How Do We Store Information to Memory?

Division of Long-Term Memory

Long-term memory stores two kinds of information - explicit and implicit. Explicit memory, or declarative memory, is composed of information that can be verbally communicated; while implicit memory, or nondeclarative memory, is composed of information that can be performed and knowing is relatively subconscious. For example, you cannot teach anyone how to dribble a basketball just by saying "how", but you need to show it; and the person who learns it cannot tell how he does it either. The distinction between explicit and implicit memory comes from the case of H.M., a man who had a severe case of epilepsy. H.M. underwent brain surgery, where the hippocampus and portions of the temporal lobes of both of his hemispheres were removed. As a result, H.M. developed partial anterograde amnesia, where prospective implicit information is preserved, and prospective explicit information is forgotten. For three days, H.M. underwent a simple training where he was instructed to trace a star-shaped figure through a mirror. Surprisingly, he was able to learn the task and performed it efficiently, even if he did not know that he has already done it.

Canadian cognitive psychologist Endel Tulving (1972, 1989, 2000) further classifies explicit memory into two - episodic memory and semantic memory. Episodic memory is memory for events. Retrospective episodic memory, also known as autobiographical memory, is memory for past events; while prospective episodic memory is memory for future activities, which is composed of timing (when the activity is intended to be done) and content (the activity itself). Semantic memory, on the other hand, is memory for meanings, or simply, meaningful memory. It is made up of information we know to be facts; for example, vocabulary, label for things, or professions. The distinction between episodic memory and semantic memory comes from reported cases of retrograde amnesia, like a young man named K.C., who developed retrograde amnesia after a motorcycle accident. K.C. only knew the things told him by his family members, like his brother drowning ten years ago, but subsequent tests revealed that he could not really remember them. Episodic and semantic memory are distinct from each other in different ways: in terms of the type of information that is stored, or the unit of information, where episodic memory is composed of events or episodes, and semantic memory is composed of facts, ideas, and concepts; in terms of how information is organized, where episodic memory is either time-based or event-based, and semantic memory is conceptual; in terms of the role of emotion, where emotion is more important in episodic memory than in semantic memory; in terms of the process of remembering, where remembering episodic memory is deliberate or effortful, and remembering semantic memory is automatic; in terms of how individuals report memories, where individuals say "I remember" with episodic memory, and "I know" with semantic memory; and, in terms of its admissibility as legal testimony, where episodic memory is admissible in court, and semantic memory is inadmissible in court. A few psychologists doubt the distinction between episodic and semantic memory, claiming that there is no such information as purely episodic or semantic, and that explicit memory lie in gray areas. However, Tulving (1983, 2000) argues that both types of explicit memory work together and not against each other for normal individuals, but are subsequently differentiated in brain-damaged individuals.

Just as explicit memory is classified, implicit memory is also classified into three - procedural, primed, and classically conditioned memory. Procedural memory is memory for skills, such as writing, dancing, and reading. Primed memory is memory activated through priming, or the process of subconsciously introducing information beforehand to aid in remembering. A stem-completion task is used to test primed memory. For example, you are introduced to the words "catatonia", "schizophrenia" and "paranoia", and then given a stem-completion task like this: _ at _ _ oni _. Because you were primed with the word "catatonia", it would be easy for you to guess the right word. Classically conditioned memory, lastly, is memory developed from classical conditioning, like fears.

Organization of Long-Term Memory

Long-term memory is organized in different ways, in terms of hierarchies, semantic networks, schemas, and connectionist networks.

The Hierarchical Organization of information in long-term memory entails classifying information as general or specific.

The Semantic Network Organization of information in long-term memory is also hierarchical, but is somewhat irregular and distorted. For example, a canary is nearer to the node of "bird" than an ostrich is. Semantic networks organize abstract concepts, in tens of thousands of units, where new memories are formed by adding new nodes. Organization by means of semantic networks explains why cramming does not put information into long-term memory and why elaboration is important. Because information need to be tied into existing networks in long-term memory, information that fails to create needed associations are therefore forgotten.

Schemas are pre-existing mental concepts or frameworks established from patterns of already stored collections of information. Schematic Organization of information in long-term memory makes use of large knowledge structures, where new memories are formed by adding new schemas or modifying old ones. By developing connections, schemas influence how new information is encoded and retrieved. A schema may be specific, as with scripts (schemas for events), where elements such as physical features, people, and typical occurrences are stored. For example, you may have a script for parties, storing such information as the common locations where they are held, the kind of people who attend them, and the typical activities done in such occasions.

The notion of schemas as a form of organizing information in long-term memory started with Sir Frederick Bartlett's (1932) studies on how people remember stories. He speculated that an individual's background serves as schemas that influence how stories are encoded, and those will reveal themselves when the individual is asked to reconstruct the stories from memory. One story, the "War of the Ghosts", an English translation of an American Indian folktale, confirmed his hypothesis. Bartlett asked British children of the upper and middle classes to read the story twice and to write down the tale from memory after a 15-minute lapse interval. The story went like this:

"Two young men from Egulac went to the river to hunt seals. The surroundings suddenly became foggy and calm, and the men heard war cries. They hid behind a log, thinking that the war cries are coming from a war party. Five men in one of the canoes invited them to make war with the people up the river to the other side of Kalama. One of the young men said that they have no arrows, but the war men said that there were arrows in the canoe. One of the young men said that he could not go because he could be killed and his relatives knew nothing about the war, so he volunteered the other. The young man went home, but the other went to war and was hit by an arrow, although he felt nothing, so that he realized that the war men were ghosts. The war men decided to retreat as the young man was hit. The young man went back to Egulac, and around a fire, told his tribesman of his story. But when the sun rose up, he fell down, and something black came out of his mouth. His face was contorted. The people jumped and cried. The young man was dead."

Bartlett's British participants had a number of inconsistencies in their reconstructed stories. The "something black" became "blood" with one participant and "condensed air" with another. Somebody wrote that the young men were hunting "beavers", instead of "seals", and another claimed that the young man's death was due to "fever". This shows how schemas influence the way information is encoded, stored, and retrieved from memory.

The theory of connectionism, or parallel distributed processing, states that long-term memory is organized by way of connectionist networks. Connectionist Network Organization refers to memory storage in terms of small unit connections among nodes of neurons - the location of neural activity, where interconnections may be due to inhibition or excitation upon reaching a critical level for activation - in tens of millions of units, where new memories are formed by increasing the strength of neural connections. Connectionism is increasingly becoming popular in memory research for two reasons: It is reliable in predicting results on memory experiments, especially when programmed in a computer; and, it supports and stimulates brain research by identifying in which areas of the brain long-term memories are typically stored.

Localization of Long-Term Memory

Besides identifying the different types of long-term memory, and how information is organized in long-term memory, memory researchers, through connectionism, also delved into discovering the specific brain locations where memories are stored. Historically, the first memory researchers who attempted to provide insight as to where long-term memory is stored are Karl Lashley and Donald Hebb. Karl Lashley (1950) spent a lifetime looking for specific regions of the brain where long-term memory is stored. He trained thousands of rats to learn a maze, and cut various portions of their brain to see if the missing brain part can affect their maze performance. Unfortunately, the loss of the brain parts did not affect the rats' ability to remember the maze's pathway, so he concluded that long-term memory is not stored in specific locations. In response to Lashley's findings, Canadian psychologist Donald Hebb (1949, 1980) suggested instead that assemblies of different cells work together to represent long-term memory. Today, memory researchers have found that long-term memory, at least for humans, is stored in the neurons, neurotransmitters, neuronal connections, and in some areas of the brain. The following research studies and expert opinions shed light into what memory researchers today know about the physical aspect of memory storage, specifically long-term memory:

• Larry Squire (1990) was the first to estimate that information in long-term memory is clustered in about 1,000 neurons;
• Single neurons activate in response to specific physical features of external stimuli, such as the face, eye, and hair color;
• Neurotransmitters act like the ink of memory. Eric Kandel and James Schwartz (1982) accustomed a sea slug, which is composed of only 10,000 neurons, to stop its reflex withdrawal. Then, through classical conditioning, Kandel and Schwartz prodded and shocked the gill of the sea slug. They then discovered that the sudden release of serotonin was responsible for the sea slug's immediate learning of the pairing, and thereby influenced it to immediately react by withdrawing its gill;
• Neuronal connections are strengthened if neurons are consistently simultaneously activated, a process called long-term potentiation. Shakesby, Anwyl and Rowan (2002) demonstrated how long-term potentiation works by means of a drug that induces simultaneous neuronal activation, and consequently strengthens neuronal connection. Service (1994) observed that when such drugs are administered to rats, they commit fewer mistakes when learning a maze. Tang et al. (1999) and Tsien (2000) showed that altering some of the mice's genes could increase long-term potentiation in the hippocampus and in some areas of the brain;
• Lynch (1990) reacted against Lashley's (1950) conclusion and claimed that different types of long-term memory involve a limited number of brain systems and pathways and with different means of working;
• The brain parts involved in explicit long-term memory are the hippocampus, temporal lobes, frontal lobe, and some areas of the limbic system (Traub, 2004; Zola & Squire, 2001). Brewer et al. (1998) found that pictures that elicit high activation of the pre-frontal lobe and some regions of the hippocampus are remembered most. Burgess et al. (2001) showed that explicit memory travels from the hippocampus to the frontal lobe. He also found that the frontal lobe is involved in both retrospective and prospective episodic explicit memory. Otten, Henson and Rugg (2001) found that encoding of explicit memory often takes place in the left frontal lobe, and that retrieval of explicit memory usually occurs in the right frontal lobe. Cabeza (2002) adds that the observed reversal of process location in old age, as elderlies typically make use of their left frontal lobes instead to retrieve explicit memory, may be a way for them to compensate for some retrieval difficulties. McGaugh (2004) and Siegle et al. (2002) found that the amygdala is involved in emotional episodic memories; and lastly,
• The brain parts involved in implicit long-term memory are the cerebellum, for procedural memory (Krupa, Thompson & Thompson, 1993); and the temporal lobes and hippocampus, for primed memory (Jernigan, Ostergaard & Tennemanotestine, 2001; Vasuno & others, 2000).