The author investigated the property of the muscle tension control in the isometric condition without external feedback. The integrated electromyogram was recorded from electrodes on the hand muscles (m. interosseus dorsalis manus primi, m. abductor digiti minimi manus). Subjects accomplished a series of tasks in which they controlled their muscular tension with movements of a target on a computer. The effects of direction of tension (increase and decrease) were studied under various conditions including different control velocity, different types of visual guide on the screen, and range of muscle tension required. The results showed that in the decrease phase the results of all subjects fitted to the exponential curves in any other conditions. However, in the increase phase the data did not fit well, and no specific patterns appeared. The results are ascribed to the property of muscle control (feedback control and feedforward control), and discussed in relation to the mechanism of the force sensation.
Four homing pigeons were trained to peck a response key to obtain food on several random-ratio and variable-interval schedules. A peck on another key(timeout key) produced a 20-second timeout period accompanying a visual stimulus change. As the variable-interval schedule was increased logarithmically through five values from 60 to 960 sec, frequency of timeout responses increased monotonicaly in two pigeons. In random-ratio schedule, no bird showed systematic change in the frequency of timeout responses when the schedule value was increased in successive stages from 50 to 300. No bitonic relationship was obtained between the schedule value and the frequency of timeout responses. The timeout responses during both schedules were found to be distributed throughout the inter reinforcement interval. The results suggest that the interpretation of timeout behavior in terms of schedule-induced or adjunctive behavior is open to considerable question.
MM, an early blind subject, received an optical iridectomy in her right eye at the age of 12. Preoperatively, her right eye had only light perception. Immediately after the operation, she could distinguish brightness, but not color. In order to improve her visionsign system activity, we planned to develop her visual discrimination between vertical and horizontal strips. On the first task, she was asked to distinguish through the matching-to-sample method, a black vertical strip from a horizontal one (7.5cm×1.0cm in size) attached to a white paper. Even after 5 months of training, she was unable to reach the criterion of successful discrimination. On the second task, she was presented with green strips (12.0cm×2.0cm in size) attached to a white paper, in addition to tactile samples aiding the discrimination task. Performance only improved, when the supplementary tactile sample was attached to the rear of the stimulus card. With such tactual aid she gradually attained the ability to discriminate between two directions over a period of 4 months. In the final stage of this task, it was observed that her activity of exploring the edge of the figure-area, and of tracing from one edge of the strip to the other, gradually became a systematic one. A theoretical implication concerning the changes in her exploratory activities is briefly discussed.
In the first experiment the apparent lengths of the minor diagonals and major diagonals of two parallelograms of the Sander figure were measured. The minor diagonals were overestimated and the major diagonals were underestimated. According to the law of good gestalt, a distorted form has a tendency to be restored to the good form, and the illusion is caused by this restoration tendency. If so, it may be hypothesized that the more distorted a figure is, the greater the illusion magnitude is. In the second experiment various distorted parallelograms were constructed, and the apparent lengths of the minor or major diagonals of those various parallelograms were measured. It was found that the greatest overestimation of the minor diagonal and the greatest underestimation of the major diagonal occur in the most distorted parallelograms irrespective of the relative sizes of the areas of the parallelograms.
Previous works have indicated that Ss could classify the sets of figures which comprised the distorted patterns from a protyotype figure. In addition, Ss classified more easily the prototypes as compared with the distorted ones out of the learning set. In the present experiment, clearly-distinguishable categories (prototypes) were used and a large number of over-learning trials were provided to the Ss. Then the prototypes were classified as accurately and quickly as the "old" distorted figures which had been comprised in the learning set. This result suggests that Ss can utilize the prototype information which were abstracted from the original learning sets, for classifying new figures.