This is a continuation of the previous study (The Japanese Journal of Psychology, 1967, Vol. 38) on the figure-effect in the third dimension of visual space. The purpose of this study is to examine the effect of a light stimulus figure upon the various positions in the depth direction of space. The figure-effect was measured by the light threshold method: a small light point (the test light) was presented in front of the stimulus figure, and the light threshold value t of the point was measured (Fig. 1, 5 and 8). Other experimental procedures were virtually the same as those in the previous report. In view of the results of the previous study, three experiments were designed to clarify the following points. 1. In experiment I, whether the gradual elevation of the threshold with increase of the depth distance from the stimulus figure can be seen in the different stimulus figure from that of previous study. 2. In experiment II, how the threshold values would change depending on the distance from the stimulus figure in frontal planes parallel to the plane containing the figure. The light stimuli used in experiment I and II were line figures and the test light was presented at positions lateral-frontal to the figure. 3. In experiment III, what aspects of threshold values can be seen in the frontal space within a contour figure. Large and small figures of circular contour were used as light stimuli. The threshold value was measured on the front space at each of the three positions p0, p1, p2 inside a circle as shown in Fig. 8. The results of these experiments were as follows: 1. The gradual elevation of the threshold value t in proportion to the increase of the depth distance from the stimulus figure was not observed in the lateral-frontal space of the line stimulus (Fig. 2, 3 and 4). 2. The threshold values in the frontal parallel planes displayed a falling logarismic curve as the distance from the figure was increased (Fig. 6, 7). 3. In the central axis (p0) of the circle, the threshold value elevated slightly as shown in Fig. 9, 10, but at places near the contour (p1, p2), an elevation was seen clearly as shown in Fig. 11, 12. From these results, it was concluded that the figure-effects in the depth, i.e., in the third dimension, of visual space were present not only in the light stimulus of two parallel lines used in the previous study, but also in other kinds of light stimulus, and the appearance of the elevation of the threshold value depended on stimulus configuration. The other finding was that the figure-effect presented a different feature in the depth dimension from that in the frontal parallel plane of visual space. But at present, it cannot easily be concluded that the figure-effect in the depth differs from that in the frontal parallel plane in terms of its function. It is supposed that the examination of the mechanism of the two effects is needed in the future.
In the information processing approaches to component aspects of human pattern recognition, it is necessary to give careful consideration to types of information to be processed and also information sources or stimulus attributes yielding information. For these information characteristics, different processing modes and processing rates may be observed behaviorally. The present study investigated two modes of spatial information processing for an array (two rows of 1 to 8 items in each) of colored alpha-numeric stimuli as measured by the “same-different” reaction time (RT). If both rows had exactly the same arrangement with respect to a predetermined stimulus attribute of either class (letters and digits) or color (red and blue), in other words, if every pair of the upper and lower items was the same with respect to a given attribute, the “same” response was to be made, and if otherwise, the “different” response was to be made. Ten kinds of stimulus arrays were prepared as well-counterbalanced examples for each condition of Item-Pairs (8)×Attributes (2)×“Same”-“Different” Responses (2). Generally, a half of each row consisted of letters and the other half digits, both in red and blue an equal number of times. Arrays for the “different” condition had one “different” pair of items at a random position. On a given trial, following a verbal instruction of “ready” and appearance of a fixation point, an array was presented, to which the S made a key-response, and a feedback light immediately informed S whether his response had been correct or not. Eight undergraduates served as Ss. Main results are summarized as follows: RTs were significantly affected by the number of item-pairs, the functional relation between the number of item-pairs and the RT being linear (Fig. 2). In the class attribute, “same” RTs increased at the rate of about 280msec/pair as item-pairs increased, and “different” RTs at the rate of about 160msec/ pair. In the color attribute, the rate was about 50msec/pair without regard to response condition. These findings indicated that the spatial informations from the class attribute of arrays were processed serially, and that those from the color attribute were processed very rapidly, though it could not be determined whether the processing mode was serial or parallel. The difference between the processing rates of “same” and “different” responses in the condition of class attribute was considered to be an evidence of the serial self-terminating mode. This interpretation received further support to some extent from the data in Fig. 3. However, the data also indicated that the serial mode did not always have a fixed order on every trial but that the serial order varied from trial to trial. Consideration of the simple RT (about 240msec) and the data in Fig. 2 suggested that a central processing time for naming and categorizing every item-pair was approximately 250msec or less, and that other internal processes required approximately 380msec or less in the present experimental conditions. The central processing time was not likely to be subjected to the speed-error tradeoff.