The subiculum: a key interface between scene representation and event memory?

When we remember events, we often bring to mind the specific location in which that event took place, often in vivid detail. These mental images are not usually single snapshots, like photographs, but something more integrated and flexible, where we can imagine different viewpoints, the spatial layout of an environment, and the relationship between objects in that environment. For this reason, understanding how different brain regions process and integrate spatial information, such as scenes, into events, is vital for understanding human memory and how it is affected in disorders of ageing. One such critical region is the subiculum, which forms part of the hippocampal formation. According to animal studies, the subiculum has a privileged position in the brain, enabling a distributed network of regions involved in memory to communicate. There is also growing evidence that this region is highly vulnerable as we age, and affected early on in disorders such as Alzheimer's disease.

Until recently it has been challenging to study the subiculum as its size and position in the brain makes it difficult to investigate using conventional imaging techniques. In this project, we will address this problem via state-of-the-art high-resolution brain scanning methods, which will be combined with novel cognitive tasks, allowing us to examine how the subiculum, via its connectivity with an extended brain network, supports scene perception and memory. This project builds on our previous work, where we found that the subiculum was critical for telling apart different scenes, but not other complex visual stimuli, such as faces.

Our first aim is to better understand the anatomical connectivity of the human subiculum, informed by work on animals. We will combine both structural and functional brain imaging to 'map out' how the subiculum is connected in the human brain, and compare this connectivity map to that seen in monkeys using an open-source dataset. We will also apply a mathematical approach, called graph theory, to see if the subiculum acts as a 'hub' in an extended memory network, as predicted from animal studies. Using this approach, we can investigate how virtually "switching off" the subiculum, relative other hippocampal sub-regions, impacts on communication between structures within the extended memory network.

Our second aim is to use high-resolution functional brain imaging to investigate the precise role of the subiculum in representing scene information. In the first task, participants will be asked to judge whether two views are from the same location or not, allowing us to ask whether the subiculum binds different views of a scene together (integration), or tells them apart (discrimination) during perception. In a second task, we will use a virtual-reality task in the scanner to see how 'integrated' scene representations in the subiculum - thought to be central to memory - are shaped by experience.

Finally, building on our second aim, we will explore how the subiculum's role in scene processing is linked to everyday event memory. To do this, participants will watch a movie in the scanner and later recall what they remember in as much detail as possible. We will test whether the subiculum, relative to other hippocampal sub-regions, responds to large shifts in the movie narrative, known as 'event boundaries'. Event boundaries are triggered by changes in the world around us, and allow us to segment our everyday experiences into discrete 'chunks' in memory. By labelling event boundaries as either spatial or non-spatial, we can see whether this brain activity is driven by spatial information (e.g., scene changes), and how this relates to event memory.

This project will generate new insights into the neuroanatomy of human memory, testing how research findings from animals generalise to humans, and providing unique knowledge about a critical brain network for scene and event memory.

The subiculum lies at the interface between the hippocampus and the rest of the brain, and thus possesses unique significance for understanding human memory and how it emerges via anatomical and functional interactions within a broader neural network supporting scene representation, event memory and navigation. The subiculum's strategic anatomical position in the brain is also thought to confer particular vulnerability to poor later life cognitive outcomes in old age, as well as increased risk of dementia. Despite this, remarkably little is known about the subiculum's functional and connectional neuroanatomy in humans.

We will address this gap in our knowledge in this proposal by applying multimodal 7T imaging to investigate how the human subiculum - via its unique anatomical connectivity within an extended brain network - contributes to scene representation and event memory.

First, we will fuse whole-brain high-resolution diffusion and functional MRI (1.2mm isotropic) to map the extrinsic connections of the subiculum and their topography, thus providing vital new knowledge about how the subiculum is organised in the living human brain.

Second, we will use ultra-high-resolution fMRI (0.8mm isotropic), focused on the hippocampus and posteromedial cortex, to test whether the subiculum integrates scenes into unified representations during perception and memory.

Finally, we will use a naturalistic movie-viewing paradigm to test whether the subiculum's putative role in scene representation underpins a broader role in event perception and memory.

The findings from this proposal have the potential to reveal new insights into the neuroarchitecture of brain circuitry supporting human event memory, and facilitate the development of novel cognitive tools for assessing brain health across the life course.