Why is developing long term memory so hard for our brain?

Electronic Memory versus Brain Memory

Anyone who has tried to learn a new language or made an attempt to memorize a poem or song, or a dialog for a play has experienced a deeply frustrating limitation of our brain — we are unable to memorize in just one or two or even half a dozen attempts. On the other hand we can instantly put almost anything into a computer or phone memory and recall it at any time. What is the key difference between our brain’s memory and the memory of an electronic device? Why is our brain’s memory seemingly so poor compared to computer memory?

Computer Memory: Localized Memory

The memory that our electronic devices use has one principal purpose — if a piece of information (in the form of 0’s and 1’s) is placed in it, the information should be kept  as long as needed unless it is erased on purpose to make room for some other information. Each 0 and 1 is assigned to specific transistors with well defined locations which are used to recall the information when needed. Every bit of information that is stored has “equal value” — the memory does not decide that some information is more important and some information is useless. “Useless information” is not wiped out or “forgotten” by the computer memory.

Even though sometimes we wish our brain’s memory was like a computer memory, evolution has decided that we have a very different type of memory. 

Most of us go to our mailbox everyday and bring back a bunch of envelopes and pamphlets. Usually a good portion of the items we bring back are junk mail and to avoid cluttering up our counters and desks we toss them out. If we don’t regularly toss out the junk mail, very soon our homes will have no space for anything. A similar phenomenon occurs with our daily experiences as we go about our lives. If we were to record on a video camera everything we see through a typical day the camera would need about 50 GB of memory. Clearly if we had the ability to store all this information in our brains, even those of us blessed with a big brain will be overwhelmed with the terabytes that we will have to store over a few years. So our brain has evolved to allow only those memories to register that are of importance to us and allow the “useless information” to fade out of our brain.

We store information and develop a long term memory of those items that are related to our well being and survival or we decide to give priority to the information to be memorized.

Our Brain’s Memory: Non-Localized Storage

We notice that we can if we are on the phone talking to someone and we are given a phone number or a name, we can recall it for a brief period of time but then it slowly gets erased from our memory. This very short term memory can hold information that takes a few seconds to speak and we can hold it for a few minutes. However, if we repeat this information several times we can “memorize” i.e. create a longer time record of it. Even an item that has been memorized gradually disappears from our memory unless we keep using it regularly.

Clearly our brain memory needs certain abilities:

1. It should not get overwhelmed by useless information so it has to have some mechanism of erasing information that is not useful for our daily life or our well being;
2. Information that is essential for our survival should be stored essentially forever.
3. It should be associative in terms of it could be triggered and recalled by seemingly unrelated triggers like a smell or song or an image.
4. It should have a lot of redundancy so that we can recall the information even with a very tiny part of the information being given to us. For example we can recognize our friends and family in a big crowd even if we see just a part of their faces.
5. The recall of information should use very little power. In computers the memory uses a lot more power than the brain does. Our brain requires just about 10 watts of power compared to computers that may require kilowatts or even megawatts of power to run.
6. The ability to store useful information should be enormous and the recall time should be as short as a fraction of a second. For example, even a child can very easily do the complex “image processing” to recognize her mother or a toy or pick out the family dog out of dozens of dogs playing in a park. To do this type of image processing requires an ability to do petabits per second processing. Even the most sophisticated modern computers cannot do this.

The brain memory is made up of neurons and synapses. A neuron is connected to other neurons through electrical signals that it can pass through a synapse. The human brain has about 100 billion neurons and over 100 trillion synapses. Information is not stored in the way it is stored on a computer memory – it is not like neurons and synapses in a certain dedicated  location store. For example, the image of a horse and the image of a car is not stored within a separate and distinct set of neurons. In the brain pieces of information are stored in overlapping neurons but the distinctness of the information is maintained by the different levels and paths of electric signals that flow through the synapses. Using this type of non-localized storage architecture allows the brain to store and recall an enormous amount of information (with a processing power as high as petabits or 1000 trillion bits per second) and the processing can occur with very little energy use.

Since the information is stored in essentially the same neuron collection the brain has to learn how to distinguish one piece of information from the other through the variation in the synapses signals. It becomes easy to make mistakes as the neural “paths” can easily get confused. So it takes a lot of repetition and association to create deep pathways between neurons to associate the path with a certain image or sound or word etc. 

While this slow and repetitive approach to storing and recalling information is frustrating it has great value for our well being. A child can recall and correctly identify images and sounds and smells at a level that even the most advanced computers of today cannot.

Encore!!! learning application is designed to help create the deep learning that is needed to create long term memory. It uses the mobile device’s memory to store and recall what we may want to learn and then the functions of 1) making a playlist of “bite-sized” elements and the use of “listen-speak-repeat” and then “test” allows one to slowly transfer the new information to a long term memory. 

Certain types of information that are directly related to our survival can be memorized in just a few repetitions. For example we can very quickly learn to remember the bark of a vicious dog. A new immigrant to a country where the language is different from his native language can very rapidly learn the basic language elements needed to hold a job that gives him funds to survive.

On the other hand items are not so essential to our survival and are “optional” such as learning a new language, learning songs and poetry, learning theatrical plays, learning sacred prayers etc. require a very large number of repetitions. When a child learns a language he or she is exposed to it hundreds of times and the parents patiently keep repeating the same words and sentences. However, when adults try to learn a new language both the learner and the teacher can get frustrated by the number of repetitions needed. The built-in functionalities of Encore!!! learning app facilitate such a learning process.  

Author: Dr. Jasprit Singh, President Gurmentor, Inc.

A Learning Company https://gurmentor.com

Professor Emeritus, Electrical Engg. & Comp. Sci. and Applied Physics

University of Michigan, Ann Arbor