SENSORY MEMORY

Sensory memory is the shortest-term element of memory. It is the ability to retain impressions of sensory information after the original stimuli have ended. It acts as a kind of buffer for stimuli received through the five senses of sight, hearing, smell, taste and touch, which are retained accurately, but very briefly. For example, the ability to look at something and remember what it looked like with just a second of observation is an example of sensory memory.

The stimuli detected by our senses can be either deliberately ignored, in which case they disappear almost instantaneously, or perceived, in which case they enter our sensory memory. This does not require any conscious attention and, indeed, is usually considered to be totally outside of conscious control. The brain is designed to only process information that will be useful at a later date, and to allow the rest to pass by unnoted. As information is perceived, it is therefore stored in sensory memory automatically and unbidden. Unlike other types of memory, the sensory memory cannot be prolonged via rehearsal.

Sensory memory is an ultra-short-term memory and decays or degrades very quickly, typically in the region of 200 - 500 milliseconds (1/5 - 1/2 second) after the perception of an item, and certainly less than a second (although echoic memory is now thought to last a little longer, up to perhaps three or four seconds). Indeed, it lasts for such a short time that it is often considered part of the process of perception, but it nevertheless represents an essential step for storing information in short-term memory.

The sensory memory for visual stimuli is sometimes known as the iconic memory, the memory for aural stimuli is known as the echoic memory, and that for touch as the haptic memory. Smell may actally be even more closely linked to memory than the other senses, possibly because the olfactory bulb and olfactory cortex (where smell sensations are processed) are physically very close - separated by just 2 or 3 synapses - to the hippocampus and amygdala (which are involved in memory processes). Thus, smells may be more quickly and more strongly associated with memories and their associated emotions than the other senses, and memories of a smell may persist for longer, even without constant re-consolidation.

Experiments by George Sperling in the early 1960s involving the flashing of a grid of letters for a very short period of time (50 milliseconds) suggest that the upper limit of sensory memory (as distinct from short-term memory) is approximately 12 items, although participants often reported that they seemed to "see" more than they could actually report.

Information is passed from the sensory memory into short-term memory via the process of attention (the cognitive process of selectively concentrating on one aspect of the environment while ignoring other things), which effectively filters the stimuli to only those which are of interest at any given time.

SHORT-TERM (WORKING) MEMORY

Short-term memory acts as a kind of “scratch-pad” for temporary recall of the information which is being processed at any point in time, and has been refered to as "the brain's Post-it note". It can be thought of as the ability to remember and process information at the same time. It holds a small amount of information (typically around 7 items or even less) in mind in an active, readily-available state for a short period of time (typically from 10 to 15 seconds, or sometimes up to a minute).

For example, in order to understand this sentence, the beginning of the sentence needs to be held in mind while the rest is read, a task which is carried out by the short-term memory. Other common examples of short-term memory in action are the holding on to a piece of information temporarily in order to complete a task (e.g. “carrying over” a number in a subtraction sum, or remembering a persuasive argument until another person finishes talking), and simultaneous translation (where the interpreter must store information in one language while orally translating it into another). What is actually held in short-term memory, though, is not complete concepts,but rather links or pointers (such as words, for example) which the brain can flesh out from it's other accumulated knowledge.

However, this information will quickly disappear forever unless we make a conscious effort to retain it, and short-term memory is a necessary step toward the next stage of retention, long-term memory. The transfer of information to long-term memory for more permanent storage can be facilitated or improved by mental repetition of the information or, even more effectively, by giving it a meaning and associating it with other previously acquired knowledge. Motivation is also a consideration, in that information relating to a subject of strong interest to a person, is more likely to be retained in long-term memory.

The term working memory is often used interchangeably with short-term memory, although technically working memory refers more to the whole theoretical framework of structures and processes used for the temporary storage and manipulation of information, of which short-term memory is just one component.

The central executive part of the prefrontal cortex at the front of the brain appears to play a fundamental role in short-term and working memory. It both serves as a temporary store for short-term memory, where information is kept available while it is needed for current reasoning processes, but it also "calls up" information from elsewhere in the brain. The central executive controls two neural loops, one for visual data (which activates areas near the visual cortex of the brain and acts as a visual scratch pad), and one for language (the "phonological loop", which uses Broca's area as a kind of "inner voice" that repeats word sounds to keep them in mind). These two scratch pads temporarily hold data until it is erased by the next job.

Although the prefrontal cortex is not the only part of the brain involved - it must also cooperate with other parts of the cortex from which it extracts information for brief periods - it is the most important, and Carlyle Jacobsen reported, as early as 1935, that damage to the prefrontal cortex in primates caused short-term memory deficits.

The short-term memory has a limited capacity, which can be readily illustrated by the simple expedient of trying to remember a list of random items (without allowing repetition or reinforcement) and seeing when errors begin to creep in. The often-cited experiments by George Miller in 1956 suggest that the number of objects an average human can hold in working memory (known as memory span) is between 5 and 9 (7 ± 2, which Miller described as the “magical number”, and which is sometimes referred to as Miller's Law). However, although this may be approximately true for a population of college students, for example, memory span varies widely with populations tested, and modern estimates are typically lower, of the order of just 4 or 5 items.

The type or characteristics of the information also affects the number of items which can be retained in short-term memory. For instance, more words can be recalled if they are shorter or more commonly used words, or if they are phonologically similar in sound, or if they are taken from a single semantic category (such as sports, for example) rather than from different categories, etc. There is also some evidence that short-term memory capacity and duration is increased if the words or digits are articulated aloud instead of being read sub-vocally (in the head).

The relatively small capacity of the short-term memory, compared to the huge capacity of long-term memory, has been attributed by some to the evolutionary survival advantage in paying attention to a relatively small number of important things (e.g. the approach of a dangerous predator, the proximity of a nearby safe haven, etc) and not to a plethora of other peripheral details which would only interfere with rapid decision-making.

"Chunking" of information can lead to an increase in the short-term memory capacity. Chunking is the organization of material into shorter meaningful groups to make them more manageable. For example, a hyphenated phone number, split into groups of 3 or 4 digits, tends to be easier to remember than a single long number. Experiments by Herbert Simon have shown that the ideal size for chunking of letters and numbers, whether meaningful or not, is three. However, meaningful groups may be longer (such as four numbers that make up a date within a longer list of numbers, for example). With chunking, each chunk represents just one of the 5 - 9 items that can be stored in short-term memory, thus extending the total number of items that can be held.

It is usually assumed that the short-term memory spontaneously decays over time, typically in the region of 10 - 15 seconds, but items may be retained for up to a minute, depending on the content. However, it can be extended by repetition or rehearsal (either by reading items out loud, or by mental simulation), so that the information re-enters the short-term store and is retained for a further period. When several elements (such as digits, words or pictures) are held in short-term memory simultaneously, they effectively compete with each other for recall. New content, therefore, gradually pushes out older content (known as displacement), unless the older content is actively protected against interference by rehearsal or by directing attention to it. Any outside interference tends to cause disturbances in short-term memory retention, and for this reason people often feel a distinct desire to complete the tasks held in short-term memory as soon as possible.

The forgetting of short-term memories involves a different process to the forgetting of long-term memories. When something in short-term memory is forgotten, it means that a nerve impulse has merely ceased being transmitted through a particular neural network. In general, unless an impulse is reactivated, it stops flowing through a network after just a few seconds.

Typically, information is transferred from the short-term or working memory to the long-term memory within just a few seconds, although the exact mechanisms by which this transfer takes place, and whether all or only some memories are retained permanently, remain controversial topics among experts. Richard Schiffrin, in particular, is well known for his work in the 1960s suggesting that ALL memories automatically pass from a short-term to a long-term store after a short time (known as the modal or multi-store or Atkinson-Schiffrin model).

However, this is disputed, and it now seems increasingly likely that some kind of vetting or editing procedure takes place. Some researchers (e.g. Eugen Tarnow) have proposed that there is no real distinction between short-term and long-term memory at all, and certainly it is difficult to demarcate a clear boundary between them. However, the evidence of patients with some kinds of anterograde amnesia, and experiments on the way distraction affect the short-term recall of lists, suggest that there are in fact two more or less separate systems.

LONG-TERM MEMORY

Long-term memory is, obviously enough, intended for storage of information over a long period of time. Despite our everyday impressions of forgetting, it seems likely that long-term memory actually decays very little over time, and can store a seemingly unlimited amount of information almost indefinitely. Indeed, there is some debate as to whether we actually ever “forget” anything at all, or whether it just becomes increasingly difficult to access or retrieve certain items from memory.

Short-term memories can become long-term memory through the process of consolidation, involving rehearsal and meaningful association. Unlike short-term memory (which relies mostly on an acoustic, and to a lesser extent a visual, code for storing information), long-term memory encodes information for storage semantically (i.e. based on meaning and association). However, there is also some evidence that long-term memory does also encode to some extent by sound. For example, when we cannot quite remember a word but it is “on the tip of the tongue”, this is usually based on the sound of a word, not its meaning.

Physiologically, the establishment of long-term memory involves a process of physical changes in the structure of neurons (or nerve cells) in the brain, a process known as long-term potentiation, although there is still much that is not completely understood about the process. At its simplest, whenever something is learned, circuits of neurons in the brain, known as neural networks, are created, altered or strengthened. These neural circuits are composed of a number of neurons that communicate with one another through special junctions called synapses. Through a process involving the creation of new proteins within the body of neurons, and the electrochemical transfer of neurotransmitters across synapse gaps to receptors, the communicative strength of certain circuits of neurons in the brain is reinforced. With repeated use, the efficiency of these synapse connections increases, facilitating the passage of nerve impulses along particular neural circuits, which may involve many connections to the visual cortex, the auditory cortex, the associative regions of the cortex, etc.

This process differs both structurally and functionally from the creation of working or short-term memory. Although the short-term memory is supported by transient patterns of neuronal communication in the regions of the frontal, prefrontal and parietal lobes of the brain, long-term memories are maintained by more stable and permanent changes in neural connections widely spread throughout the brain. The hippocampus area of the brain essentially acts as a kind of temporary transit point for long-term memories, and is not itself used to store information. However, it is essential to the consolidation of information from short-term to long-term memory, and is thought to be involved in changing neural connections for a period of three months or more after the initial learning.

Unlike with short-term memory, forgetting occurs in long-term memory when the formerly strengthened synaptic connections among the neurons in a neural network become weakened, or when the activation of a new network is superimposed over an older one, thus causing interference in the older memory.

Over the years, several different types of long-term memory have been distinguished, including explicit and implicit memory, declarative and procedural memory (with a further sub-division of declarative memory into episodic and semantic memory) and retrospective and prospective memory.

DECLARATIVE (EXPLICIT) & PROCEDURAL (IMPLICIT) MEMORY

Long-term memory is often divided into two further main types: explicit (or declarative) memory and implicit (or procedural) memory.

Declarative memory (“knowing what”) is memory of facts and events, and refers to those memories that can be consciously recalled (or "declared"). It is sometimes called explicit memory, since it consists of information that is explicitly stored and retrieved, although it is more properly a subset of explicit memory. Declarative memory can be further sub-divided into episodic memory and semantic memory.

Procedural memory (“knowing how”) is the unconscious memory of skills and how to do things, particularly the use of objects or movements of the body, such as tying a shoelace, playing a guitar or riding a bike. These memories are typically acquired through repetition and practice, and are composed of automatic sensorimotor behaviours that are so deeply embedded that we are no longer aware of them. Once learned, these "body memories" allow us to carry out ordinary motor actions more or less automatically. Procedural memory is sometimes referred to as implicit memory, because previous experiences aid in the performance of a task without explicit and conscious awareness of these previous experiences, although it is more properly a subset of implicit memory.

These different types of long-term memory are stored in different regions of the brain and undergo quite different processes. Declarative memories are encoded by the hippocampus, entorhinal cortex and perirhinal cortex (all within the medial temporal lobe of the brain), but are consolidated and stored in the temporal cortex and elsewhere. Procedural memories, on the other hand, do not appear to involve the hippocampus at all, and are encoded and stored by the cerebellum, putamen, caudate nucleus and the motor cortex, all of which are involved in motor control. Learned skills such as riding a bike are stored in the putamen; instinctive actions such as grooming are stored in the caudate nucleus; and the cerebellum is involved with timing and coordination of body skills. Thus, without the medial temporal lobe (the structure that includes the hippocampus), a person is still able to form new procedural memories (such as playing the piano, for example), but cannot remember the events during which they happened or were learned.

Perhaps the most famous study demonstrating the separation of the declarative and procedural memories is that of a patient known as “H.M.”, who had parts of his medial temporal lobe, hippocampus and amygdala removed in 1953 in an attempt to cure his intractable epilepsy. After the surgery, H.M. could still form new procedural memories and short-term memories, but long-lasting declarative memories could no longer be formed. The nature of the exact brain surgery he underwent, and the types of amnesia he experienced, allowed a good understanding of how particular areas of the brain are linked to specific processes in memory formation. In particular, his ability to recall memories from well before his surgery, but his inability to create new long-term memories, suggests that encoding and retrieval of long-term memory information is mediated by distinct systems within the medial temporal lobe, particularly the hippocampus. The fact that he was able to learn hand-eye coordination skills such as mirror drawing, despite having absolutely no memory of having learned or practised the task before, also suggested the existence different types of long-term memory, which are now known as declarative and procedural memories

There is strong evidence, notably by studying amnesic patients and the effect of priming, to suggest that implicit memory is largely distinct from explicit memory, and operates through a different process in the brain. Studies of the effects of amnesia have shown that it is quite possible to have an intact implicit memory despite a severely impaired explicit memory. Priming is the effect in which exposure to a stimulus influences response to a subsequent stimulus, so that, for instance, if a person reads a list of words including the word “concert”, and is later asked to complete a word starting with “con”, there is a higher probability that they will answer “concert” than, say, “contact”, “connect”, etc. Studies from amnesic patients indicate that priming is controlled by a brain system separate from the medial temporal system that supports explicit memory

EPISODIC & SEMANTIC MEMORY

Declarative memory can be further sub-divided into episodic memory and semantic memory.

Episodic memory represents our memory of experiences and specific events in time in a serial form, from which we can reconstruct the actual events that took place at any given point in our lives. It is the memory of autobiographical events (times, places, associated emotions and other contextual knowledge) that can be explicitly stated. Individuals tend to see themselves as actors in these events, and the emotional charge and the entire context surrounding an event is usually part of the memory, not just the bare facts of the event itself.

Semantic memory, on the other hand, is a more structured record of facts, meanings, concepts and knowledge about the external world that we have acquired. It refers to general factual knowledge, shared with others and independent of personal experience and of the spatial/temporal context in which it was acquired. Semantic memories may once have had a personal context, but now stand alone as simple knowledge. It therefore includes such things as types of food, capital cities, social customs, functions of objects, vocabulary, understanding of mathematics, etc. Much of semantic memory is abstract and relational and is associated with the meaning of verbal symbols.

The semantic memory is generally derived from the episodic memory, in that we learn new facts or concepts from our experiences,

and the episodic memory is considered to

support and underpin semantic memory. A gradual transition from episodic to semantic memory can take place, in which episodic memory reduces its sensitivity and association to particular events, so that the information can be generalized as semantic memory.

Both episodic memory and semantic memory require a similar encoding process. However, semantic memory mainly activates the frontal and temporal cortexes, whereas episodic memory activity is concentrated in the hippocampus, at least initially. Once processed in the hippocampus, episodic memories are then consolidated and stored in the neocortex. The memories of the different elements of a particular event are distributed in the various visual, olfactory and auditory areas of the brain, but they are all connected together by the hippocampus to form an episode, rather than remaining a collection of separate memories.

For example, memories of people’s faces, the taste of the wine, the music that was playing, etc, might all be part of the memory of a particular dinner with friends. By repeatedly reactivating or “playing back” this particular activity pattern in the various regions of the cortex, they become so strongly linked with one another that they no longer need the hippocampus to act as their link, and the memory of the music that was playing that night, for example, can act as an index entry, and may be enough to bring back the entire scene of the dinner party.

Our spatial memory in particular appears to be much more confined to the hippocampus, particularly the right hippocampus, which seems to be able to create a mental map of space, thanks to certain cells called "place cells". Episodic memory does also trigger activity in the temporal lobe, but mainly in order to ensure that these personal memories are not mistaken for real life. This difference in the neurological processing of episodic and semantic memory is illustrated by cases of anterograde amnesia cases (a good example being a case known as “C.L.”) in which episodic memory is almost completely lost while semantic memory is retained.

A further category of declarative memory, referred to as autobiographical memory, is sometimes distinguished, although really it is just one area of episodic memory. Autobiographical memory refers to a memory system consisting of episodes recollected from an individual’s own life, often based on a combination of episodic memory (personal experiences and specific objects, people and events experienced at particular times and places) and semantic memory (general knowledge and facts about the world).

One specific type of autobiographical memory is known as a "flashbulb memory", a highly detailed, exceptionally vivid “snapshot” of a moment or circumstances in which surprising and consequential (or emotionally arousing) news was heard, famous examples being the assassination of John Kennedy, the terrorist bombings on 9/11, etc. Such memories are believed by some to be highly resistant to forgetting, possibly due to the strong emotions that are typically associated with them. However, a number of studies also suggest that flashbulb memories are actually not especially accurate, despite apparently being experienced with great vividness and confidence.

RETROSPECTIVE & PROSPECTIVE MEMORY

An important alternative classification of long-term memory used by some researchers is based on the temporal direction of the memories.

Retrospective memory is where the content to be remembered (people, words, events, etc) is in the past, i.e. the recollection of past episodes. It includes semantic, episodic and autobiographical memory, and declarative memory in general, although it can be either explicit or implicit.

Prospective memory is where the content is to be remembered in the future, and may be defined as “remembering to remember” or remembering to perform an intended action. It may be either event-based or time-based, often triggered by a cue, such as going to the doctor (action) at 4pm (cue), or remembering to post a letter (action) after seeing a mailbox (cue).

Clearly, though, retrospective and prospective memory are not entirely independent entities, and certain aspects of retrospective memory are usually required for prospective memory. Thus, there have been case studies where an impaired retrospective memory has caused a definite impact on prospective memory. However, there have also been studies where patients with an impaired prospective memory had an intact retrospective memory, suggesting that to some extent the two types of memory involve separate processes.