Final Report Summary - ENSEMBLE (Neural mechanisms for memory retrieval)
The aim of ENSEMBLE was to explore how a cognitive function like memory works. Memory is one of the most extraordinary phenomena in biology. The mammalian brain stores billions of bits of information and can retrieve stored information incredibly fast without mixing the billions of items. How is this possible? In ENSEMBLE, we provide new understanding of the key mechanisms of memory formation in neural networks of the hippocampus and associated regions. We identify mechanisms explaining the dynamics of memory at multiple time scales, and found evidence for functional contributions of neural networks in multiple interlinked cortical and subcortical brain regions.
First, we showed that the brain´s 8 Hz theta rhythm enables the hippocampus to separate different spatial maps stored in the memory. Differential maps are expressed as short memory chunks in multiples of 125ms which compete until map selection settles. We also probed the capacity of the spatial memory system, recording from ensembles of CA3 cells in rats exposed to very similar spatial environments. The capacity was large; eleven rooms were encoded after only one exposure to each and kept for the future, with no overlap between maps beyond chance level. Inactivation of the medial entorhinal cortex (MEC) caused remapping in the hippocampus, pointing to a role of the MEC in remapping.
We then asked how events like odours can be associated to certain places. We showed that the lateral entorhinal cortex (LEC) and the hippocampal CA1 synchronize with the brain´s low gamma frequency rhythm (20-40Hz). This facilitates communication between the LEC and the hippocampus. Odour maps, selective for different odours develop during acquisition of the task in both structures. If the animal makes the wrong decision, the maps are not evident, suggesting they are important for the memory of the odour-place association.
We also asked whether we can understand the controlling mechanism for map transitions in the CA1 when animals make choices. We found that nucleus reuniens plays a key role in relaying information between the prefrontal cortex and the hippocampus. If the reuniens activity was inhibited, no map transitions happened in the hippocampus. We then asked how these three areas are coordinated. Supramammilare nucleus in the hypothalamus were found to synchronize the theta activity in the prefrontal cortex and the reuniens to hippocampal theta. In this way, information transfer is facilitated at the point the animal decides which future path to select, and since it is an alternation task, the memory of the last path has to be kept in mind until the choice is made.
We then showed that there is a difference in the amount of remapping expressed, or the ability for the cell groups to keep maps separate, along the CA3-CA2 axis. CA2 expressed the least remapping, suggesting that separated maps must be communicated to the rest of the brain from the CA3.
Finally, we asked whether the spatial memory system is mature when the animals start exploring around age 15 days. Earlier my lab showed that grid cells do not express a perfect hexagonal pattern at this age. We showed that border cells and head directions cells are mature very early on and found that these cells are sufficient for supporting path integration; the youngest animals had CA1 cells that fired at positions controlled by self-motion cues and not only external landmarks. To conclude, we addressed whether the spatial system needs certain experiences to become adult-like later. Recording data from the medial entorhinal cortex from rats raised in opaque spheres indicated that the animals need to be exposed to straight to anchor efficiently a mature grid map to the environment.
In conclusion, ENSEMBLE has been successful, contributing important understanding of the spatial memory system in mammals. Most projects are published in highly-ranked journals like Nature, PNAS and Neuron, with two more papers still to come.
First, we showed that the brain´s 8 Hz theta rhythm enables the hippocampus to separate different spatial maps stored in the memory. Differential maps are expressed as short memory chunks in multiples of 125ms which compete until map selection settles. We also probed the capacity of the spatial memory system, recording from ensembles of CA3 cells in rats exposed to very similar spatial environments. The capacity was large; eleven rooms were encoded after only one exposure to each and kept for the future, with no overlap between maps beyond chance level. Inactivation of the medial entorhinal cortex (MEC) caused remapping in the hippocampus, pointing to a role of the MEC in remapping.
We then asked how events like odours can be associated to certain places. We showed that the lateral entorhinal cortex (LEC) and the hippocampal CA1 synchronize with the brain´s low gamma frequency rhythm (20-40Hz). This facilitates communication between the LEC and the hippocampus. Odour maps, selective for different odours develop during acquisition of the task in both structures. If the animal makes the wrong decision, the maps are not evident, suggesting they are important for the memory of the odour-place association.
We also asked whether we can understand the controlling mechanism for map transitions in the CA1 when animals make choices. We found that nucleus reuniens plays a key role in relaying information between the prefrontal cortex and the hippocampus. If the reuniens activity was inhibited, no map transitions happened in the hippocampus. We then asked how these three areas are coordinated. Supramammilare nucleus in the hypothalamus were found to synchronize the theta activity in the prefrontal cortex and the reuniens to hippocampal theta. In this way, information transfer is facilitated at the point the animal decides which future path to select, and since it is an alternation task, the memory of the last path has to be kept in mind until the choice is made.
We then showed that there is a difference in the amount of remapping expressed, or the ability for the cell groups to keep maps separate, along the CA3-CA2 axis. CA2 expressed the least remapping, suggesting that separated maps must be communicated to the rest of the brain from the CA3.
Finally, we asked whether the spatial memory system is mature when the animals start exploring around age 15 days. Earlier my lab showed that grid cells do not express a perfect hexagonal pattern at this age. We showed that border cells and head directions cells are mature very early on and found that these cells are sufficient for supporting path integration; the youngest animals had CA1 cells that fired at positions controlled by self-motion cues and not only external landmarks. To conclude, we addressed whether the spatial system needs certain experiences to become adult-like later. Recording data from the medial entorhinal cortex from rats raised in opaque spheres indicated that the animals need to be exposed to straight to anchor efficiently a mature grid map to the environment.
In conclusion, ENSEMBLE has been successful, contributing important understanding of the spatial memory system in mammals. Most projects are published in highly-ranked journals like Nature, PNAS and Neuron, with two more papers still to come.