[Reading Science] The Reason We Forget: It's in the 'Engram Connections'

IBS Identifies Synaptic Strengthening Between Engram Cells
as a Key Condition for Memory Recall

Although memories are certainly stored in the brain, why do they sometimes disappear when we try to recall them? A new study has found that it is not the engram cells themselves-specific neurons responsible for memory-but rather the connections linking these cells that determine whether a memory can be retrieved.

The research team led by Bongkyun Kang, Director of the Center for Memory and Glial Cells at the Institute for Basic Science (IBS), experimentally demonstrated that not only the engram cells that store memories, but also the number and structural strengthening of synapses formed between these cells, play a decisive role in actual memory recall.

Summary schematic diagram of the core research content. After learning, the number of synapses and structural size between engram cells increase. When protein synthesis is inhibited once, the strengthening of synapse structure is reduced, but the increase in number is partially maintained, resulting in weakened natural memory retrieval but possible artificial retrieval. On the other hand, when inhibition is repeated, both the increase in synapse number and structural strengthening are blocked, and neither natural nor artificial memory retrieval occurs. Provided by the research team

Engrams refer to the traces of specific experiences left in the brain, and in neuroscience, the group of neurons responsible for memory is called "engram cells." While previous memory research has mainly focused on identifying which cells store memories, the research team emphasized that "even if engram cells exist, if a sufficiently connected circuit is not formed between them, the memory may not manifest as behavior."

Synapses Linking Engram Cells Revive Memories

The researchers analyzed the ventral hippocampus (vCA1)-basal amygdala (BA) brain circuit, which is closely associated with fear memory. After conducting fear conditioning experiments in which mice associated a specific environment with an unpleasant stimulus, they observed synaptic changes in the brain circuit before and after learning at high resolution. During this process, they quantitatively analyzed synapses derived from engram cells and those that were not.

The results showed that synaptic changes after fear learning did not occur randomly. Only the synapses directly connecting engram cells exhibited a selective increase in synapse number, and the size of the spine structures forming these synapses also increased. In contrast, such changes were rarely observed in synapses between engram and non-engram cells.

The research team then conducted experiments inhibiting protein synthesis immediately after learning to determine how these synaptic changes affect actual memory recall. When protein synthesis was inhibited once, most structural strengthening of the synapses was blocked, but the increase in synapse number was largely maintained. Under these conditions, the mice exhibited reduced natural recall of fear memory, but when the engram cells were artificially stimulated to forcibly reactivate the memory, recall responses were maintained.

Research team photo. From the left, Bongkyun Kang, Director of the IBS Center for Memory and Glial Cells (corresponding author), Ilgang Hong, Postdoctoral Researcher at the Center for Memory and Glial Cells (first author), Yeonjun Kim, Postdoctoral Researcher at the Center for Memory and Glial Cells (first author). Provided by IBS

On the other hand, when protein synthesis was repeatedly inhibited for an extended period, blocking both the increase in synapse number and structural changes, not only natural memory recall but also artificially induced recall responses were significantly reduced. This result demonstrates that even if engram cells storing the memory are formed, if sufficient synaptic connections linking them are not established, the memory may not be recalled at the behavioral level.

Bongkyun Kang stated, "While it is known that engram cells store memories, the conditions required for those memories to actually be recalled were not clear. This study shows that memory recall can be determined by the state of synaptic connections between cells, providing important clues for understanding memory retrieval failure and memory disorders."

The results of this study were published in the Proceedings of the National Academy of Sciences (PNAS) on January 12, 2026. The paper is titled "Protein Synthesis Blockade Prevents Fear Memory Reactivation via Inhibition of Engram Synapse Strengthening."