The functional architecture of memory representations in the parahippocampal-hippocampal system

Abstract

Episodic memories are inherent to our lives. We naturally remember past experiences and become unsettled when their recollection falls short. Recollection is a fascinating and complex process. A memory is not a single piece of information. A memory reflects our rich experiences and comprises many aspects of information, some item-related (e.g. objects and their features) and others contextual (e.g. the scenery). Our brain must be configured to represent and process all these aspects so that we can later access and recollect the entire experienced episode. In the brain’s medial temporal lobe lies a key assembly of regions for episodic memory, the parahippocampal-hippocampal system. The aim of this dissertation is to provide evidence and discuss characteristics of this system’s functional architecture that serve episodic memory. First, I focus on the representation and processing of experienced episodes in the parahippocampal-hippocampal system. Item and context information reach the system from largely segregated cortical processing streams. To what extent the information continues to be communicated in a segregated manner is unclear. With ultra-high field functional imaging, I provide novel empirical evidence, notably on a subregional level in humans, for two functional routes throughout the system. One route specifically processes scene information, in functionally connected parahippocampal, posterior-medial entorhinal cortices, and the distal subiculum. Another route, that connects perirhinal Area 35 and the retrosplenial cortex to anterior entorhinal subregions and the subiculum/CA1 border, shows no selectivity between scene and object processing. Additionally, I review evidence across species and conclude that the perirhinal cortex processes and integrates item-related features, irrespective of their nature, into unitized multidimensional item representations. Together, these insights suggest topographically specific routes through the human parahippocampal-hippocampal system, characterized by organized item-context convergence and unique context processing, respectively. Subsequently, I examine which part of the system is particularly involved when we recollect episodes. Computational and animal models suggest hippocampal subfield CA3 plays a crucial role in completing a cue towards a whole memory representation. With ultra-high field functional imaging, I provide the first empirical evidence in humans for the involvement of subregion CA3 in the cortical reactivation of the information that makes up an episode. This insight bridges the gap between model-based observations and human brain function. It shows the specific functionality of subregion CA3 in accessing memory representations and reinstating them in the cortex for holistic recollection. Finally, I discuss what happens to our memories when disease distorts their functional architecture. I provide a novel conceptual link between the information-specific architecture of the parahippocampal-hippocampal system, memorability, and altered memories with progressing Alzheimer’s pathology. I propose that memory representations reflect the pathological distortion of the system along information-specific routes. Certain aspects may thus withstand decline in early pathology stages, causing a profile of fragmented representations with potentially diagnostic value. My thesis advances insight into the functional architecture of memory representations in the human parahippocampal-hippocampal system at a rare level of subregional detail. I leverage ultra-high field functional imaging, translate long-held hypotheses from computational and rodent research to the human brain and incorporate insights across clinical and basic cognitive neuroscience. The findings sketch a specific representational architecture with subregional dynamics set up to keep information together that belongs together. This organizational scaffold has implications for the nature of memories and their recollection. My work contributes to insights on how cognitive functions like episodic memory emerge from the design of the human brain, and hence how we remember our past experiences.