Quantitative differences in performance on abstract discriminations using one or both hemispheres

Abstract

Cats were trained to a stringent criterion on abstract two-choice visual pattern discriminations and were then tested with task information relayed to both hemispheres, or to only one hemisphere. Essentially normal, high-level response was possible with cue inputs to one eye only (the other occluded by a mask), even following section of the optic chiasm or the corpus callosum. If both structures were cut, so that unilateral inputs were restricted to one hemisphere, however, there was a sizable permanent decrease in accuracy of performance (accuracy under the bilateral inputs condition remained high). Performance on simple discrimination tasks, employing patterns with the same or greater acuity demands, was not affected by such restriction of inputs to one hemisphere. This finding, and the fact that some subjects can perform at a normal preoperative level after chiasm section, rule out explanation of the loss in single-hemisphere abstract task performance solely in terms of destruction of visual fibers. A more plausible interpretation of the results is provided in terms of a principle of cortical mass action, which assumes that the asymptotic level for performance on complex tasks may depend on the amount of cortex participating in the processing of task information.