Flexible use of allocentric and egocentric spatial memories activates differential neural networks in mice.

Goal-directed navigation can be based on world-centered (allocentric) or body-centered (egocentric) representations of the environment, mediated by a wide network of interconnected brain regions, including hippocampus, striatum and prefrontal cortex. The relative contribution of these regions to navigation from novel or familiar routes, that demand a different degree of flexibility in the use of the stored spatial representations, has not been completely explored. To address this issue, we trained mice to find a reward relying on allocentric or egocentric information, in a modified version of the cross-maze task. Then we used Zif268 expression to map brain activation when well-trained mice were required to find the goal from a novel or familiar location. Successful navigation was correlated with the activation of CA1, posterior-dorsomedial striatum, nucleus accumbens core and infralimbic cortex when allocentric-trained mice needed to use a novel route. Allocentric navigation from a familiar route activated dorsomedial striatum, nucleus accumbens, prelimbic and infralimbic cortex. None of the structures analyzed was significantly activated in egocentric-trained mice, irrespective of the starting position. These data suggest that a flexible use of stored allocentric information, that allows goal finding even from a location never explored during training, induces a shift from fronto-striatal to hippocampal circuits.

Learning induced epigenetic modifications in the ventral striatum are necessary for long-term memory.

Epigenetic modifications such as histone acetylation in cortical or allocortical regions have been shown to be necessary for the formation of long-term memories. Here we investigated whether similar changes were occurring also in the ventral striatum and whether they are necessary for the consolidation of aversive memory. To this purpose we performed immediate post-training focal administrations of the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA, 5, 10 or 15 µg/side) or the DNA methyltransferase (DNMT) inhibitor, 5-aza-2'-deoxycytidine (5-AZA, 0.0625 or 0.125 µg/side) in the ventral striatum of mice trained in one-trial inhibitory avoidance task. Intra-ventral striatal SAHA administrations, immediately after training, improved memory retention. Opposite effects were found with 5-AZA. We also found that training in the one-trial inhibitory avoidance is accompanied by increased acetylation of specific residues that can be further increased by intra-VS SAHA administrations. Intra-VS 5-AZA administrations on the other hand reduced training-induced histones acetylation at the same residues. These findings imply the occurrence of histone acetylation in the ventral striatum in order to store aversive memory. Moreover, they suggest that the effects induced by the DNMT inhibitor 5-AZA may at least partially due to blockade of H3 and H4 acetylation. These results suggest that the contemporary activation of similar molecular mechanisms might be needed in different brain regions to enable the formation of long-term memories.