Shifting between response and place strategies in maze navigation: Effects of training, cue availability and functional inactivation of striatum or hippocampus in rats

Response and place memory systems have long been considered independent, encoding information in parallel, and involving the striatum and hippocampus, respectively. Most experimental studies supporting this view used simple, repetitive tasks, with unrestrained access to spatial cues. They did not give animals an opportunity to correct a response strategy by shifting to a place one, which would demonstrate dynamic, adaptive interactions between both memory systems in the navigation correction process. In a first experiment, rats were trained in the double-H maze for different durations (1, 6, or 14 days; 4 trials/day) to acquire a repetitive task in darkness (forcing a response memory-based strategy) or normal light (placing response and place memory systems in balance), or to acquire a place memory. All rats were given a misleading shifted-start probe trial 24-hr post-training to test both their strategy and their ability to correct their navigation directly or in response to negative feedback. Additional analyses focused on the dorsal striatum and the dorsal hippocampus using c-Fos gene expression imaging and, in a second experiment, reversible muscimol inactivation. The results indicate that, depending on training protocol and duration, the striatum, which was unexpectedly the first to come into play in the dual strategy task, and the hippocampus are both required when rats have to correct their navigation after having acquired a repetitive task in a cued environment. Partly contradicting the model established by Packard and McGaugh (1996, Neurobiology of Learning and Memory, vol 65), these data point to memory systems that interact in more complex ways than considered so far. To some extent, they also challenge the notion of hippocampus-independent response memory and striatum-independent place memory systems.