How does the brain encode memories?

The nervous system is an aggregate of neurons and glial cells that allows information to circulate. There are no less than 100 billion neurons in the human brain and ten times more glial cells, which makes our brain an extraordinarily complex machine. In its reductionist enterprise, science wishes to pinpoint the specific functions of certain groups of neurons. One of the themes of interest in neuroscience and psychology is the study of memory.

Some reminders about memory

Memory corresponds to a set of systems responsible for encoding, storing, recalling, sorting and forgetting information. “ At the biological level, memory is underpinned by what is called the engram which refers to a network of neurons which underlies information. explains Alexandra Gros, doctor in neuroscience, specialist in the study of the hippocampus and its memory functions in the adult brain.

At the cognitive level, we distinguish several memory subsystems. From a temporal point of view, we differentiate short-term memory (or working memory) from long-term memory. The first allows you to remember information for a short time and use it, for example remembering a phone number to add it to your contacts.

The second contains non-declarative (or implicit) and declarative (or explicit) information. Non-declarative memory refers to elements whose recall does not require the use of language. Procedural memory is an example of non-declarative memory: when we ride a bike, our muscles move without us having to be aware of and put into words each movement.

In this article, it is declarative memory that will interest us. It is also differentiated into two distinct types of memories, the discovery of which is attributed to Endel Tulving, an Estonian and Canadian psychology researcher who died last September. “ We distinguish between semantic memory, which is the memory of raw, decontextualized facts, and episodic memory, which is the reliving memory of specific episodes experienced. recalls Lucile Rey, doctor in neuroscience, specialist in episodic memory in humans. Let’s illustrate this with an example: you can remember precisely a weekend that you spent in Paris (episodic memory) while knowing, independently of this weekend, that Paris is the capital of France (semantic memory) .

Two major theories of memory

To account for the functioning of the episodic memory system, there are two main theoretical models: “ Historically, Tulving’s model suggests that episodic memory is opposed to semantic memory, and refers to the questions what, where and when. More recently, Conway’s model has come to challenge this simplistic vision. According to him, our memories are a complex mixture of episodic information which would be built on semantic foundations, says Lucile Rey. She adds that the distinction is important to grasp, because in most studies which fall within the Endel Tulving model, in humans, memory is studied in a very supervised manner (everything is controlled) but very simplified. Our memories can be reduced to recalling a list of words associated with a feeling of recollection. People interested in autobiographical memory will study real memories. These memories will be complex, full of details and information. But, memory is not an exact copy of a past event, and it will be impossible for the researcher to detect recall errors. However, these two types of memory do not activate the same networks. »

We understand that the first method is easily used in the laboratory while the second includes numerous biases (already mentioned in an article on childhood amnesia). In a recent research article published in Natural Human Behavior, American, British and German researchers set out in search of the neurons which code for a particular episode using a methodology close to that recommended by the Tulving movement. They want to know if hippocampal neurons code for specific elements of an episodic memory or if they code for an entire memory. “ Their experiment seeks to decide between two theories. The first one – L’indexing theory – proposes that the role of the hippocampus in information storage is to form and maintain an index of neocortical areas activated by the episode. The second – the concept neuron theory – suggests that the hippocampus reactivates neurons independent of the context of the episode referring to invariant elements (concepts) which aggregate to form the episodic memory. explains Lucile Rey. The authors subscribe to the first theory and will seek to show the existence of neurons specific to episodes. Remember that even if they succeed, this will not be able to absolutely determine the veracity of this theory.

Epileptic patients as an experimental model

Obviously, we cannot do this kind of experiment in the general population because it is far too invasive. We then take advantage of patients in whom electrodes have been implanted in order to identify the area of ​​the brain where the epileptic focus is located (the “sick” area of ​​the brain). Regarding the experiment, here is how the researchers proceeded: “ they presented an image of an animal to the participants for a duration of 2 seconds. Next, they present an image of a face, a place, or both with the animal image and ask participants to tell a story in their head involving that composition. The authors suggest that this mimics the creation of an episodic memory. Finally, they presented the animal image again and asked participants how many other images it was associated with: 1, 2 or 0. Among those who answered that one or two images had been associated with the animal, They presented two target images (already presented) and two distractor images (never presented) and asked the participants to choose the one(s) which were associated with the animal image. explains Lucile Rey. As for electrophysiological measurements, the authors measured the action potentials of neurons in the targeted area during the presentation of images and during recall, and demonstrated that certain neurons activate in a similar way during these two phases. .

Neurons not so episodic after all

This study represents an enormous methodological work that is extremely dense and difficult to summarize. However, like any experience, it is not without limits. The first concerns the memory task: “the task used is an associative memory task which is close to episodic memory but which does not include reliving and context”, points out Lucile Rey. Alexandra Gros also adds that temporality raises questions: “ it is questionable whether participants had time to create an episodic memory given that there are only a few minutes between the encoding phase and the recall phase. It would be interesting to see if the identified neurons reactivate a few weeks after the task.” The second concerns the method of detecting episodic neurons: “ we can ask ourselves the question of the reliability of the measurement made on diseased brains and the identification of specific episodic neurons. To illustrate, let’s say that they planted stakes to measure the electrical activity of each of the small pebbles that make up the quarry that is our brain. They differentiated them using complex statistical models. But even with such a methodology, differentiating two similar stones side by side remains very difficult. The authors are aware of this and mention it in their article. However, let us remember that research in epileptic patients is valuable because it is the only way to directly measure brain activity in human beings. explains Lucile Rey.

The third experiment is inherent to all studies in neuroscience – and we have already mentioned it in this article – and concerns the size of the engram and the impossibility of measuring the simultaneous activity of all the neurons of a or several structures: “ we still don’t know how big the engram we were talking about at the beginning is. Historically, the study of cognitive functions began with brain-damaged patients (Lucile Rey reminds us that for episodic memory it is thanks to patient HM that we were able to understand the importance of the hippocampus in episodic memory ). This is how we can deduce which areas are involved in which cognitive functions. Perhaps other non-episodic neurons participate in the recall of a memory but their action potentials were not measured in the experiment. suggests Alexandra Gros. Imagine a puzzle with a piece missing. Until you finish it, you won’t be able to realize that a part is missing. It’s kind of the same problem here.