Auditory frequency generalization in normal and prefrontal dogs trained in varieties of active avoidance reflexes.

Twelve dogs were trained to avoid shock with response contingent termination of a 1000 Hz tonal CS on all trials, whereas 12 additional dogs received the same training except that on 50 percent trials there was a fixed CS duration and no shock. Initial acquisition revealed that CS prolongation resulted in longer response latencies compared to subjects trained with response contingent CS termination. Both training groups were then assigned to subgroups (n=4) that received either lateral or medial lesions of the prefrontal cortex or a rest pause of 10 days. Postsurgical reacquisition indicated a medial deficit in response latencies for both training groups whereas performance levels of normal and lateral groups were comparable. Two generalization tests to tones of 600-, 800-, 1000-, 1200-, and 1400-Hz followed. During the sampled test only one block of the five test frequencies occurred with the regular 15 training trials, while the tonal frequencies were presented on all 20 daily extinction trials during massed generalization testing. No evidence of extinction was observed during the sampled generalization test in either training group. However, during massed testing, all subjects trained with response contingent CS termination showed an overall extinction influence, which was most pronounced in the medial subgroup, although the laterals showed frequency control as well. In the group trained with CS prolongation the controls had clear, discriminative performance, while little stimulus control or extinction influence were found in the medial or lateral subgroups. Comparison with alimentary data indicated much broader tonal frequency generalization in dogs trained in avoidance. However, when the procedures reduced avoidance levels, double dissociation of medial and lateral lesion effects on stimulus control and susceptibility to extinction emerged.