Neural Correlates and Behavioral Underpinnings of Remote Memories during Systems Consolidation

Abstract

Remote retrieval of declarative memories has been under investigation for a long time. After acquisition, long-term encoding of memories happens through crosstalk across multiple brain regions. Through systems consolidation, memory retrieval has been shown to become progressively less dependent on hippocampus and more dependent on cortex. During such processes, it has been observed that memories become generalized over time and their details are not recalled in their entirety. In contrast, modifications in training and testing protocol have been used to show that under certain circumstances, memories can be recalled in detail even after long time of encoding. At the same time, it has been shown that retrieval of memories results in memory updating in terms of enhancement or distortion. Although, the neural correlates and behavioral processes involved in such phenomenon are still being investigated. We implement behavior tests, lesion experiments and in vivo imaging to understand retrieval-dependent interactions between old memories and related new information, in terms of their elemental and configural nature. First, we establish a more sensitive behavior paradigm and analyses methods for social transmission of food preference task to investigate remote retrieval of relatively simple, hippocampus dependent non-spatial associations. We find that the retention and remote retrieval of such memories depends on the feature content information. Secondly, we test the role of feature overlap in detailed retrieval of remote contextual memories using contextual fear conditioning paradigm. We find that the specificity of remote memory is dependent on the testing order. At the same time, we find evidence for remote retrieval-dependent, second order conditioning (SOC) based new learning for ABC testing order and not for BAC testing order. Thirdly we perform timed hippocampus lesions for animal sub-groups tested in ABC order, to investigate acquisition and systems consolidation of SOC based new learning. We find that hippocampus is required for SOC based new learning and such memory can be retrieved if hippocampus is lesioned after ~4 weeks of acquisition. Surprisingly, we find that SOC based new learning undergoes rapid systems consolidation as it can be retrieved even if hippocampus is lesioned just after 24 hours of acquisition. Finally, we perform in vivo imaging of dendritic spines in RSC after acquisition of SOC based new learning and find that the increase in RSC spine density correlates with acquisition and early encoding of remote retrieval-dependent SOC based new learning