On this page, I plan to collate as much information that I can on how our memory works; how we form and retrieve memories, how different parts of the brain is responsible for different memory functions. For example, the hippocampus is responsible for our ability to learn new things.
Table of Contents
How does Memory Work to Store and Retrieve Information?
Memories are not “stored” and “retrieved” the way files are on a computer.
To the extent that memories are “stored” at all, they are stored as millions of molecular changes at the connection points (synapses) between neurons.
How is Human Memory Stored?
In general, the brain does not store new memories by a centrally controlled process, the way one might file something away in a file cabinet. Memory is a decentralized, distributed process in the brain. All parts of the brain are adapting to new information all the time. The main criteria the neural circuits use to decide when to alter themselves to accommodate new information is repetition. The more frequently two or more unusual pieces of information co-occur, the more the brain decides they must be important and prioritizes them for associative retention by altering neural circuits. The brain does not really store “information” but rather adapts to statistical relationships among signals.
The Uniqueness of a Stimuli Determines our Ability to Remember It
Another criteria by which the brain selects memory for storage is perceived value. If something unexpected happens, the brain records everything that led up to it. This process is mediated by the neurotransmitter dopamine, and is one of the reasons why dopamine-related drugs (cocaine, amphetamine) are addictive.
We Need the Hippocampus to Learn New Things
There is one region of the brain that is specialized for the formation of new declarative memories, and that is the hippocampus. A famous patient, commonly called HM, had part of his hippocampus removed to prevent epilepsy, and from that day forward, he was unable to form new memories. He still fully remembered his past. This led to the realization that the hippocampus is specialized for the formation of new memories.
One current theory is that memories are coded temporarily by the hipppocampus throughout the day, and then at night while we sleep (or dream?), the memories are transferred to the rest of the brain via a process called memory consolidation. The specifics of how this might work are still being investigated.
To Remember is to Connect the Dots
Most recall occurs by association. An event or situation that is somehow similar to a memory from the past, or a piece of knowledge, causes that memory or knowledge to spontaneously re-enter consciousness. The spontaneous retrieval of memory by association is closer to a process of “reconstruction” than retrieval. Clues about a past mental state are activated until something coherent emerges. That coherent “recreated” brain state may or may not be an accurate reproduction of the past, as the high error rate of eye-witness testimony shows.
What are the Different Types of Human Memory?
There are many different types of memory. Each type of memory works according to different neural mechanisms, has different properties, and in most cases is managed by different parts of the brain.
Two of the more commonly known types of memory are procedural memory (memory for skills, such as driving) and declarative memory (memory for abstract knowledge, such as calculus).
What is Procedural Memory?
Procedural memory — and in the case of driving, what some would call “muscle memory” — makes use of a large number of diverse and older brain areas. The skill of driving, as with riding a bicycle, has many components to it. There is a coordination skill — how to control the muscle movement patterns necessary to turn the steering wheel, hit the right pedal, etc. There are visual interactive action skills, such as responding to seeing another car or an obstacle come too close. There are automatic motor reflex skills, such as compensating for the resistance of the steering wheel or looking over your shoulder before changing lanes. This web of interlocking behavior is mastered by many different regions of the brain working in coordination.
The prefrontal cortex (part of the cerebral cortex) handles driving goals and task breakdown. The basal ganglia learns interactive motor patterns and visual obstacle avoidance. The cerebellum learns low-level coordination and motor action smoothing, like not over-correcting the steering wheel. These systems learn together and learn from each other, forming a deeply embedded, hierarchically organized, mutually-reinforcing ecosystem of related actions and reactions via thousands mastered microrelationships. Once this interactive motor skill system has established itself, reactivating it “brings it all back”, hence the saying that you never forget how to ride a bicycle.
What is Declarative Memory?
Declarative memory, which is required to learn calculus, appears to be stored primarily in the cerebral cortex. Calculus involves a lot of abstract knowledge. The process of solving an equation is a skill, but it is primarily a cognitive skill involving abstract manipulations in short-term memory using explicitly memorized steps. The only visuo-motor aspects of calculus involve moving the pencil on the paper and recognizing the shape of mathematical symbols. Abstract declarative memories fade faster if there is not regular repetition to keep them refreshed. Declarative memory is intertwined with language.
Other than procedural and declarative memory, other types include (but are not limited to) episodic, semantic, and working memories.
How Does the Brain Forget Memories?
As far as “where does learning vanish,” it doesn’t go anyplace. It just fades and gets disrupted with the addition of new information (a process called “interference”). The disruption is not a replacement. It’s more like graffiti on a wall constantly being painted over by more graffiti. Whatever images were put there most recently are the easiest to make out. Of the earlier images, often the most interesting part is left exposed when new designs are added, while the least interesting part gets painted over. So as the wall evolves, it gets richer and richer with a high density of interesting design, while the old stuff is mostly obliterated unless it is refreshed with new paint. And yet the first large image, while possibly gone, shaped everything that came after it, so its presence is still felt even if it is no longer visible (think early childhood experiences).
Rather than specks of paint on concrete, the memory substrate used by the brain are the trillions of microscopic chemical changes scattered across trillions of synapses in the brain’s neural networks. Still, the idea of interference and fading is approximately the same.