healthHealth and Medicinehealthneuroscience

From Synapses To Switches: A Journey Through The Mystery Of Memory

Theories abound as to how memories are stored in our brains, but what is the current consensus?

Laura Simmons - Editor and Staff Writer

Laura Simmons

Laura Simmons - Editor and Staff Writer

Laura Simmons

Editor and Staff Writer

Laura is an editor and staff writer at IFLScience. She obtained her Master's in Experimental Neuroscience from Imperial College London.

Editor and Staff Writer

clockDec 6 2022, 14:34 UTC
human brain made of wires and circuit boards

Is the brain really so different from a computer? Image credit: Zaleman/ 

The question of how memories are stored in the human brain is an age-old puzzle. Over the years, there have been plausible theories – but the experimental evidence to back these up has not always followed close behind. With diseases that affect memory posing a huge threat to public health, there is arguably a greater need than ever before to unlock the secrets of how our brains hold onto information.

The beginnings of memory research

Santiago Ramón y Cajal was the first to suggest that synapses might be the key to memory formation, way back in the 1890s. Ramón y Cajal used a technique developed by Italian scientist Camillo Golgi to produce unbelievably detailed images of brain tissue like this one. 

Purkinje cell drawing
One of Ramón y Cajal's drawings, showing a Purkinje cell from the cerebellar cortex of a cat (public domain).

The pair went on to share a Nobel prize for their efforts, and Ramón y Cajal’s drawings are a familiar sight in neuroscience lectures to this day.

Synapses are the connections between neurons. A theory – the beginnings of which were first developed much later in 1949 by Canadian Donald Hebb – held that persistent changes in the strength of these connections over time could allow the brain to store information. This idea was called synaptic plasticity – we’ll come back to that in a minute.

Other experiments, including those of Canadian-American neurosurgeon Wilder Penfield, helped the next generation of scientists figure out where to look for memory storage in the brain. Penfield pioneered a new way of treating epilepsy. It might sound a bit extreme, but he would open up patients’ skulls to expose their brains, and – while they were awake – probe the brain with a small electrode to try to figure out which specific bits of tissue were causing the seizures. Penfield noted that stimulating certain parts of the brain triggered memories in his patients.

There was one brain region that would come to be thought of as most important of all when we're talking about memory – and to understand that, you have to know the story of Patient H.M.

Patient H.M. and the hippocampus

As a child, the person who came to be known as H.M. was involved in an accident that caused him to experience seizures. By the time he reached his late twenties, the seizures had become so debilitating that, despite taking medication, he had to give up work. 

As a last resort, H.M. underwent drastic surgery to remove part of his brain. It did the trick in terms of controlling his seizures, but there was one very notable side effect – H.M. had developed amnesia. Evocatively, he described his condition as “like waking from a dream...every day is alone in itself...”

It turned out that the part of H.M.’s brain that had been removed included one key region – the hippocampus. Buried deep inside the brain, the seahorse-shaped structure has since been considered the cornerstone of learning and recollection, and this discovery is all thanks to the legacy of patients like H.M.

We now know that the picture is a little more complex, and that no one part of the brain is responsible for everything to do with memory – for example, a recent study revealed that the cerebellum helps us store memories associated with strong emotions. However, what all this still doesn't address is how the memories are formed in the first place.

Synaptic plasticity and long-term potentiation

We promised we’d get back to this, didn’t we?

For many decades, the favorite theory among neuroscientists researching memory storage relied on the idea that synaptic connections can be made stronger when repeatedly activated. When this effect happens in a long-lasting way, it is called long-term potentiation (LTP).

For an electrical signal to be transmitted from one nerve cell to another, it has to somehow get across the synapse. This happens in a series of three main steps. First, chemical neurotransmitters are released into the synaptic cleft, or gap. These bind to receptors on the neuron at the opposite side of the cleft. This binding triggers the opening of ion channels, which are what allow electrical currents to flow.

diagram of synaptic transmission
The diagram depicts the transmission of a nerve impulse between two neurons across a synapse. Image credit: Ali DM/

LTP can be achieved either when there are more neurotransmitter molecules released, or when there are more receptors available for them to bind to. Either way, more electrical current is going to get through the ion channels, making the synaptic connection stronger.


The idea is that strong synapses are formed at the point at which a new memory is created and that this pattern of potentiation is enough to encode and store the memory.

It has been difficult to experimentally prove that LTP is the master key to unlocking the secrets of memory storage in the brain, although the evidence that shows it is very important is building. Over the years, however, many scientists have suggested that it is not the whole story.

Memory engrams

Not just the preserve of science fiction, the term "engram" goes way back to the early days of memory theory. Proposed by Richard Semon in 1904, an engram was thought to be a collection of cells that have undergone lasting chemical or physical changes, and which, when reactivated, allow for the recall of a specific memory. Just as amino acids are the individual units that make up proteins, engrams could be considered the base units of memories.

During Semon’s lifetime, his theory did not get a lot of traction. More recently, though, advances in technology may be paving the way for a discovery that Semon himself described as a “hopeless undertaking” – the biological basis of engrams.


A series of studies from a group at the Massachusetts Institute of Technology provided compelling evidence for the existence of so-called "silent engrams". The researchers were able to artificially retrieve memories of events that were stored in the brains of mice with retrograde amnesia.

The team then published a subsequent paper, adding further weight to their theory that while synaptic plasticity is involved in memory formation, it is not necessary for long-term memory storage.

“One of our main conclusions in this study is that a specific memory is stored in a specific pattern of connectivity between engram cell ensembles that lie along an anatomical pathway,” said senior author Susumu Tonegawa in a 2017 statement. "This conclusion is provocative because the dogma has been that a memory is instead stored by synaptic strength."

As a 2020 review of the subject points out, "many questions remain." However, it is exciting to think that an idea that was seeded at the turn of the last century could now be edging closer to accepted scientific consensus.

Mechanical memory

In 2021, British scientist Ben Goult spoke to IFLScience about a brand-new theory, in which the human brain is likened to a computer.

The MeshCODE, as this computer-like machinery has been dubbed, works like binary code – in this case, the 1s and 0s are different structural states of a protein called talin.

Since the initial hypothesis was published, Goult has expanded further on the potential role of talin, plus other proteins that form the eponymous "meshwork" of proteins at each synapse. Take the diagram of synaptic transmission above; the meshwork is a crisscrossing protein skeleton that sits inside the neurons at each edge of the synapse.

Most recently, the team has built to-scale models of the MeshCODE complexes to visualize how these molecules interact in real life, after discovering that all our current models are dramatically incorrect in scale. Analysis of the new models has led to some exciting updates to the theory.


"With our new analysis we suddenly realized a new discovery that these switches are introducing a large increase in the length of these molecules (the '1' state is approximately 10 times longer than the '0' state)," Goult told IFLScience. Since one talin molecule has 13 switches along its length, this means that the protein could theoretically be stretched out to nearly 1 micron.

"We realized that as the switch patterns change they would be moving the enzymes bound to the MeshCODE after that switch a quantized distance away from the active zone...of the synapse," continued Goult. 

And here we are again – back to the synapses. These new models have helped Goult and his team discover how the workings of the MeshCODE could have a knock-on effect on synaptic activity, which we know has an important role in memory.

"Here we discovered that the binary code, will be reorganizing the enzymes within the synapse...and that leads to an idea of how a binary code could coordinate synaptic activity. As the switches spatially organize enzymes relative to each other and their targets, some codes will line them up perfectly for very rapid activity and some would hold these molecules far apart preventing activity!"


The next step will be actually visualizing the talin molecules changing shape inside the neuron, something which the team hopes is not far away. They are also researching a number of nervous system diseases, plus some cancers, that they think could be impacted by altered talin signaling.

Goult has high hopes that the MeshCODE theory could be the unifying force that has so far been lacking for all of the aspects of memory we've journeyed through.

"This is the big idea of this theory, that all our memories are physically written into the shape of these molecules in our brain. I find it quite exciting to imagine that every memory we recall has a physical location where it is written in the brain. Ultimately...what I am proposing is the location of engrams, the physical substance of memory, that they are written in binary format in the shape of the MeshCODE memory molecules in the synaptic scaffolds. And they are read by altering the transmission through each synapse in the neuronal circuits. "

What does the future hold?

This is an exciting time in the field of memory research. Modern technological advances mean that theories and ideas that have been postulated since the earliest days of neuroscience can now be tested experimentally in ways they have not been before. As Goult points out, some of the biggest breakthroughs come when scientists from different backgrounds work together: "I think a lot of exciting discoveries are made at the interface between disciplines, for example mechanobiology is at the interface of engineering and biology."


Much of the functioning of the human brain still remains elusive. However, as long as there are new theories being proposed – and older ones remembered! – we will continue to edge towards a greater understanding of our minds and ourselves.

healthHealth and Medicinehealthneuroscience
  • tag
  • memory,

  • neurons,

  • synapses,

  • neuroscience,

  • neurotransmitters,

  • recall,

  • engram,

  • Neurobiology