STM = short term memory.
LTM = long term memory.
It seems me these days that there is a fundamental unanswered question sitting in the middle of all of neuroscience and cognitive science. It's sort of like an 800 lb gorilla in the room that noone is admitting is there. Even in the true research of artificial intelligence, this question is avoided with a kind of petrifying fear that thinking about it would hurt us in some way.
Anyway... this unanswered question is how LTM and STM relate to one another. Which one dominates the other? Is LTM a "repository" of memories like a hard drive, which the STM then "queries" for needed information which is returned to it like some sort of database? Or is STM like a slave of LTM, such that the STM is like the go-fetch dog whose responsibility is only to go grab data from the immediate environment by perception? Undoubtedly this relationship between LTM and STM is far more mysterious than anyone could have ever dreamed.
As scientists we admit to each other that there only exists the present moment of time. The past moments are already gone, and the future does not yet exist. Yet as a human being sitting at your computer reading this message you have a very convincing illusion that you exist within the "space" of an hour. Or that you are somehow "Straddling" a space between your past life and what you may do in the future years. This is all a very tricky illusion created by your powerful brain. But in the really-real world, you only ever exist in a perfect present moment. So now that you have realized that, and I have you in the basic conceptual frame that I need you to be, read the following facts:
1) You are perceiving now.
2) You are engaging in active behavior now.
3) You are thinking now. You are planning future behavior.
4) You are feeling emotion, now.
Perception, action, sensing, planning, feeling and thinking are all happening simultaneously. If not that, then some very powerful aspect of your brain is perfectly SCHEDULING them into little time slots best suited to giving you effective intelligent behavior.
Now lets move onto some even more troubling facts.
1) STM does not store anything in your brain. It does not place bits of info into "RAM" like in a computer. The way that it works is by creating signal feedback between cells in the prefrontal cortex and cells in regions of the limbic system, in particular the hypothalamus. As the minutes pass by this feedback becomes progressively weaker, which is why you cannot remember a phone number that you read an hour ago, even though you could repeat it easily after having read it.
2) The attentional focus is a behavioral (active) method of perception. Human consciousness cannot deal with every changing detail in a room full of people. This is why our focus will be in one place and not another. Children who come into a room will have their focus on particular things in that room. The adults will have their focus on something else. LTM has affected the very action of STM, in an autonomic manner. Because the adults have a larger history with the world, their focus goes automatically to those aspects of the room that are "concerning" or 'important" for adults. This example casts serious doubts on whether the STM is some sort of master over the LTM, and suggests that possibly the opposite is true.
Given that the STM does not store "data", this does not stop neuroscientists from still using erronious and mythological language when talking about LTM. They often say that the short-term memories are then "copied over" to the LTM which is stored in different locations of the cortex. What is copied? Some data? There is no data. STM is a feedback loop of signals. They then go on to use language that is even more peculiar. Saying that the STM then "retrieves" said memories out of the cortex, as if data is somehow stored there.
Of course, this theoretical paradigm of store-and-retrieve must have proper doubts cast on it. If you are presented with any problem in your life, there is an aspect of thinking about the problem using logic within an abstract framework of your STM. To what degree do you call upon your past wisdom in making sense of the problem? To what degree CAN YOU call upon the bulk of your entire life as it relates to the current situation? It's absurd to think that you may be retreiving each individual item of your LTM to compare and contrast. If that were the case, then it would take you hours to decide whether to have coffee with breakfast or not.
It does not take us hours to make decisions like that. LTM's action on our current behavior is immediate and effortless. One is almost led to believe that LTM is not "informing" our current behavior, but is SHAPING it. This again leads us to the idea that perhaps the purpose of the LTM is to shape the actions of the STM, whose only real job is work-a-day perception of the environment.
Researchers in robotics have found this problem so intellectually frightning that they avoid it by assuming, a priori, that their agents must schedule LTM and STM. It's amazing to me that actual scientists are doing intellectually dubious things like assuming something must be true because their own minds cannot come up with an alternative. The usual method here is to have their robot engage in a sense-think-act cycle, where each is done separately in its own scheduled timeslot. You want examples? Look at the DARPA grand challenge cars. This is precisely how they operate.
Indeed, in some respects the battle between LTM and STM has broken AI research down into two camps. Analogously, it has split neuroscience into two distinct camps who do not talk to each other. They are the Developmental Neuroscientists and the Cognitive Neuroscientists. (You know who you are.) This ever-growing rift between these two camps may have its roots in our inability to talk objectively about how wisdom and immediate logic fight against each other, or cooperate, as it were.