In the first half of the past century, the basal ganglia were mainly implicated in motor behavior [1]. Subsequent research challenged the view of the basal ganglia as a pure motor system and revealed that the basal ganglia are involved in a variety of cognitive functions, ranging from learning and memory to attentional processes or the building of action plans 1, 2, 3. In particular, the dorsal striatum, composed of caudate nucleus and putamen, was shown to be relevant for a form of learning, often referred to as ‘habit’ learning, in which rather rigid stimulus-response (S-R) associations are incrementally built 1, 4, 5. This putative dorsal striatal S-R ‘habit’ memory system has been dissociated from a flexible but cognitively more demanding (‘cognitive’) memory system that processes the relationship between multiple cues to build a cognitive map and depends mainly on the hippocampus 4, 6. These dorsal striatal and hippocampal memory systems operate in parallel [7] but may lead to distinct responses and compete for control of behavior 8, 9. A key question thus concerns the factors that determine which of these memory systems gets the upper hand and can thus guide behavior.

Research over the past two decades demonstrated that arousal or stress may bias the preferential engagement of striatal and hippocampal memory systems 10, 11, 12. Stressful events have been known for long to be a powerful modulator of learning and memory 13, 14. Although major stress mediators, such as glucocorticoids and noradrenaline, were initially thought to act primarily on medial temporal and prefrontal areas, more recent evidence suggests that these stress mediators impact also striatal processes 15, 16, 17. Beyond its effects on the functioning of a single memory system, stress has been shown to modulate the balance of multiple memory systems. Most of the research in this area focused on the preferential recruitment of hippocampal ‘cognitive’ memory and dorsal striatal ‘habit’ memory during learning. I will briefly summarize these findings in the first part of this review. Recent research revealed that stress may modulate, in addition to its effects on the engagement of multiple memory systems during learning, the preferential recruitment of coexisting hippocampal and dorsal striatal memory traces during remembering. I will focus on these recent discoveries in the second part of this review. Finally, I will suggest that modulatory effects of stress on the acquisition and expression of ‘cognitive’ vs. ‘habitual’ memories can be found across domains of memory, pointing to a general mechanism that may facilitate adaptation to stressful events.