Study shows that dendritic plasticity contributes to the integration of memories

Past neuroscience studies suggest that memories of events that occurred at short time intervals from one another are often connected, via a process referred to as memory linking. While memory linking is now a well-documented phenomenon, its neural underpinnings have not been fully elucidated.

Researchers at the University of California Los Angeles (UCLA) recently carried out a study aimed at better understanding the neural processes that contribute to memory linking in the mouse brain. Their findings, published in Nature Neuroscience, suggest that dendritic plasticity, the adaptation of dendrites (i.e., branch-like extensions of neurons) over time, plays a key role in the linking of memories.

“A few years back, in a landmark study published in Nature in 2016, we demonstrated that memories formed a few hours apart are linked because they are stored in a common set of neurons in the hippocampus,” Alcino Silva, senior author of the paper, told Medical Xpress. “We wanted to know: Where within these neurons are these memories stored and linked? What was causing these neurons to be recruited?”

While answering these research questions in an experimental setting was unfeasible a few years ago, since then Silva and his colleagues have developed new tools and technologies to probe subcellular mechanisms. In collaboration with Yiota Poirazi’s laboratory at the Foundation for Research and Technology in Crete, they then set out to investigate how dendritic and synaptic dynamics could contribute to memory linking, employing modeling techniques.

“In an earlier theoretical study also published in 2016, we predicted that in addition to being stored in common neuronal populations, linked memories should also reside within common dendrites within these neurons,” said Poirazi.

“Inspired by these findings and with the help of our multidisciplinary team, we set out to reveal whether and how dendritic mechanisms may allow the linking of memories across time in rodent brains,” added Silva.

As part of their recent study, the researchers employed three different but complementary imaging techniques. Using these techniques, they visualized three distinct subcellular compartments in living mice, namely the soma, dendrites and spines from neurons.

“We showed that when mice form two memories close in time, we can see that many of the same somas, dendritic branches, and spines are involved in forming these two memories,” explained Megha Sehgal, the first author and a co-corresponding author of the paper.

“In a second class of experiments, we used sophisticated genetic tagging techniques to manipulate these neuronal somas and dendrites. When we forced independent memories to be stored in the same neuronal somas or even the dendrites and found just this simple intervention in one brain region, the retrosplenial cortex, was enough to link these memories!”

Silva, Sehgal and their colleagues found that, following their experimental intervention, mice became scared of a box that was previously unimportant to them, simply because the memory of this box was stored in the same dendrites that stored memories of a box in which they experienced an electric shock. In collaboration with Poirazi and her lab, they then used computational modeling techniques to explain their observations.

“By simulating a bio-realistic network of neurons with dendrites and localized plasticity, the model showed that localized dendritic plasticity mechanisms are necessary for replicating key properties of linked memories, such as their recruitment of the same dendrites, clustering of synapses, and stability over time,” said Sehgal.

The findings gathered by this research team suggest that the linking of memories in the mouse brain is supported by highly localized changes (i.e., within a few micrometers) on neuronal dendrites. Silva, Sehgal and their colleagues hypothesize that similar localized dendritic changes could also play a role in other types of memory processes.

“Although such localized changes have been reported in previous studies in cell cultures and brain slices, we did not know their function,” said Sehgal. “To the best of our knowledge, this is the first demonstration of their usefulness in animal behavior.”

This recent study could soon pave the way for further research exploring the contribution of dendritic plasticity to specific well-documented memory processes. In addition, it could help to better understand disorders associated with an impaired ability to link memories.

“Our findings are important for understanding how memories are linked across time to form memory episodes as well as for addressing memory deficits whereby memory linking is impaired, such as those linked to Alzheimer’s disease,” explained Poirazi.

“By providing a mechanistic understanding of memory linking, our work serves as a first step towards the development of new treatments that may target such mechanisms in order to remedy respective memory deficits.”

In their future research, Sehgal and her lab at The Ohio State University will continue exploring the underpinnings of dendritic plasticity that contributes to the linking and encoding of memories. In addition, they plan to further investigate the memory-related plasticity patterns that they observed as part of their recent study.

“We discovered that compartmentalized plasticity plays a critical role in dictating how memories are stored in the future, but we do not know the underlying mechanisms,” said Sehgal. “My lab is now digging deeper into the circuit and molecular processes that allow this plasticity.”

Poirazi and her collaborators at the Foundation for Research and Technology are now working to extend their computational models to simulate other brain areas and their contribution to different cognitive tasks. They hope that these models will help them to better understand the role of dendritic mechanisms in learning and memory functions, while also unveiling some of their most important features.

“In parallel, given the key role of dendrites in biological learning and memory, we initiated a new research line whereby we adopt dendritic mechanisms in artificial neural network systems (e.g., Chavlis and Poirazi, Nature Communications 2025), with the aim of making machine learning and artificial intelligence systems more robust, intelligent and efficient, like the brain.”

© 2025 Science X Network

Past neuroscience studies suggest that memories of events that occurred at short time intervals from one another are often connected, via a process referred to as memory linking. While memory linking is now a well-documented phenomenon, its neural underpinnings have not been fully elucidated.

Researchers at the University of California Los Angeles (UCLA) recently carried out a study aimed at better understanding the neural processes that contribute to memory linking in the mouse brain. Their findings, published in Nature Neuroscience, suggest that dendritic plasticity, the adaptation of dendrites (i.e., branch-like extensions of neurons) over time, plays a key role in the linking of memories.

“A few years back, in a landmark study published in Nature in 2016, we demonstrated that memories formed a few hours apart are linked because they are stored in a common set of neurons in the hippocampus,” Alcino Silva, senior author of the paper, told Medical Xpress. “We wanted to know: Where within these neurons are these memories stored and linked? What was causing these neurons to be recruited?”

While answering these research questions in an experimental setting was unfeasible a few years ago, since then Silva and his colleagues have developed new tools and technologies to probe subcellular mechanisms. In collaboration with Yiota Poirazi’s laboratory at the Foundation for Research and Technology in Crete, they then set out to investigate how dendritic and synaptic dynamics could contribute to memory linking, employing modeling techniques.

“In an earlier theoretical study also published in 2016, we predicted that in addition to being stored in common neuronal populations, linked memories should also reside within common dendrites within these neurons,” said Poirazi.

“Inspired by these findings and with the help of our multidisciplinary team, we set out to reveal whether and how dendritic mechanisms may allow the linking of memories across time in rodent brains,” added Silva.

As part of their recent study, the researchers employed three different but complementary imaging techniques. Using these techniques, they visualized three distinct subcellular compartments in living mice, namely the soma, dendrites and spines from neurons.

“We showed that when mice form two memories close in time, we can see that many of the same somas, dendritic branches, and spines are involved in forming these two memories,” explained Megha Sehgal, the first author and a co-corresponding author of the paper.

“In a second class of experiments, we used sophisticated genetic tagging techniques to manipulate these neuronal somas and dendrites. When we forced independent memories to be stored in the same neuronal somas or even the dendrites and found just this simple intervention in one brain region, the retrosplenial cortex, was enough to link these memories!”

Silva, Sehgal and their colleagues found that, following their experimental intervention, mice became scared of a box that was previously unimportant to them, simply because the memory of this box was stored in the same dendrites that stored memories of a box in which they experienced an electric shock. In collaboration with Poirazi and her lab, they then used computational modeling techniques to explain their observations.

“By simulating a bio-realistic network of neurons with dendrites and localized plasticity, the model showed that localized dendritic plasticity mechanisms are necessary for replicating key properties of linked memories, such as their recruitment of the same dendrites, clustering of synapses, and stability over time,” said Sehgal.

The findings gathered by this research team suggest that the linking of memories in the mouse brain is supported by highly localized changes (i.e., within a few micrometers) on neuronal dendrites. Silva, Sehgal and their colleagues hypothesize that similar localized dendritic changes could also play a role in other types of memory processes.

“Although such localized changes have been reported in previous studies in cell cultures and brain slices, we did not know their function,” said Sehgal. “To the best of our knowledge, this is the first demonstration of their usefulness in animal behavior.”

This recent study could soon pave the way for further research exploring the contribution of dendritic plasticity to specific well-documented memory processes. In addition, it could help to better understand disorders associated with an impaired ability to link memories.

“Our findings are important for understanding how memories are linked across time to form memory episodes as well as for addressing memory deficits whereby memory linking is impaired, such as those linked to Alzheimer’s disease,” explained Poirazi.

“By providing a mechanistic understanding of memory linking, our work serves as a first step towards the development of new treatments that may target such mechanisms in order to remedy respective memory deficits.”

In their future research, Sehgal and her lab at The Ohio State University will continue exploring the underpinnings of dendritic plasticity that contributes to the linking and encoding of memories. In addition, they plan to further investigate the memory-related plasticity patterns that they observed as part of their recent study.

“We discovered that compartmentalized plasticity plays a critical role in dictating how memories are stored in the future, but we do not know the underlying mechanisms,” said Sehgal. “My lab is now digging deeper into the circuit and molecular processes that allow this plasticity.”

Poirazi and her collaborators at the Foundation for Research and Technology are now working to extend their computational models to simulate other brain areas and their contribution to different cognitive tasks. They hope that these models will help them to better understand the role of dendritic mechanisms in learning and memory functions, while also unveiling some of their most important features.

“In parallel, given the key role of dendrites in biological learning and memory, we initiated a new research line whereby we adopt dendritic mechanisms in artificial neural network systems (e.g., Chavlis and Poirazi, Nature Communications 2025), with the aim of making machine learning and artificial intelligence systems more robust, intelligent and efficient, like the brain.”

© 2025 Science X Network