تفاوت نیمکره غربی در آموزش معکوس سریالی: یک مطالعه با بیماران کامیشروتومی

A serial reversal learning task involving tactile-proprioceptive discrimination and manual responses was presented alternately to the left and the right hemisphere of four ‘‘split-brain’’ patients. In each trial, one of two rods that differed in both diameter and surface texture was placed in the patient’s hand out of view. The patient was trained to match it to samples according to either size or texture and non-verbal audio–visual feedback was used to signal the correctness of each response. After reaching five consecutively correct responses, the feature to be matched was switched. When the patient again made five consecutively correct responses, the feature to be matched was reversed back. This procedure was repeated until the end of a 200-trial training run. The two hemispheres learned equally readily on the first learning task. The right hemisphere had much greater difficulty in learning the reversals than the left hemisphere and this was not attributable to a strong tendency to stay with a previously correct match. Learning with the left hemisphere showed relatively stable performance across successive reversals, whereas that with the right hemisphere showed high lability. Control trials showed that the hemispheres were equally competent in making the basic tactile-proprioceptive discriminations. Comparisons with the findings on (a) three control patients and (b) training with unrestricted visual input showed that learning with two hemispheres was easier than learning with either one alone; performance regulated by both hemispheres was also more stable.

Most studies on the functional asymmetry between the two cerebral hemispheres have employed a testing approach commonly practiced in the assessment of human abilities: Each response is recorded and scored, but there is no immediate feedback about its correctness to the subject. Although much has been learned about the propensities and abilities of each hemisphere (e.g. see 11, 15 and 16for general reviews), this testing approach has yielded little direct information on how the two hemispheres might differ in their ability to correct errors when given feedback. The readiness with which behavior can be modified to suit changing conditions is important for adaptation and survival and this learning ability has often been used to distinguish different animal species along the phylogenetic scale. In the present study, a training procedure commonly used in non-human animal studies was employed to compare (a) the learning abilities of the left and the right hemisphere and (b) learning with either hemisphere alone vs learning with both hemispheres. A two-choice learning task was presented, followed by a series of reversal training. Comparisons were made on the abilities to learn (a) the initial task and (b) the serial reversals.