Human galanin (0.32 nmol per rat, i.c.v.), an endogenous neuropeptide, administered 30 min before acquisition or retention trials, significantly impaired the acquisition of learning and recall of memory in a step-through type passive avoidance performance. 2. The role of dynorphin A (1-13) in learning and memory is controversial. Dynorphin A (1-13) (0.5 nmol per rat, i.c.v.) administered 5 min before galanin injection, completely antagonized these impairments. 3. Galanin significantly decreased acetylcholine release in the hippocampus 40 to 120 min after injection as determined by in vivo brain microdialysis. This peptide also decreased acetylcholine release, albeit to a lesser extent, from the frontal cortex. 4. Dynorphin A (1-13) (0.5 nmol per rat, i.c.v.) 5 min before galanin injection, completely blocked the decrease in extracellular acetylcholine concentration induced by galanin. 5. These antagonistic effects of dynorphin A (1-13) were abolished by treatment with norbinaltorphimine (5.44 nmol per rat, i.c.v.), a selective kappa-opioid receptor antagonist, 5 min before dynorphin A (1-13). 6. Dynorphin A (1-13) (0.5 nmol) itself had no effect on learning and memory and on the acetylcholine concentration in the hippocampus or the frontal cortex in normal rats. 7. These results suggest that the neuropeptide dynorphin A (1-13) ameliorates the galanin-induced impairment of learning and memory accompanied by abolition of reductions in acetylcholine release via kappa-opioid receptors.