Recent studies in neuroscience have deepened our understanding of the processes through which memories, particularly those formed in close temporal proximity, become linked together. This process, known as memory linking, is now recognized as a well-established phenomenon in the brain, but the intricate neural mechanisms driving this phenomenon are not yet fully understood. A team of researchers at the University of California, Los Angeles (UCLA), sought to address this gap by investigating the neural basis of memory linking in the mouse brain. Their research, published in Nature Neuroscience, provides new insights into the role of dendritic plasticity — the ability of dendrites (the tree-like extensions of neurons) to adapt over time — in the process of memory linking.
A few years prior, a groundbreaking study by Alcino Silva and his colleagues had demonstrated that memories formed within a few hours of each other are often linked because they are stored in overlapping populations of neurons in the hippocampus. This discovery raised important questions about how these memories are stored within individual neurons and what processes trigger the recruitment of these neurons. In their latest research, Silva and his team aimed to answer these questions, employing advanced techniques to explore the subcellular mechanisms that underpin memory linking.
The challenge of directly observing memory storage within neurons had long hindered progress in this area, but advances in technology allowed for a more detailed investigation. The team collaborated with Yiota Poirazi’s laboratory at the Foundation for Research and Technology in Crete to explore how dendritic and synaptic dynamics could play a role in the linking of memories. Poirazi’s earlier theoretical work in 2016 had predicted that linked memories should not only be stored in common neurons but also within specific dendrites of those neurons. This prediction laid the foundation for their latest research, which sought to validate and expand on this hypothesis.
To carry out their study, the researchers employed a multi-faceted approach, using three complementary imaging techniques to visualize the subcellular compartments of neurons in living mice. These compartments included the soma (the cell body), dendrites, and spines (tiny protrusions on dendrites where synapses form). The results from these experiments were striking. The researchers found that when mice formed two memories in close succession, many of the same somas, dendritic branches, and spines were involved in encoding these memories. This observation strongly suggested that specific dendritic structures play a crucial role in linking memories across time.
In a second set of experiments, the team used sophisticated genetic tagging techniques to manipulate the neuronal somas and dendrites of the mice. By forcing independent memories to be stored within the same somas or dendrites, the researchers were able to link these memories in the rodents’ brains. Notably, this intervention took place in the retrosplenial cortex, a brain region that is known to be involved in spatial memory and navigation. The results were dramatic: the manipulated mice developed an aversion to a previously neutral box simply because the memory of this box was now stored in the same dendrites that held memories of an earlier electric shock. This experiment provided compelling evidence for the hypothesis that dendritic plasticity plays a vital role in linking memories.
The researchers also turned to computational modeling to help explain their findings. By simulating a bio-realistic network of neurons with dendrites and localized plasticity, the model showed that dendritic plasticity mechanisms are essential for replicating key properties of linked memories. These properties include the clustering of synapses, the recruitment of the same dendrites for multiple memories, and the stability of the linked memories over time. The model demonstrated that localized changes in dendrites — occurring within a few micrometers of each other — are necessary for memory linking. This finding is significant, as it suggests that small, localized plasticity changes at the level of individual dendrites are the critical neural processes driving the linking of memories.
The study also offered new insights into the role of dendritic plasticity in memory processes beyond memory linking. Though localized dendritic changes had been observed in earlier studies in cell cultures and brain slices, this study is one of the first to demonstrate their functional relevance in the context of animal behavior. As a result, the findings suggest that dendritic plasticity could be involved in a wide range of memory-related processes, from the encoding of new memories to the formation of complex memory episodes that span different times and contexts.
One of the most exciting implications of this research is its potential to inform our understanding of memory disorders, particularly those that involve impairments in memory linking. In conditions like Alzheimer’s disease, patients often experience difficulty linking past memories or recalling sequences of events, which can severely disrupt their sense of continuity. By revealing the specific mechanisms behind memory linking, this research could pave the way for the development of new treatments aimed at restoring or enhancing memory processes in individuals with neurodegenerative diseases. Understanding how memories are linked across time could also help researchers design interventions for other conditions where memory processes are disrupted, such as post-traumatic stress disorder (PTSD) and schizophrenia.
In future studies, the researchers plan to further investigate the mechanisms of dendritic plasticity that contribute to the encoding and linking of memories. Megha Sehgal, the first author of the study, is particularly interested in understanding the circuit and molecular processes that enable this plasticity. Her lab at The Ohio State University is already working to explore these aspects in greater detail. The team’s work has also opened up new avenues for investigating how dendritic plasticity can be manipulated in other brain regions to impact different types of cognitive processes, such as learning and decision-making.
In parallel, Poirazi and her collaborators in Crete are extending their computational models to simulate other brain areas and their contributions to memory processes. Their aim is to refine these models further to account for the complex interactions between dendritic mechanisms and other cognitive tasks. These models may ultimately shed light on how different brain regions work together to support memory and learning, and could even lead to new insights into how the brain adapts to new information over time.
An exciting development in this area is the potential application of these dendritic mechanisms to artificial intelligence (AI) systems. Given that dendrites play a critical role in biological learning and memory, Poirazi’s lab has initiated research to incorporate these principles into artificial neural networks. Their goal is to make machine learning systems more robust, intelligent, and efficient, mimicking the adaptive processes seen in the human brain. This line of research could lead to more advanced AI systems capable of learning in ways that are similar to human cognition, offering promising implications for fields like robotics, natural language processing, and autonomous systems.
The findings from this study contribute significantly to our understanding of the neural mechanisms underlying memory linking and memory encoding. By uncovering the critical role of dendritic plasticity, the research opens up new possibilities for addressing memory-related disorders and improving our understanding of the brain’s ability to adapt to new experiences. As the field continues to evolve, the insights gained from these studies will likely form the foundation for new treatments and technologies aimed at enhancing both human and artificial cognitive capabilities.
More information: Megha Sehgal et al, Compartmentalized dendritic plasticity in the mouse retrosplenial cortex links contextual memories formed close in time, Nature Neuroscience (2025). DOI: 10.1038/s41593-025-01876-8