Subchronic cocaine produces training paradigm-dependent learning deficits in laboratory rats.

The effect of cocaine on spatial learning was investigated by exposing male Sprague-Dawley rats to 0, 20, or 40 mg/kg cocaine prior to and during training on a water maze task. Half the animals were pretrained on cued trials prior to hidden platform trials, while the remaining animals completed hidden platform trials immediately. Escape latencies for all animals improved with training, but pretrained animals located the hidden platform faster than untrained animals (P<.001). Pretraining also decreased the effect of cocaine. In pretrained animals, only the high dose of cocaine caused significant increases in escape latency (P<.001), while in the untrained group the lower dose of cocaine also caused a significant increase (P<.001). On working memory measures, cocaine affected both the pretrained (P<.01) and untrained (P<.001) groups. Dwell ratio measurements indicated unaffected reference memory in both pretrained (P<.001) and untrained (P<.001) animals, and no significant differences were detected among the treatment conditions in either group (P>.05). Thus, while cocaine did not abolish learning, the efficiency with which the task was learned was compromised. However, this effect was reduced by pretraining.