Abstract

A new avoidance-motivated detour-maze in which memory for an immediately preceding sample event could be assessed was evaluated by testing seven 6-month-old male F-344 rats with a delayed matching-to-sample (DMTS) paradigm. Rats first received extensive pretraining with this paradigm over several months and after a minimum of 1,390 choice-trials demonstrated great proficiency in this maze. Studies were then conducted to establish cholinergic and hippocampal involvement in the DMTS task by using drug manipulations (scopolamine and physostigmine), and after lesions to the fimbria-fornix (FF) pathway. A high dose (1.0 mg/kg) of scopolamine but not a low dose (0.3 mg/kg) significantly interfered with choice accuracy as measured by errors and trials to criterion; physostigmine (0.01 and 0.03 mg/kg) had no significant effect; and fimbria lesions significantly disrupted both choice accuracy and runtime performance. Disruption was most pronounced on difficult problems (different paths to the goal). After lesions only, considerable within-trial perseverative errors occurred during the early postlesion weeks on four difficult problems from among the 18 tested. Results were discussed in terms of (a) specificity of this disruption, (b) indications of proactive interference effects, and (c) the movement-related excitation component of maze learning. The present results accord with earlier findings of disruption by scopolamine and FF lesions in a 14-unit T-maze, both mazes having similar performance requirements of shock avoidance and multiple 90-degree turns along the paths to the goal. The present results affirm that this new detour maze provides a viable approach for assessing cognitive performance in a within-subject design and thereby offers new possibilities for testing various aspects of cognitive processing, particularly for aged rodent models, in a complex aversive situation.