Effects of natural categories on pictorial short term memory in pigeons.

四只鸽子被训练在相同/不同的任务中，并依次展示成对的幻灯片。在相同的试验中，一张与前一次试验相同的幻灯片被展示了10秒，而消失是有效的。在不同的试验中，对与前一次试验不同的幻灯片的键击以不规律的间隔强化，平均为6秒。幻灯片包括120张图片，如山、建筑物、动物和各种物体。在实验一中，训练结束后延长了试验间隔。尽管表现作为保持间隔的函数而下降，但鸽子即使在最长的50秒内也保持了相对较好的表现，比颜色或线条方向等物理刺激的时间更长。在实验2中，在相同的实验中旋转幻灯片，并对受试者进行正常定向、左右颠倒、上下颠倒、左右颠倒和上下颠倒的测试。受试者先测试了90张新奇图片，然后再测试了90张熟悉的图片。受试者在反转新奇图片时的表现不太准确，而且上下颠倒对表现的负面影响比左右颠倒更大。被试与熟悉的图片保持着准确的操作，而熟悉图片的颠倒对操作没有影响。因此，鸽子对幻灯片识别的记忆相对较长，熟悉图片的旋转不会改变识别幻灯片所需的信息。受试者可能使用整个幻灯片的信息来识别幻灯片，而不是使用图片的一小部分。

Four pigeons were trained in a same/different task with successively presented pairs of slide pictures. On same trials, a slide identical to that of the previous trial was presented for 10 sec while extinction was in effect. On different trials, keypeckings to a slide different from that of the previous trial were reinforced at irregular intervals averaging 6 sec. The slides were a wide variety of 120 pictures such as mountains, builidings, animals, and miscellaneous objects. In Experiment 1, intertrial interval was lengthened after completion of the training. Although per-formance decreased as a function of retention interval, the pigeons retained relatively good performance even at the longest in terval of 50 sec, longer than with physical stimuli such as color or line orientation. In Experiment 2, the slides were rotated on same trials, and the subjects were tested with normally oriented pictures, left-right reversals, top-bottom reversals, and left-right and top-bottom reversals. The subjects were tested with 90 novel pictures and then with 90 familiar pictures. The subjects performed less accurately with reversed novel pictures, and top-bottom reversal more adversely affected performance than left-right reversal. The subjects maintained accurate performance with familiar ones and reversals of familiar picture had no effect on performance. Therefore, relatively long memory was in effect for pigeons' slide identification, and rotation of familiar pictures did not change the information necessary for slide identification. The subjects may have used the information of a slide as a whole, rather than small parts of picture, for slide identification.