CONCEPT I: Sensory Memory

Psychologists usually distinguish three types of memory: sensory memory, short-term memory, and long-term memory.

Sensory memory is a very brief (perhaps only a second or two) and temporary store for information coming in through sensory channels. We have a sensory memory to correspond to each of our five senses. George Sperling's experiments demonstrated that sensory memory may be able to process more information than was previously thought. It is during this brief period of sensory memory that we select the information that will warrant further processing in short-term memory. Some psychologists believe that difficulty in reading may be linked to poor sensory memory. Some children have an unusually good ability to remember vivid visual images, called eidetic images. This ability to form eidetic images usually declines with age.

CONCEPT II: Short-Term or Working Memory

Selective attention allows us to select and attend to only certain information from any single sensory channel. It has been examined by using a dichotic listening technique, in which subjects hear two different messages at the same time, one in each ear. If one of the messages is "shadowed" (repeated aloud by the subject), only the physical characteristics (such as pitch) of the messages that are not shadowed are remembered. This is called selective attention.

Encoding is the act of converting sensory stimuli into a form that can be placed in memory. The method we use in encoding is called an encoding strategy. Encoding often draws upon previously stored information.

Automatic encoding occurs without intentional effort and is distinguished from effortful encoding, or deliberate memorizing.

Short-term memory has a limited duration (about twenty seconds), but information can be maintained in short-term memory through rehearsal, or repetition. Short-term memory rehearsal is believed to be acoustical—that is, the sounds of the words are repeated and stored. Our short-term memory probably holds "seven plus or minus two" (that is, five to nine) pieces of information at any one time, according to George Miller and most other psychologists. Consequently, if we can "chunk" information—that is, organize it into a smaller number of units—we can increase the capacity of short-term memory.

CONCEPT III: Long-Term Memory

Long-term memory is the permanent, unlimited-capacity part of our memory; it is the repository of all our accumulated information. Three kinds of information are stored in long-term memory: Associations between stimuli and responses are called procedural memories; conceptual knowledge and mental representations of objects and qualities are called semantic memories; and personally experienced events are called episodic memories.

Although we do not understand completely how information enters long-term memory, we do know that storage depends on both the amount and the kind of rehearsal. Shallow processing, or maintenance rehearsal, is more likely to reinforce short-term memory. Deep processing, or elaborative rehearsal, is more likely to produce long-term storage. When elaboration stresses some meaning, memory is enhanced.

Retrieval from long-term memory can be in the form of recognition (you are presented with a stimulus and asked if you have encountered it before) or recall (you are asked to remember an event). Recognition may explain the déjà vu phenomenon, which is a feeling of having experienced something before although you know you could not have. Recognition is generally easier than recall, since it involves only a matching rather than a memory search process. Retrieval cues (hints) are helpful in recall. When memories are stored, we typically create a number of retrieval pathways we can use to access the memory. This fact is demonstrated by the tip-of-the-tongue phenomenon, where subjects feel they know a correct response, but can't quite access it.

Memory can be improved by using memory aids, called mnemonic devices. One such technique is the method of loci, in which items are visually associated with a series of well-known places. A peg word system involves memorizing a list of simple words or sentences and then visually associating each item of a sequential list with the ordered key words. Recall in both cases involves going back through the place or key words and stating the visually associated objects. Both techniques make use of organization and imagery.

Another memory aid is reconstructing the context in which information was learned, thereby establishing cues to help retrieve the memory. In this way, we find that mood can affect memory. Mood-dependent memory, for example, is a condition in which recall is enhanced if the person is in the same psychological state as when the memory was originally stored. Mood-congruent recall, on the other hand, refers to the tendency to recall information that matches the current mood.

Mood-congruent recall can be understood as an example of a network model of memory, which specifies that our memories are linked together in a complex network of pathways. Network models fit into the connectionist framework, which emphasizes the connections between stimuli and responses.

Reconstruction may lead to memory distortion if gaps in memory are filled in with what seems to be true rather than what actually occurred. Distortions are often consistent with our expectations, or schemas, and inconsistent information may simply be forgotten. Also, when we are set to perceive a situation in a particular way, we tend to recall the event as we were ready to perceive it, rather than as it actually occurred.

Memory distortions have obvious impact on eyewitness testimony. Indeed, Elizabeth Loftus has demonstrated that eyewitness recall can be influenced by the kinds of questions the eyewitness is asked. Distortions of events and misidentification of faces may occur as a result of subsequent information which is added after the original memory was encoded. Loftus first proposed that the original memories were changed by the subsequent information. Other researchers have suggested that subjects' distortions resulted not from changed memories, but rather from guessing about details they had no memory of. This misinformation acceptance comes from their willingness to add to their original memories details which were suggested later.

CONCEPT IV: The Physiology of Memory

Many neurophysiologists believe short- to long-term memory involves temporary circulation of electrical impulses through interconnected neurons. This view is consistent with retrograde amnesia, in which the victim cannot remember events that occurred just prior to a head injury.

Apparently, the transfer from short-term to long-term memory, or consolidation, takes time, as demonstrated by an experiment in which rats shocked up to an hour after a learning trial appeared to have no memory of their learning. Many scientists also believe that long-term memory involves some sort of permanent synaptic change—structural, chemical, or both. Brief, high-frequency stimulation can produce long-lasting change in a neuron's communication across synapses, apparently enhancing the neural circuits. One view of how this might happen suggests that a sudden inflow of calcium changes the neural membrane and makes it more sensitive to neurotransmitters. This and other findings have been understood to indicate that a physiological change does accompany learning.

Neuroscientists now believe that the hippocampus, a part of the brain's limbic system, may be involved in memory. Anterograde amnesia—a loss of memory for new events and information—provides evidence that semantic and episodic memories involve brain circuitry different from procedural memory. For example, H. M., who suffers anterograde amnesia, cannot form new semantic or new episodic memories, but apparently does form new procedural memories. Other parts of the brain also are involved in memory functions, and complex information is probably stored in more than one brain location, oftentimes being in the same areas of the brain involved in originally learning the data.

CONCEPT V: Forgetting

Psychologists have suggested three causes of forgetting: decay, interference, and motivated forgetting. Decay theory states that as time passes, memories that are not rehearsed simply fade away. Interference theory postulates that forgetting occurs because new events become confused with old memories. Two examples are retroactive interference, in which a new event prevents us from remembering an old one, and proactive interference, in which an already learned event interferes with our ability to learn a new one. Motivated forgetting, or repression, involves forgetting anxiety-arousing or unpleasant events.

Most psychologists believe that long-term memories are never really forgotten; they are just inaccessible at times. This notion is consistent with theories of interference and motivated forgetting. It is demonstrated by brain stimulation in which subjects report long-forgotten memories and by improved recall under hypnosis and psychoanalysis. One problem is that this research involves the possibility of confabulation—creating new information to fill in the lost details.

Some research suggests that long-term memories are permanently stored. Other research, however, indicates that certain memories may be entirely lost. Loftus, for instance, demonstrated that once subjects have incorporated false information into a memory, they cling to that misinformation and appear to forget what actually happened. In fact, forgetting is a valuable process; if we never forgot anything (like Luria's famous subject), our memories would be very confusing and unmanageable.