Thursday, January 22, 2015

Memory Pixelization

The New York Times recently published and article entitled, "How The Brain Stores Trivial Memories"  worth reading in its entirety here.  The article describes that a study has suggested that there is a kind of "just-in-case" file in our minds which contains seemingly useless information, which we store for a time and then get rid of. However, it is difficult to get a bird's eye view of memory simply by reading the article. I endeavor to clarify this, as well as answer a question offered in the comments.
"Where visual stimuli is remembered, and how does the memory system process it? How is it for instance, that you can decide a particular plant is a tree, even though you may have never seen that kind of tree before? So where are visual memories stored, or are they?"
Storing "images" as a kind of guide to organizing the world around us stores a lot of unnecessary information. It is very inefficient, and slows our cognitive processes.

To understand the answer to "How do we know a tree we have never seen before is a tree?" is to understand concepts, not mass mental image storing. And to understand concepts is to understand that the conceptualization process is one of what the mind omits --- not of how much detail it commits, to memory --- leaving behind only what is essential.

Are you familiar with video encoding? Why does a Blu-Ray authored movie look better than a DVD authored movie? Because there is more space to include more data. But we also can shrink that information down without a noticeable loss in quality via encoding. How is that done? You may already have an idea, but you can read an easy guide about it here. Ever watch an encoded movie or video on your PC, pressed pause and then when you resumed there were a few seconds of pixelization? This is because the video file does not have the data to display the full image. And yet when we watch the same video without stopping, we don't notice any difference. Why is this?

What has been seen, cannot be unseen.
This is going to be highly simplified, but let's say everyone's favorite character Jar Jar Binks is on screen. For one glorious moment, he is not moving and is totally still for 100 frames. When he is still, it becomes no longer necessary to store his portion of the image 100 times into the video, only necessary to store 1 image, or one frame and leave a small pointer of how long the image is to be repeated. In this way, we have shed shrunk this video file of 99 "frames" of Jar Jar and a resulting X amount of data. Apply this to the whole movie, and this how we can shrink file sizes from their raw, unencoded state.

Modern encoders are complex. They have the ability to set a certain threshold for motion and color, so that Jar Jar need not be totally still, or have exactly the same pixel color from frame A to frame B in order to get compressed. The encoders have been programmed to isolate which datum is essential to the image, relative to the previous image, and which is inessential. In other words, it decides what datum it writes, (or "commits") and which datum it omits.

From the above Trusted Reviews article, by Gordon Kelly.

Our minds have the capacity to act as high-level video encoders (although no matter how hard we try, there will always be that one last frozen frame of Jar Jar left). However it is not really accurate to say that what we do is leave these kinds of appended "images" behind. The encoding process to the mind is called "conceptualization". And each frame is called a "percept" (that includes information from all 5 senses). Conceptualization is omission. It is omitting as much as we can. It is shrinking our "data" down to the most bite-sizable chunks, it is reducing everything to its essentials. How do we know a tree we have never seen before is a tree? Because compared to how we have encoded the image of the other trees we have seen, all that is left in our minds, conceptually speaking, is something that can't tell the difference between that old one and the new one. To the conceptual faculty, we have certain essential requirements for what it means to be called a "tree", and if it meets those, it gets called a tree. Perhaps this 100 frame new tree is simply compared to a set of 5 "frames", linked together in the mind by this conceptual pointer system: something like: "thing that exists in the open --- growing out of the ground --- starts to spread out in the middle --- brown and green --- hard." Here we have visual shape percepts, a visual-color percept and a tactile percept. But even among those, the color is not essential, as the color-blind would tell you. Anyway this is just an example.

The article describes a study where some people were given mild shocks when doing a memorization task, and the researchers found that there was a small increase in memory performance among those who were shocked while studying the cards. So why is this?
"But the experiment said nothing about the effect of trauma, which shapes memory in unpredictable ways. Rather, it aimed to mimic the arousals of daily life: The study used mild electric shocks to create apprehension and measured how the emotion affected memory of previously seen photographs."
If we want to focus on why people tend to memorize certain otherwise "trivial" events (or flashcards), it is because they are only trivial to man "the rational animal", but they are not to man "the animal". This is one aspect of survival information. Why are people drawn to look at fights? Why do they rubberneck? It is because man the animal is "wired" to seek survival information on a base level. It's evolution. If someone is given a small shock, this a very minor threat to survival, but among some personality types, that may enough to trigger the state of labeling their experience as some level of SI, and not only trying to integrate those percepts into their conceptual framework, but instead as storing those percepts as "images" --- as somewhere near the full 100 frames of Jar Jar, or as simply 10 or 20 or 30, depending on the degree of survival value determined by the body.

Where a survival information response is stimulated, and when it is not will differ among people. SI is also not a boolean function. The "arousals" of daily life exist on a scale from 0 to 100. It doesn't take being physically abused or being in a car crash or going to war to stimulate an SI response. It does not even have to be a negative or threatening experience to trigger it, it can be an affirming one, because positives can be as useful as negatives for survival. I would wager that some of this "just-in-case" memory system is simply low level SI stimuli sticking around in the mind for a time. (It's important to not to confuse conceptualization of ideas, with the storing of memory. However, these two functions overlap in the mind.)

As for trauma, SI memory is different from conceptualization. It contains a lot more "data", there are fewer omissions to the measurements that the the mind would otherwise be making. Yet, the mind cannot handle too much of it. Our brains are magnificent machines be we are limited. When people sometimes imagine "aliens who are beyond our level comprehension", the real way they would be superior to us, is to be able to handle more SI bandwidth. Where we are, to borrow more PC tech analogies, is the old PC/AT bus while the aliens are NVLink. So if that ostracized anyone the meaning is: we are bottlenecked, and in a way forced to conceptualize because of our limited hardware. A more advanced being would still be a conceptual creature, by necessity, but would also have great capacity to manage stimuli, percepts, and memory without as much stress.

We are traumatized, not simply by the event itself, but by our mind's biological inability to process and store that much data without omission.

No comments:

Post a Comment