Memory engrams – physical substrates of memory in the brain

Memory engrams – physical substrates of memory in the brain

What is the best present you have ever received? Thanks to the activation of a certain combination of brain cells, you will now vividly remember the best gift you once got. In my case, it is my first bike: bright yellow with an orange flag, for my fifth birthday. Even though this memory was formed over twenty years ago, I can still clearly remember the anticipation when I saw the flag peeking through the wrapping paper. And the wonderful first ride through the living room is imprinted in my brain.

The remarkable processes of memory formation, consolidation, and retrieval have intrigued great thinkers over many centuries, like Plato and John Locke. And indeed, memories are very special conceptions. They allow you to sing along to all those torturous Christmas songs, remember your all-time favourite pet, and permit a squirrel to relocate his buried nuts after hibernation. Besides, memories were a source of inspiration for numerous movies like ‘Eternal sunshine of the spotless mind’, ‘Fifty first dates’, and Leonardo DiCaprio’s ‘Inception’. While in these movies memories can be erased or implanted in the blink of an eye, neuroscientists are actually taking baby steps in unravelling the elusive physiological substrates of memories, the so-called memory engrams.

Already in 1921, the German zoologist and biologist Richard Wolfgang Semon postulated his theory on memory engrams: “The enduring though primarily latent modification in the irritable substance, produced by a stimulus”.1 In other words: memories are formed by lasting biophysical changes in brain cells. Furthermore, memories could subsequently be reactivated, leading to retrieval of the memory. In his famous book in 1949, psychologist Donald Hebb proposed a neurobiological mechanism based on plasticity of synapses, as a substrate of memory.2 He hypothesised that when the activation of one cell leads to the activation of another cell, the connection between these two cells is reinforced. In other words: “cells that fire together, wire together”. His theory was confirmed in numerous studies a few years later and it remains to date one of the most influential theories in neuroscience. However, in a brain containing over 80 billion neurons, how do you find the cells that are connected and contain a memory?

For decades, memory research focussed on damaging parts of the brain to find out their role in memory formation and retrieval. This approach is courteously illustrated by the remarkable case of epileptic patient H.M. Back in the fifties of the previous century, H.M. underwent surgery with the aim of alleviating his epileptic seizures. In both hemispheres of his brain, considerable parts of the medial temporal cortex were removed. The surgery appeared to be successful in mitigating the epileptic seizures, but H.M. was unable to form new memories. Like the main character of romantic comedy ‘Fifty first dates’, he suffered from anterograde amnesia. This important case study highlighted the crucial role of the hippocampus, part of the removed medial temporal lobe, in the formation of new memories. The next decades of research confirmed the vital importance of the hippocampus in several memory processes and indicated that memory loss in psychopathologies, like Alzheimer’s disease, can often be traced back to damage in the hippocampus. However, if you want to study memory formation or retrieval on a cellular level, a more detailed and precise analysis is required.

Recently, researchers have been able to zoom in on the networks of neurons that contain memories. By using markers of neuronal activity, researchers estimated that only about ~10% of neurons in certain brain areas are active upon memory retrieval. This indicates that only a small subset of neurons constitutes a memory engram. By using modern molecular methods, combined with for example optogenetics, researchers can nowadays find and manipulate activated, memory-harbouring neurons.3 In case of optogenetics, a technique awarded prestigious titles like Science’s ‘Breakthrough of the Decade’ 4 and Nature’s ‘Method of the Year’,3 genes for light-sensitive proteins are inserted in selected neurons. Neurons can be selected based for example on their cell-type or activation. In these cells, light-sensitive proteins will be inserted in the cell membrane and can subsequently be activated by (laser) light of a certain wavelength. Depending on whether you opted for an exciting or inhibiting protein, this will lead to activation or suppression of the cell, respectively. That way, you can determine the contribution of these cells to a certain action or behaviour, even in living animals. If you insert these light-sensitive genes into memory-harbouring cells, you can investigate how these cells contribute to memory expression or reconsolidation.

Beeld: Thomas Briggs

In 2012, a group of researchers at the Massachusetts Institute for Technology (MIT) utilized optogenetics to confirm the Engram theory that was postulated nearly a decade before.5 They exposed a mouse to a context, referred to as A, where it received mild foot shocks. By using a combination of a transgenic mouse line and viral vectors, activated cells in the hippocampus encoding context A were tagged with a light-sensitive protein. When these cells were later optogenetically reactivated in a different context, animals displayed freezing behaviour, an indicator of fear. This shows that a fearful memory can be tagged and reactivated. Following up on this phenomenon, the researchers performed another study in which they showed that reactivation of cells could lead to a false memory. Using a similar strategy as before, mice were placed in context B, which they could freely explore, while activated cells were tagged. Then, mice were placed in context A, where they received mild foot shocks and optogenetically reactivated neurons encoding context B. When the mouse was then placed back into context B, where it had never received foot shocks, it showed increased freezing behaviour. By optogenetically reactivating cells encoding the context of B, they attached a false memory to it. These experiments revealed proof for the existence of memory engrams and were therefore rightly greeted by ‘oohs and aahs’ of fellow neuroscientists.

In the years that followed, neuroscientists have made significant progress in manipulating not only fearful memories in healthy mice, but also disrupted drug-associated memories.6 Recently, it was shown that in a mouse model of Alzheimer’s disease, apparent lost long-term memories could be retrieved when optogenetically stimulated.7 While natural cues were unable to evoke retrieval of memories in these Alzheimer mice, optogenetic stimulation of the memory engrams returned memory retrieval to normal levels. This effect was seen in both fearful and food-reward memory, which implies that memories in Alzheimer may not be lost, but rather insufficiently retrieved. This research has therefore important implications for treating memory loss of Alzheimer’s disease.

It should be noted however, that simple contextual memories in a mouse might not be directly translatable to the significantly more complex human memories. Nevertheless, the ability to investigate the cellular correlates of memory formation and retrieval may reveal the underlying neurobiological mechanisms of numerous memory-related psychopathologies, like Alzheimer’s disease, addiction or post-traumatic stress disorder. This may in the long-term lead to the development of novel therapies that specifically target memory engrams in mice and men. The first steps towards unravelling memory correlates are to illuminate how the brain encodes memories and how do memory engrams differ from their neighbours on a molecular and cellular level? This will bring us closer to the roots of memory engrams and thereby the development of memory engram-based therapies. But for now, artificial memory formation or erasure in humans will be strictly limited to the fantasies of Hollywood.


1. Semon, R. W. The Mneme. George Allen & Unwin Ltd. (1921).

2. Hebb, D. The organization of behavior; a neuropsychological theory. Wiley & Sons (1949).

3. Deisseroth, K. Optogenetics. Nat. Methods 8, 26–29 (2011).

4. Staff, T. new. Insights of the decade. Stepping away from the trees for a look at the forest. Introduction. Science 330, 1612–1613 (2010).

5. Liu, X. et al. Optogenetic stimulation of a hippocampal engram activates fear memory recall. Nature 484, 381–5 (2012).

6. Hsiang, H.-L. et al. Manipulating a ‘Cocaine Engram’ in Mice. J. Neurosci. 34, 14115–14127 (2014).

7. Roy, D. S. et al. Memory retrieval by activating engram cells in mouse models of early Alzheimer’s disease. Nature 531, 508–12 (2016).

Esther Visser


Esther Visser is alumna van de interdisciplinaire onderzoeksmaster Brain & Cognitive Sciences aan de Universiteit van Amsterdam. Momenteel doet ze promotieonderzoek bij het Centre for Neurogenomics and Cognitive Research (CNCR) aan de Vrije Universiteit. Hier werkt ze in de onderzoeksgroep ‘Memory Circuits’ aan het ontrafelen van de moleculaire en cellulaire basis van verslavingsherinneringen.

One thought on “Memory engrams – physical substrates of memory in the brain”

  • Jan Boom says:

    Interessant onderzoek, en bijzonder dat de Mneme-hypothese van Semon hierin weer actueel wordt.
    Ben zelf bezig met het afronden van een paper over membranen waarin dat ook het geval is. Zou het op prijs stellen om over dit aspect contact te krijgen. Mijn website bevat hierover meer informatie.

Leave a Reply

Your email address will not be published. Required fields are marked *