AIMS: Fluoxetine, a selective serotonine reuptake inhibitor, is commonly used to treat psychiatric disorders. Available data show that fluoxetine has limited side effects and, more importantly, may improve patient’s cognitive abilities. However, little is known about the mechanisms by which fluoxetine affects learning, especially appetitively motivated one. Thus, in the present project we investigated the effects of a long-term fluoxetine treatment on appetitively motivated discrimination learning. METHODS: We used fully automated behavioral assessment of discrimination learning in group-housed subjects, DI-staining for determining changes in morphology of dendritic spines and gel zymography for measurement of activity of MMP-9 (matrix metaloproteinase 9, an enzyme involved in synaptic plasticity). RESULTS: We showed that above-described learning is severely impaired in mice subjected to the long-term fluoxetine treatment. Since we have previously shown that such learning depends on MMP-9 activity in the central amygdala (CeA), we examined MMP-9 activity in the CeA of the fluoxetine treated mice. We found decreased MMP-9 level. Further, we tested fluoxetine influence on dendritic spine morphology in the CeA and observed that behavioral performance of the control wild type mice was highly correlated with a size and of mature, mushroom-shaped dendritic spines. No such correlation was found in MMP-9 knock out mice. Applied treatment abolished this correlation in wild type mice and did not reinstated it to a significant level in MMP-9 knock outs. CONCLUSIONS: Obtained results show that chronic fluoxetine treatment impairs appetitive discrimination learning in healthy controls, decreases MMP-9 activity and disrupts correlation between subjects’ performance in appetitive learning and structural synaptic plasticity in the CeA. The data shed light on dendritic spines’ dependent learning mechanisms, that may be disarrayed in the CeA by commonly applied fluoxetine treatment in patients.