• A M Schrier and M L Povar.
• J Exp Psychol Anim Behav Process. 1982 Jan 1; 8 (1): 33-48.

AbstractFour experiments were conducted on stumptailed monkeys (Macaca arctoides) to determine whether the high levels of visual scanning (shifts in fixation from one discriminative stimulus to the other) seen during discrimination learning play a necessary role in this learning. In Experiments 1 and 2, the monkeys were given a series of two-choice, dot-pattern discrimination reversal problems. Normal visual scanning before a choice response was allowed during all but the reversal trials of half of the problems. On these latter trials, the discriminative stimuli were replaced by an uninformative stimulus after the animals made more than one visual fixation on each discriminative stimulus. Thus, on these trials, the animals were limited, in terms of the information received, to the empirically determined minimum number of scans necessary to maintain high levels of performance on such problems. In both experiments, which differed primarily in the number and type of uninformative stimuli used, the rate of reversal learning was markedly retarded by the experimental condition, with the effect persisting over a long series of problems. The magnitude of the effect was unrelated to the similarity of the uninformative stimulus to the discriminative stimuli. In Experiment 3, the monkeys were given a series of discrimination problems without reversals, during half of which the experimental condition was in effect. The results were similar to those of the first two experiments. Experiment 4 was similar to the preceding experiment except that, under the experimental condition, each trial began with the uninformative stimuli, which were replaced by the discriminative stimuli when visual scanning occurred. The uninformative stimuli had no clear-cut effect on discrimination learning in this experiment. These experiments indicate that the information provided by above-minimum levels of scanning is not necessary for discrimination learning per se, but it dose appear necessary for efficient discrimination learning.

10 keywords...

hide…