Given that one hemisphere is better than the other at performing certain tasks, it might be expected that the memories related to these tasks would be associated with (stored in?) the more proficient hemisphere. There are, however, many complications in endeavoring to establish the relevant facts (see Doty and Ringo, 1994). In the first place, it is clear that in macaques, either hemisphere has access to the mnemonic store of the other via either the anterior commissure or the corpus callosum (e.g., Doty, Ringo, and Lewine, 1988, 1994). For some human patients the anterior commissure seems incompetent in this regard (e.g., McKeever, Sullivan, Ferguson, and Rayport, 1981), although this may reflect merely a sampling problem in that the physical size of the human anterior commissure has a sevenfold variation (Demeter, Ringo, and Doty, 1988).
Being able to forestall the creation or utilization of an effective memory trace in either hemisphere by teta-nization of perihippocampal structures, Ringo (1993) has greatly clarified the situation. In animals with a transected optic chiasm and either of the forebrain commissures, it could be shown that visual input into one hemisphere established a memory trace in both; and when, by use of the tetanization, the trace was limited to the initially viewing hemisphere, it could subsequently be accessed via the other hemisphere. Because of the possibility that any input with the forebrain commissure intact might establish a bilateral engram, this was the first clear demonstration that in "normal" circumstances a unilaterally established engram could be activated via an "ignorant" hemisphere.
Unilateral engrams have been established in other circumstances. Perhaps the most interesting is that of Kucharski and Hall (1988), who established an olfactory discrimination via one nostril in rat pups. Before maturation of the anterior commissure the memory was inaccessible via the other nostril, but subsequently, it could be activated via this untrained side. Thus, a memory trace that is established even before the existence of effective interhemispheric connections could still be found and utilized once the appropriate fibers were functional. Such was not the case, however, when more complex behavior was involved. Rudy and Paylor (1987) found that young rats that were trained on a water maze with one eye open before maturation of the callosum did not benefit when tested with the other eye after the callosum was functional.
Doty, Negrao, and Yamaga (1973) used macaques having either anterior commissure or callosum cut before experimentation but with intact optic chiasm. When trained to respond to electrical excitation of striate cortex in one hemisphere, animals continued doing so to excitation of striate cortex in the other hemisphere only as long as the corpus callosum was intact, while continuing to respond to nontrained striate loci in the original hemisphere. Such unilaterality was not true in the case of the anterior commissure, in which responding continued unabated for stimulation of the untrained striate cortex after the commissure was cut (completing transection of both forebrain commissures). It is speculated that it is the continuing input to both striate cortices, both eyes being open and callosum intact, that produced the confinement of the engram to the stimulated hemisphere. Unlike the situation in the Ringo (1993) experiments with split chiasm, in which the non-viewing eye is occluded, the nontrained hemisphere here would be occupied by its normal visual processing and unavailable for further encoding. On the other hand, it is equally plausible that the absence of callosal connections between striate cortices might underlie the unilat-erality. Ringo, Doty, and Demeter (1991) showed with split-chiasm macaques with intact forebrain commissures that if the two hemispheres concurrently viewed congruent images (superimposed so that the correct response could be obtained via either or both hemispheres), there was an advantage over viewing the same material by either hemisphere alone; but if the images were noncongruent, that is, different images were presented to each hemisphere, this advantage was lost. This is somewhat reminiscent of the above in that two greatly differing activities, electrical excitation and normal vi sual input, cannot be processed concurrently via the callosal path. However, also in line with the above, this was not true when only the anterior commissure was available. In that case binocular viewing was advantageous even though each hemisphere concurrently viewed a different image!
Perhaps even more surprising is the fact that binocular viewing held a slight advantage even with split-brain animals. This might simply arise because of absence of interference from the hemisphere that is literally kept in the dark on monohemispheric trials. However, there is evidence for participation of brain stem elements in mnemonic processing, such that there is a limited-capacity system shared by the two hemispheres (Lewine et al., 1994). The experiments on macaques are somewhat complicated, since they involved giving the animal a set of two to six target images that it had to remember temporarily and distinguish from nontarget items. Such a task is noted for the fact that the time required to reach a decision on a given item shows an increment that is linearly proportional to the number of target items. Accuracy of the decision also diminishes proportionally with the number of targets. The experiments sought to determine what happens, in split-chiasm animals with either the splenium of the corpus callosum or the anterior commissure intact, when some of the targets to be remembered are presented to one hemisphere and some to the other. The result was that either commissural system was effective in unifying mnemonic load, so that response via either hemisphere reflected the total number of targets regardless of how they had been distributed between the hemispheres initially. However, when the hemispheres were isolated from each other by completion of the commissural transection, the reaction time for each hemisphere reflected only the number of targets that it had been given; but the accuracy of the response depended on the total mnemonic load given to the two hemispheres together! Thus, even in the absence of the forebrain commissures, some subcortical system was unifying an aspect of the mnemonic processing.
We (Kavcic, Fei, Hu, and Doty, 2000) have recently derived further evidence, on two split-brain macaques, that processes are in play that endeavor to form a unified engram from images viewed simultaneously by the isolated hemispheres. First, if the two hemispheres viewed the same image, there was commonly better performance than if each hemisphere viewed an entirely different image. However, if both hemispheres had concurrently viewed different images and were then asked to recognize a repeat, simultaneous presentation of these same two images, performance was excellent. Indeed, as was noted above, it was significantly better than when either hemisphere alone viewed an image on each occasion. The critical test arose when the two hemispheres at first viewed images simultaneously (parallel processing), and then only one of them alone was called upon subsequently to recognize the image that it had seen in conjunction with the other. Performance in such case was significantly worse than when either hemisphere worked alone on a single image. Thus, in sum, bihemi-spheric performance is excellent, showing that there is no deficiency arising from parallel processing per se, but when only half of the simultaneously experienced visual input is present, recognition is worse than if there was only a single input to a single hemisphere on each occasion. This strongly suggests that when the two isolated hemispheres view two images concurrently, there is some amalgamation of the separately derived memory traces, yielding a deficiency in recognition when only half of that scene is viewed (by either hemisphere).
Was this article helpful?