A remarkable feature of the human mind is the ability to retain memories at different levels of abstraction. We not only have detailed, vivid memories of specific experiences, but also store generalized memories that are based on distinct experiences. This symposium will track the emergence and transformation of such diverse memory representations in the brain. Deniz Kumral will first demonstrate how the formation of memories is reflected in alterations of the brain's microstructure. Anne Bierbrauer will then show how consolidation leads to an emphasis of an episode’s gist at the cost of specific details. Afterwards, Heidrun Schultz will elucidate how reward influences the emergence of generalized memory representations, and Mona Garvert will examine how they are influenced by conditioned fear. Finally, Roland Benoit will provide causal evidence for the neural basis of such memory representations. Together, the five presentations will provide a current perspective on the nature of neural memory representations and how they evolve over time.

Neuroplasticity, the brain’s capacity for functional or structural changes following a learning experience, allows humans to learn and form new memories. With new neuroscientific methods, such as diffusion-weighted MRI (DW-MRI), we can now characterize brain microstructural changes that signal neuroplasticity by assessing motion profiles of water molecules. Recent studies have shown that during repeated encoding and retrieval of an object–location learning task, a physical memory trace is formed rapidly in the parietal cortex, which can be observed already 90 minutes after training. The exact temporal dynamics of microstructural changes following a learning experience are, however, unclear. Here, we investigated functional and structural changes in the brain over the course of learning. Seventy-six participants completed an object-location learning task with repeated encoding and retrieval of image pairs and their locations, while a no-learning control condition (N=38) was tested at corresponding times. To assess structural changes, DW-MRI was acquired at 22 time points across three hours: one hour before the learning, during learning, and a one-hour-long wake rest phase after learning. We observe a decrease in mean diffusivity (MD), reflecting neural plasticity, directly following the learning task in the precuneus, cerebellum, and lateral occipital cortex, while we observe no structural changes in the control group. Further, these MD decreases further develop during the rest phase. Notable, within-person associations reveal that microstructural changes during learning are linked to memory performance. We confirm the rapid build-up of long-term memory representation, highlighting the dynamic nature of neuroplasticity and its crucial role in memory formation.

Systems consolidation suggests a transformation of memory, where episodic details decrease, and gist information increases over time. Previous studies indicate this process involves a shift from the posterior to the anterior hippocampus, as well as from the hippocampus to cortical regions. However, most studies have focused on pictures or movies as proxies for real-life episodic memories. We present a 3-year longitudinal study on memory consolidation of real-life episodic memories. Ninety-two participants experienced twelve distinct, standardized real-life episodes. A wearable eye-tracker captured the ground truth of their encoded experiences, allowing us to track the transformation of these memories during consolidation. Participants underwent fMRI scanning the day after, 8 months, and 3 years later, while recalling the episodes, enabling us to quantify the involvement of brain regions during retrieval and how episode representations changed. An autobiographical memory interview (AMI) was conducted to assess the accuracy and detail richness of the memories. We also measured spontaneous memory reactivations, with preliminary results indicating a link between reactivations and the shift from episodic to gist-based information. fMRI data showed a shift in activation from the posterior to anterior hippocampus, as well as from hippocampus to cortical areas during long-term consolidation. Finally, we examined how stress and emotions influence episodic memory transformation by inducing stress in some episodes. Elevated cortisol, heart rate, and more negative emotional ratings confirmed successful stress manipulation. Preliminary results suggest a stress-driven centrality effect in fixation patterns, consistent with prior findings.

An integral part of episodic retrieval is the reinstatement of neural activity that was present in the medial temporal lobe during encoding. However, neural memory representations do not remain static. Consolidation promotes the transformation of representations that are specific to individual episodes toward more generalized representations that reflect commonalities across episodes. Moreover, reward has been shown to augment episodic memory by enhancing consolidation, and it may accelerate the transformation of neural memory representations. We investigated this account with n=40 human participants using fMRI and an associative memory task. They encoded pictures of objects, each with one of four recurring scenes. Two scenes led to high reward, and two led to low reward. The next day, participants encountered the objects again and retrieved the scenes from memory. Using representational similarity analysis, we demonstrate that retrieval is concurrently accompanied by the reinstatement of original neural representations and the activation of transformed, more generalized memories. Specifically, the parahippocampal cortex reinstates scene-specific patterns from the encoding phase during successful retrieval. In contrast, activity patterns in the medial prefrontal cortex and anterior hippocampus reflect transformed memories: They become more similar to each other for memories sharing the same scene, independent of memory success. Importantly, high reward enhances memory transformation in the anterior hippocampus. The brain thus maintains complementary memory representations: An episodic representation that resembles the original encoding pattern, and a generalized representation that summarizes commonalities across memories - in part for particularly valuable information. Published as Schultz et al. (2025), Imaging Neuroscience, doi: 10.1162/imag_a_00476.

Avoiding danger is vital for survival, yet directly learning from dangerous situations often entails significant risks to our health and well-being. To mitigate these risks, the brain uses models of the world to generalise learned information to similar situations. Such generalisation is essential for adaptive behaviour, enabling efficient decision-making based on limited experiences without the need for constant relearning. However, not all generalisations are beneficial or accurate. Exaggerated generalisation of aversive information, in particular, can result in excessive avoidance of anticipated dangers, profoundly impairing daily life. In this study, we integrated a cognitive map formation task with a fear-conditioning paradigm to examine how aversive information spreads across cognitive maps. Pupil dilation revealed that anxious-depressive traits are associated with the formation of less precise cognitive maps and distinct map-based generalisation patterns. Heightened attention to pain, on the other hand, selectively influences learning about safety cues. These results shed light on the mechanisms underlying the formation of memory representations used for generalisation and their disruption in anxiety-related disorders.

The medial prefrontal cortex (mPFC) is thought to support the recollection of past episodes and the simulation of future episodes - specifically by mediating the reinstatement of memory schemas. Memory schemas provide general knowledge about what to expect in a given scenario (e.g., the sequence of events when preparing a meal). Such general knowledge may scaffold the (re)construction of specific episodes. One methodological challenge in elucidating the contribution of the mPFC to schema reinstatement is the difficulty in assessing the schematic content of recollected and simulated episodes. Here, we first used natural language processing on a large data set to infer the schematic content of several everyday scenarios (e.g., preparing a new meal). Specifically, using topic-modelling, we determined the key reoccurring content across a large number of reported memories and simulations. We then examined whether people with mPFC lesions use less of that schematic content than matched controls. Traditional measures of autobiographical memory revealed that lesions to the mPFC caused impoverished recollections and simulations. This was– at least partly – due to a reduced reliance on schematic knowledge that otherwise seems to help scaffolding retrieval and simulation. Notably, as expected, the deficiency in using schematic content was particularly pronounced for future simulations. This is consistent with a stronger scaffolding demand for the simulation of novel episodes. Together, the data provide evidence for the critical contribution of the mPFC to the reinstatement of memory schemas.