The partitioning problem

It is natural to conceive of the corpus callosum as a collection of function-specific channels. But how do we define them? One logical way is by the cortical modules that the channels interconnect. In principle we can trace such cortico-cortical connections with retrograde degeneration staining methods, but such studies are scarce in humans (see S. Clarke's commentary in this section). A complementary source of data are behavioral discon nection effects associated with localized callosal lesions. Another approach is by the regional morphology of callosal fibers. This is the approach taken by Aboitiz and his associates. They classified callosal axons by their size and myelination and showed a differential distribution of different fiber types in different callosal regions, thus supporting the channel doctrine. Small-diameter fibers that interconnect association cortex predominate in anterior and posterior callosal regions, whereas large-diameter myelinated fibers predominate in the body of the callosum. This leaves open two critical questions. First, which fibers interconnect homotopic areas and which interconnect heterotopic areas? Aboitiz and colleagues focus their discussion on the functional significance of homotopic connections, but S. Clarke challenges this. Thus, she challenges the channel doctrine, with the possible exception of the genu and the sple-nium. Second, the anatomy does not tell us which callosal fibers serve to facilitate interhemispheric transfer and which serve to inhibit or modulate it.

A third approach to decomposition is by factor analysis of regional size differences in the corpus callosum. This has the advantage that the data themselves, rather than our preconceptions, generate the hypothesis. The rationale is that callosal regions that are functionally independent vary more across individuals than do regions that are functionally homogeneous. This, in turn, may reflect differential developmental trajectories for different callosal channels. This approach has been applied to callosal area but not yet to callosal shape, where points of maximum change may signal regional boundaries. Factor analysis of regional size has not been rewarding about function (cf. Cowell, this section; Clarke, 1990). Clarke (1990) partitioned 60 corpus callosum into 30 radial regions. Principal component analysis of absolute area measures revealed one main factor associated with overall callosum size. Principal component analysis of areas normalized for total callosum size revealed two main factors. The first factor includes the anterior half of the corpus callosum without the rostrum. This region contains fibers that originate primarily from the frontal lobes. It is surprising that the factor does not distinguish a subregion associated with motor functions from a subregion associated with prefrontal, executive functions. The second factor, including the posterior midbody, may be associated with somesthetic functions (Pandya, Dye, and Butters, 1971). Clarke studied the correlations of these two callosum factors with behavioral laterality effects observed in a set of multimodal tests that tap hemispheric specialization and callosal connectivity in vision, audition, somesthesis, and language. Out of 44 correlations, the only significant one (negative) was between the left-hand condition of a roughness discrimi nation test and the anterior callosal factor normalized for total callosal area (see below). Clarke also administered a battery of cognitive tests that measured handedness, visual, perception, verbal production, memory, and executive functions. There were no significant correlations between any of the cognitive measures and the two factors. Thus, this anatomical-behavioral approach to callosum partitioning was not successful.

Note that Clarke's factor analytic approach assumes that functionally unitary regions of the callosum form a continuous surface, with no holes. This might not be the case. For example, sensorimotor integration appears to be controlled by a frontal premotor-inferior parietal circuit, which has strong callosal connections (Iacoboni, Woods, and Mazziota, 1998). It is not known whether the interhemispheric connections of the circuit are implemented through a single channel that sends both homotopic and heterotopic projections to parietal and frontal areas or whether the connection consists of two or more noncontiguous channels.

A fourth approach is to partition the corpus callosum by its behavioral correlates. The idea is to obtain cal-losum morphometry and behavioral laterality measures tests for many subjects and intercorrelate them to find which behaviors are associated with which regions. There is a dual rationale for this approach. First, behavioral tests of transfer of information in a specific modality or a specific code should be associated with a specific callosum channel. Unfortunately, this assumption is not likely to be true, at least for sensorimotor fibers. That is because Aboitiz and his associates (see their chapter in this section) found that regional cal-losum size is not correlated with the number of large-diameter myelinated fibers that interconnect primary sensorimotor cortex. The assumption may be valid for small-diameter fibres and more abstract codes, although our understanding of abstract callosum channels is rudimentary.

The second rationale for the function/structure correlations is the hypothesis that anatomical-functional asymmetry of cortical modules is inversely related to the degree of connectivity of the callosal channels that interconnect the modules. This hypothesis was first articulated by Geschwind and Galaburda (e.g., Galaburda and Geschwind, 1980). It has received support from anatomical studies in animals (Rosen, Sherman, and Galaburda, 1989), from anatomical studies in humans (Aboitiz et al., 1992b, and chapter in this section), and from behavioral studies in humans (Clarke, 1990; Clarke and Zaidel, 1994). The chapter by Aboitiz and colleagues is distinguished by bringing an evolutionary perspective to bear on this issue. One particular hypothesis discussed by Aboitiz and colleagues and due to

Ringo and colleagues (1994) posits that as brain grows in size through phylogeny, callosal connectivity decreases owing to increased distance, and this, in turn, leads to increased hemispheric specialization. The chapter by Jancke and Steinmetz in this section cites data in support of this hypothesis.

Was this article helpful?

0 0

Post a comment