigations of the callosum suggested similar specificity. In this view, it did not matter that the visual field was split vertically and that the representations of the primary and secondary visual areas (areas 17 and 18) were joined at the representation of the vertical meridian, the precise point at which the visual field representations were divided between the two cortical hemispheres. Symmetrical corticortical connections could have joined any split in the representation of the visual field. Recent work has shown that the real connection between the hemispheres is different from this simple model in at least three ways. First, although the cells that project through the corpus callosum are fairly widespread, their terminals are strongly focused on a narrow transition zone (TZ) between areas 17 and 18 that is part of neither area (Payne, 1990, 1991; Payne and Siwek, 1991). This transition zone represents the ipsilateral visual field principally through the contralateral eye, owing to the fact that a region of temporal retina near the area centralis projects both contralaterally and ipsilaterally. Second, connections between the two hemispheres are not bilaterally symmetrical. Instead, these connections link topographically corresponding positions, that is, positions in the two hemispheres that are concerned with the same area of the visual field (Olavarria, 1996). For example, callosal cells located laterally in the callosal zone, i.e., in the TZ where the ipsilateral visual field is represented, project to a medial region of the callosal zone of the other hemisphere, containing a representation of the same area of the visual field. Third, on page 441–457 of the present issue, Olavarria shows that callosal connections are not just specific for the same part of the visual world; they also specifically connect ocular dominance (OD) columns serving the same eye. This stands in contrast to intrahemispheric connections that are predominantly orientation-selective. Strikingly, the eye specificity of the callosal pathway results in cells in the TZ occupying contralateral eye ocular dominance columns, whereas those in immediately adjacent portions of areas 17 and 18 proper are found primarily in the ipsilateral eye columns. Without the careful distinction made between the TZ and areas 17 and 18 proper, it would have appeared that callosal connections were not specifically related to ocular dominance columns, and their fine topography would have been overlooked. Figure 1 summarizes Olavarria’s findings; it illustrates the anatomical organization of callosal connections at the area 17–18 border in relation to topography and ocular dominance. The ocular preference of callosal cells has prompted the author to suggest a mechanism for the selection and stabilization of callosal connectivity during development. According to Olavarria’s “uniocular” hypothesis, the key players are the bilateral projections originating from the temporal retina. Coherent activity between ganglion cells in the same eye is conveyed through the geniculocortical pathway to different points in the cortex in the two hemispheres, and callosal fibers connecting these points are stabilized.