Pasternak T
J Comp Physiol Psychol. 1977 Jun;91(3):472-84. doi: 10.1037/h0077350.
Pigeons were trained preoperatively on a mixed delayed matching-to-sample (DMTS) task in which six different conditions were presented randomly; simultaneous matching, 0-, 1-, 2-, 4-, and 8-sec delays. Subjects that sustained extensive or complete damage to the visual Wulst and moderate damage to hyperstriatum ventrale showed a decrease in accuracy of performance to chance levels at all of the delay conditions as well as on simultaneous matching. After extensive retraining on the 0-sec-delay matching alone, performance on 0-sec-delay and simultaneous matching, presented in mixed DMTS, improved to between 70% and 90% correct. However, performance on delay conditions remained at chance level. All but one bird failed to show signs of postoperative improvement on delay problems in the course of the final testing. The data suggest that the conditional property of the task was a critical factor in the initial drop in accuracy on all of the presented problems. The relatively permanent loss of accuracy on all delay conditions is attributed mainly to the temporal separation of sample and comparison stimuli.
术前对鸽子进行了混合延迟匹配样本（DMTS）任务训练，其中随机呈现六种不同条件：同时匹配、0秒、1秒、2秒、4秒和8秒延迟。视觉上纹状体遭受广泛或完全损伤且腹侧超纹状体遭受中度损伤的受试者，在所有延迟条件以及同时匹配任务中，表现准确性均降至随机水平。仅对0秒延迟匹配进行广泛再训练后，在混合DMTS中呈现的0秒延迟和同时匹配任务的表现提高到了70%至90%的正确水平。然而，延迟条件下的表现仍维持在随机水平。在最终测试过程中，除一只鸽子外，所有鸽子在延迟问题上均未表现出术后改善的迹象。数据表明，任务的条件属性是所有呈现问题初始准确性下降的关键因素。所有延迟条件下准确性的相对永久性丧失主要归因于样本和比较刺激的时间分离。
