March 21, 2025

How scientists uncovered memory's hidden architecture New structural hallmarks of memory storage discovered by Scripps Research could lead to treatments for memory loss.

March 21, 2025

LA JOLLA, CA Using cutting-edge genetic tools, 3D electron microscopy, and artificial intelligence, Scripps Research scientists and colleagues have uncovered key structural hallmarks of long-term memory, called an engram. Their study, published in Science on March 20, 2025, paves the way for future advances in treating memory loss and other cognitive impairments associated with aging and neurodegenerative diseases.

Our work leverages recent technological developments across multiple fields, says Marco Uytiepo, a Scripps Research graduate student and the study's lead author. We used high-resolution 3D imaging to reveal the intricate architecture of brain circuits that store memory traces with unprecedented detail. Since analyzing these images with conventional computer programs could take years, we relied heavily on AI algorithms to accelerate data processing by several orders of magnitude.

Uytiepo and his colleagues focused on the hippocampus, a brain region crucial for learning and memory in both animals and humans. They used mouse models to label and identify neurons activated during a specific learning task and, at nanometer-scale resolution, then reconstructed their synaptic connections (the junctions where neurons communicate).

We hoped to uncover something interesting since no similar approaches had been implemented before, says Anton Maximov, professor of neuroscience and the study's senior author. What we did not expect was that our findings would challenge two long-standing dogmas.

At neuronal synapses, chemical signals are typically transmitted from a single nerve terminal a swollen region of an axon filled with vesicles that secrete these signals to a single postsynaptic site on the dendrite of a receiving cell. Many previous studies (using lower-resolution optical imaging methods) have suggested that learning requires a bulk increase in synapse number. However, Maximov's team found that this is not always the case the total number and arrangement of isolated synapses remained unchanged after memory formation. Instead, neurons allocated to an engram expanded their connectivity through multi-synaptic boutons (MSBs) specialized axonal terminals that simultaneously signal to up to six different dendrites rather than just one. These MSBs were not only more abundant along the axons of activated neurons but also structurally more complex.

Secondly, Maximov's team discovered that engram neurons in adjacent hippocampal regions do not preferentially connect with each other, counter to what is widely believed in the field. Instead, the expansion of their network through MSBs resulted in the recruitment of other neurons that were not engaged during learning. Moreover, the researchers found that engram neurons exhibited fine-scale alterations in the architecture of their individual synapses, including changes in intracellular organelles such as mitochondria and smooth endoplasmic reticulum. Additionally, these neurons displayed enhanced interactions with astrocytes glial cells that regulate synaptic function and provide metabolic support.

Researchers now aim to determine whether similar mechanisms operate in other brain circuits and whether their dysfunction contributes to memory loss. Furthermore, MSBs have emerged as promising therapeutic targets.

We are excited about the possibility of targeting MSBs with drugs to develop new and effective treatments for memory disorders, says Maximov. However, achieving this goal will require designing new tools to dissect the molecular composition of MSBs, which remains entirely unexplored. We are already making progress in this direction, but much work still lies ahead.

As part of this effort, the researchers are also continuing to refine their AI pipelines to improve the efficiency and accuracy of analyzing large-scale imaging data.

This study was conducted in collaboration with the National Center for Microscopy and Imaging Research (NCMIR) at UC San Diego, directed by Distinguished Professor of Neurosciences Mark H. Ellisman. As an NIH BRAIN Initiative National Resource for Technology Integration and Dissemination, NCMIR provides cutting-edge imaging tools that advance neuroscience research.

We feel incredibly fortunate to have joined forces with Mark and his team, says Maximov. Their deep knowledge, technical expertise, and access to state-of-the-art microscopes were instrumental to our success.

This work was supported by funding from the National Institute of Mental Health, the National Institute of Neurological Disorders and Stroke, and The Brain Research Through Advancing Innovative Neurotechnologies Initiative, or The BRAIN Initiative .

In addition to Uytiepo, Ellisman and Maximov, authors of the study, Synaptic architecture of a memory engram in the mouse hippocampus, include Yongchuan Zhu, Katherine Chou, Filip Souza Polli, Elise Zhao, Danielle Luu, Lyanne Chang, Dong Yang, Tsz Ching Ma, Mingi Kim, Yuting Zhang, Grant Walton, Tom Quach, Luca Patapoutian, Arya Shahbazi, Yuxuan Zhang, Elizabeth Beutter, Weiheng Zhang, Brian Dong, Aureliano Khoury, Alton Gu, Elle McCue and Lisa Stowers of Scripps; and Eric Bushong, Keun-Young Kim and Matthias Haberl of UCSD.

Neuroscience Maximov, Anton

LINK:	https://www.scripps.edu/news-and-events/press-room/2025/20250321-maxim...
	See more stories from scripps

Most recent headlines