Jan 14 2019

Our Memories Work Backwards

One more piece to the memory puzzle seems to be falling into place. The question is – what steps do our brains go through when recalling a memory? Researchers have been focusing on visual memory, because it is easiest to model and image, and they have found that memories are recalled in a reverse of the process by which they are formed.

A recent study in Nature Communications replicates the overall findings of a previous study published in PNAS. Both studies looked at the visual system and found essentially the same thing.

When we perceive an object, first our brain receives an image from the retina. By the time this image gets to the visual cortex some basic image processing has already occurred at the subcortical level. Then the cortex puts the image together, sharpens up contrast and lines, interprets size and distance, shadows and movement, etc. The brain then tries to find a match in its catalogue of known things. Once a match is found, actually, that information is then communicated back down to the more basic visual layers and the image is adjusted to enhance the match – lines are filled in, extraneous details are suppressed, assumptions of size and distance are adjusted.

Then the now identified object is sent to even higher brain areas (higher in this network) to afford meaning to the object. If your brain thinks the object has agency, this connects to the emotional centers in order to remember what you feel about the object. Connections are also made to memories about the object. Let’s call these thematic memories. So our brains build the image up from basic details, to complex shapes, then to known objects, and finally to feelings, connections, meaning and memories.

But what about when you recall the object that you previously saw? Both of these studies, using visual memories, found that the brain works backwards. First the thematic areas of the brain light up, then progressively more basic areas of visual processing. Media reporting on these studies emphasize that this is backward from how visual memories are made in the first place. However, this is only sort-of true. Remember – even when perceiving things, information goes simultaneously from the details to the themes, but then back down from the themes to the details. Perception and memory formation is bidirectional.

Memory recall, at least according to these studies, seems to be primarily from thematic (technically called “semantic”) to detailed (technically called “episodic”) memory. First we remember the big picture – how do we feel about the object (I say “object” but I am including people, places, or anything visually processed)? What is it’s meaning for us? Is this something threatening, something to eat, something I love? The details get reconstructed later.

This has important consequences for how our memories work, and is completely in line with prior research. Let’s say I am trying to remember having seen a giraffe. First I remember that what I saw made me feel a certain way, and then that it was a giraffe, and then I fill in the details about what the giraffe looked like. But here’s the thing, and this has also been demonstrated by other research, what we think we are seeing or remembering (the theme or gestalt – it was a giraffe) has more of an influence on what we both see and remember than the actual details of what we saw.

So while the media reporting about these studies say that this is “backwards” and different than what researchers expected, I disagree. This is completely in line with existing research. Both perception and memory work as much or more from theme to details as they do from details to theme.

Of course, this applies not just to mundane things like seeing a giraffe at the zoo. This applies to being an unexpected witness to a crime or serious event. It applies to experiencing something unusual, and then interpreting it in line with cultural expectations. If you think you saw a flying saucer, then that is what you remember. Then when you recall the event, you start with the theme that you saw a flying saucer, and then you fill in the details, and those details are determined more by the remembered theme than the actual details of what you saw. Therefore those details in your memory will morph over time, to get further and further in line with the theme (something which has also already been well established).

Why would our memories work this way? It’s always hard to make firm conclusions about evolutionary causes like this, there are just too many variables, and the causes are likely to be many and complex. First, causes are not restricted to adaptation. Sometimes biology works a certain way because it was constrained, or simply because of the quirky details of history.

But we can ask what possible adaptation these features of memory might have. The implication is that the broad themes of memory are more important for survival than the tiny details. Something dangerous lives over there, so don’t got there (at least not alone and unarmed).

We can, of course, still accurately remember some details. Sometimes details do matter, and if we focus on them we may be able to remember them. But we cannot rely upon our memory for details. Also, it may be true that in our complex technological society, accurately perceiving and remembering details is increasingly important, and our memories are just poorly suited to this.

We need to compensate for our theme-biased memories with analytical tools and aids. This is obvious to see when it comes to things like remembering phone numbers – we just store them in our phone’s address book. But it is also important when it comes to understanding the world and having opinions about the important questions. Our “gut feelings” are not enough. They are basic, flawed, and biased – but they can give us the powerful illusion of knowledge. Understanding how our brains work (and this is one example) will help break the illusion.


No responses yet