How are human memories formed and retrieved? The clearest evidence to date emerges!

Written by: Liu Fang

Editor: Huang Shan

Layout: Li Xuewei

Remember the amnesiac girl played by Drew Barrymore in "50 First Dates"?

Due to a car accident, she would forget everything that happened the previous day when the sun rose the next day, so Adam Sandler had to give her fifty first loves so that she would fall in love again every day.

This sounds like an extremely romantic story, but for people who have lost their episodic memory function, forgetting is a personal pain.

So how do humans synthesize episodic memories? And what goes on in our brains when we recall past events?

Photo | 50 first dates stills

Researchers at UT Southwestern's Peter O'Donnell Jr. Brian Institute have discovered 103 special neurons in the human brain's hippocampus that play a key role in retrieving episodic memories. The discovery could point the way to new deep brain stimulation (DBS) therapies that could benefit patients with traumatic brain injury (TBI), Alzheimer's disease and mood disorders.

On December 15, the related paper will be published in the journal Science Direct under the title Neurons in the human medial temporal lobe track multiple temporal contexts during episodic memory processing.

Dr Bradley Lega, one of the lead authors of the paper, said: "This is the clearest evidence yet on how memories are formed and retrieved."

Exploring episodic memory in the human brain

First, let us understand what episodic memory is.

Episodic memory refers to memories of personal experiences that occurred in a certain time and place. The memory of your first kiss or a drink with friends at Christmas are both examples of episodic memory. The other concept, semantic memory, refers to the general information and facts that our brains are able to store.

Episodic memory belongs to the category of remote memory and is the most advanced and maturest memory system of human beings. A large amount of experimental evidence shows that the medial temporal lobe (MTL) of the brain plays a major role in episodic processing. The MTL structure of the brain is composed of the hippocampus and its surrounding areas (parahippocampal gyrus, entorhinal cortex, and olfactory cortex). It can gather sensory information from the occipital, parietal, temporal, and frontal cortices and integrate and process this information.

Schematic diagram of memory encoding and retrieval experiment | Source: Paper

In the experiment, neuroscientists recruited 27 patients diagnosed with intractable epilepsy at Thomas Jefferson University Hospital (TJUH) and the University of Texas Southwestern (UTSW). Before surgery, they needed to implant microelectrodes in the brain to find the epileptic focus, which provided scientists with experimental data at the neuronal scale to study the encoding mode of episodic memory.

The memory encoding experiment was a word association task, in which 308 different English nouns appeared on the screen.

Each word appeared only once during the experiment and remained on the screen for 1.6 seconds, after which the subjects entered the distractor period.

The researchers randomly gave the subjects some arithmetic problems to divert their attention. Then it was time for the wonderful memory retrieval. The subjects needed to try their best to recall these 308 fleeting English nouns. Based on the electrode recordings of the neurons, the researchers eventually identified 103 "Subsequent memory effect" (SME) neurons in the hippocampus and entorhinal cortex of the brain, which had significant characteristics in terms of discharge frequency when encoding was successful, including 51 anterior hippocampal neurons (49.5%), 31 posterior hippocampal neurons (30.1%), 4 middle or unspecified hippocampal neurons (3.9%), and 17 entorhinal cortex neurons (16.5%).

SME neuron diagram | Source: paper

When memory encoding is successful, the activity of these 103 memory-sensitive neurons increases.

When the subjects tried to retrieve the memory, especially when they recalled a large amount of detail, these neurons produced the same firing pattern. In other words, the firing frequency of the neurons in the medial temporal lobe when the subjects encoded the words was highly consistent with the firing frequency when they recalled the words, and the encoding pattern was different for each word.

The key to distinguishing memories from hallucinations

This result confirms the "serial memory effect" theory of the human brain at the neuronal scale, that is, the activity of the brain's nervous system, especially the medial temporal lobe structure and the prefrontal lobe, can predict to a certain extent whether the experienced events can be correctly recalled, and the brain activities corresponding to successful and failed memories are significantly different.

Interestingly, the order in which the words appeared, rather than their semantics, was important for recall.

In neurology, this is called the temporal cluster effect, which means that the brain tends to encode information that is adjacent in time. Therefore, when we successfully recall the past, we are usually impressed by the circumstances before and after an event. In contrast, the 103 memory-sensitive neurons discovered this time failed to predict the memory retrieval methods of different semantic types. For this reason, the researchers specifically found 100,000 documents from Wikipedia to quantify the semantic information of the words given in the experiment.

The authors believe that this may be because the experimental setup did not require subjects to retrieve memories by category, or because the activity of semantically sensitive neurons is outside the MTL area of ​​the brain, such as in the temporal lobe and prefrontal cortex.

(Source: UT Southwestern Peter O'Donnell Jr. Brian Institute website)

In addition, this study also found for the first time in the human brain that the firing time of the neurons responsible for retrieving memories is different from the firing time of other neurons in the brain. They named this phenomenon "phase offset".

The researchers believe that it is this "phase difference" that allows us to clearly distinguish between memories and reality.

Dr Bradley Lega said: "Our findings shed important light on the question, how do you know you are remembering a past memory, rather than experiencing a new one?"

The findings are particularly important for patients with Alzheimer's disease and schizophrenia.

Carol Tamminga, a professor of psychiatry at UT Southwestern, said that dysfunction in the hippocampus is the root cause of schizophrenia patients' inability to distinguish between memories and hallucinations: "The hallucinations and delusions (in patients with schizophrenia, etc.) are actually real memories that have been damaged. They are processed by the neural memory system just like 'normal' memories. It is important to understand how to use this 'phase difference' mechanism to correct these damaged memories."

We hope that in the near future this major discovery will be able to help people who are troubled by hallucinations and reality.

(Source: UT Southwestern Peter O'Donnell Jr. Brian Institute website)

References:

https://www.sciencedirect.com/science/article/pii/S1053811921009629?via%3Dihub

https://www.sciencedaily.com/releases/2021/12/211206215953.htm

http://icanbrainlab.bnu.edu.cn/webpic/image/20180910/20180910175322.pdf

Source: Academic Headlines