Chapter Outline

Chapter Outline

1. MEASURES OF MEMORY

   1. There are two explicit measures of memory: recall and recognition. Recall requires the subject to generate the details of a memory, similar to an essay exam. Free recall involves generating the details without any specific directions or cues such as the request to name the fifty United States. In cued recall, information is presented to assist in the generation of the memory; instead of simply requesting that all fifty states be named provide the first letter of each state. It should be noted that in any exam a cue is presented to assist in memory retrieval, this cue is the question itself. Recognition measures the ability to identify information that had been experienced at an earlier time. A multiple-choice exam is analogous to recognition as a measure of memory. The student is presented with a question and asked to recognize the correct answer.

   2. There are at least two implicit measures of memory: savings and reaction time. Created by Ebbinghaus (1885), savings refers to the fact that fewer trials (savings) are required to "re-learn" material, implying the existence of some residual memory from original learning. When required to retrieve stored information, the amount of time necessary (reaction time) to adequately access the information varies with the accessibility of the memory; faster reaction times are associated with stronger memory traces.

2. A THREE-STAGE VIEW OF MEMORY

   According to Atkinson and Shiffrin (1971) memory consists of three interrelated systems: sensory store (also known as the sensory register), short-term store, and long-term store.

   1. The sensory store holds external input for a brief period of time, usually 500 msec up to 1-2 sec. This information is an initial impression of the external stimulus and decays rapidly unless it undergoes further processing.

   2. The short-term store is the second stage of information processing and is a temporary, limited-capacity facility for our experiences. The length of time information remains in the short-term store depends on whether it undergoes rehearsal or not. Without rehearsal information is lost after approximately 15-30 sec. New incoming information may displace older information if the capacity of the short-term store is exceeded. An important role played by this system is the organization of material for meaningful storage and subsequent retrieval.

   3. The long-term store is presumed to be an unlimited, relatively permanent storage system. Information stored in this system may or may not be easily accessed. The ease with which the material is accessed correctly depends on the degree of interference from other memories and the presence of memory attributes.

3. SENSORY REGISTER

   Research on the sensory stores has focused on two sensory systems: the visual system and the auditory system. The visual representation in sensory memory is termed an icon and visual sensory memory is known as iconic memory. The auditory sensory experience is called an echo and auditory sensory memory is known as echoic memory.

   1. Iconic Memory

      1. In Search of an Icon

         Sperling (1960, 1963) has examined the nature of iconic memory. He visually presented to subjects a matrix of letters for a brief period of time and then asked for recall of the matrix. This procedure, known as the whole-report technique, indicated that subjects reported on average 4.5 items from the matrices that had as many as 12 items originally. Sperling conducted another version of the experiment, using the partial-report technique. In this version, the subjects are asked to recall information presented in one of three rows. A tone given at the end of the retention interval indicated which row had to be recalled. Each row contained four items and Sperling discovered that the subjects could remember most of the letters. However, if the retention interval was greater than 0.25 sec, subject performance declined to the level that only one item could be remembered in each cued row. Sperling's results suggest that the stored icon is subject to a rapid decay process.

      2. Duration of an Icon

         Sperling's (1960) original research suggested that icons last for 0.25 sec. However, as the intensity of the visual stimulus increases so does the duration of the icon, at least up to 1 sec or more (if the display is very dark following presentation of the letters). In addition, if the display is very bright before or after matrix presentation the icon may last less than 0.25 sec.

   2. Echoic Memory

      1. In Search of an Echo

         Neisser (1967) called the sensory memory of an auditory stimulus an echo. Moray, Bates, and Barnett (1965) conducted research indicating that the properties of echoic memory are similar to those for iconic memory.

      2. Duration of an Echo

         The duration of an echo is longer than an icon, extending to 2 sec. Research by Crowder and Morton used the serial position effect to demonstrate the longer duration of echoic memory.

   3. The Nature of Sensory Memory

      Reseachers initially thought that a sensory memory was an exact replica of an event; that is, an icon is an exact copy of a visual stimulus and an echo is an exact duplicate of a sound (Posner, 1982). More recent evidence reveals that some processing occurs in the sensory store, in particular some form of categorization. Although auditory and visual sensory memories are important, it should be noted that sensory memories exist for all the senses.

4. SHORT-TERM STORE

   The short-term store receives information from the sensory stores but unlike the sensory registers, the short-term store actively processes the information. The short-term store holds information by rehearsal and its function is to organize and analyze information. It has a limited capacity and is susceptible to disruption when rehearsal is interrupted. The function of rehearsal is to increase the recall of information by allowing the information to enter the long-term store.

   1. The Span of Short-Term Memory

      Peterson and Peterson's (1959) study suggested that short-term memory rapidly loses information when it is prevented from being rehearsed.

   2. Disrupting Short-Term Memory

      The short-term store has limited capacity but is required to constantly process incoming information. Therefore, it is susceptible to disruption from external and internal sources. Unless the information is stored in a meaningful way in the long-term store it will be lost forever.

   3. Limited Storage Capacity

      Miller (1956) proposed that the short-term store holds on average 7 items of information plus or minus 2 (5 to 9 items). Watkins (1974) proposed that the actual capacity of the short-term store is about 3-4 items based on the idea that some of the items represent information retrieved from the long-term store.

   4. The Organization of Our Experiences

      One type of organization that the short-term store accomplishes is chunking. A second type is coding and a third type is the formation of associations.

      1. Chunking

         Chunking involves combining individual elements into meaningful units of information. It is an automatic organizational process of the short-term store and significantly enhances our ability to recall past experiences. Chunking also causes the information to be recalled from the long-term store in a specific order based on categories. The recall of information in terms of categories is called clustering.

      2. Coding

         The coding of information involves transforming information into a totally new form. Visual experiences are also coded in terms of an acoustic code presumably to enhance the thinking process. Information is also coded as a visual code. Some individuals are able to generate a very distinct visual image of a past event, a visual memory known as an eidetic image. A vivid, detailed, and long-lasting memory for an emotionally arousing event is called a flashbulb memory.

      3. Association of Events

         Collins and Quillian suggest that semantic memory consists of interconnected concepts and properties associated with each concept node. This perspective is known as the hierarchical approach. Spreading activation theory assumes that a property may be associated with more than one concept but will be more closely associated with some concepts. Once a concept is activated, the activation spreads to associated concepts or properties, a perspective that explains the effect of priming. The notion that knowledge is not localized but rather is distributed throughout the entire memory system is called the parallel processing model.

   5. The Rehearsal Function of the Short-Term Store

      Rehearsal serves two functions (Eich, 1985). First, it retains information in the short-term store so it is not forgotten and second, it provides the opportunity to make the information more meaningful. The first function corresponds to maintenance rehearsal which involves simply repeating the information until it is needed. The second function is accomplished by elaborative rehearsal which involves embellishing the material to create a more meaningful memory.

5. ALTERNATIVES TO THE ATKINSON-SHIFFRIN MULTI-STAGE MODEL

   The Atkinson-Shiffrin model assumes that while in the short-term store, the memory is rehearsed, which results in organization and an increase in the likelihood it will be retrieved later. Two alternatives to this position have been proposed. One, the rehearsal systems approach is a modification of the Atkinson-Shiffrin model of the short-term store and the second, levels of processing dispenses with the stage notion altogether.

   1. Rehearsal Systems Approach

      The rehearsal systems approach, developed by Baddeley (1990) argues that while memories are retained in the sensory systems, those memories are acted on by a process he called working memory. Working memory retains all the characteristics of the short-term store but also contains several rehearsal systems: the phonological loop for auditory information, the visuospatial sketch pad for imaging, and a central executive to coordinate the activities of the rehearsal systems.

   2. Levels of Processing Approach

      The levels of processing view posed by Craik and Lockhart (1972) assumes that a memory can be processed at different levels. Information can receive only a shallow, superficial analysis or it can be interpreted at a deeper, semantic level. The deeper or more elaborate the information processing, the more likely it is that it will be recalled later.

6. LONG-TERM STORE

   The long-term store preserves information on a relatively permanent basis. There are several different types of long-term memories including episodic, semantic, declarative, and procedural. The latter two are posited by those investigating the physiological bases of memory formation and forgetting.

   1. Episodic vs. Semantic Memories

      Tulving suggested a distinction between episodic and semantic long-term memories. An episodic memory consists of information about events and their associated time and place. Semantic memory contains knowledge associated with the use of language (meanings expressed through the use of symbols).

   2. Types of Information

      1. Separate Operations

         The episodic memory system holds sensory experiences; however, the semantic memory system records knowledge conveyed by language. The temporal order of events can be detected by episodic memory in contrast to semantic memory which must infer temporal sequencing. Recollection from episodic memory is deliberate and often requires conscious effort whereas recall from semantic memory is often unconscious and automatic. Thus, humans remember episodic information but know semantic information.

      2. Two Functionally Different Memory Systems?

         Although rCBF studies support Tulving's notion of separate long-term systems, several researchers disagree and posit a single memory system with content that can vary from highly context-specific episodes to abstract generalizations.

   3. Procedural versus Declarative Memories

      Squire has proposed the term procedural memory for knowledge and recall of motor skills. Procedural memories are acquired slowly through repeated experience and we tend to be unaware of them. In contrast, declarative memory represents factual memory which is conscious and can exist as a verbal thought or non-verbal image.

   4. In Search of the Engram

      The engram represents the physiological correlate of long-term memory. An engram may be the result of cellular modification, a change in neural responsiveness due to structural changes at the synapse, chemical enhancement, or the establishment of new connections.

      1. Learning in the Aplysia californica

         Kandel and his associates (1991, 1993) have conducted extensive research indicating that modification of neurotransmitter function is important for memory formation. Kandel's research involves the investigation of habituation, sensitization, and classical conditioning of the gill reflex in the sea mollusk Aplysia californica.

      2. Structural Changes and Experience

         Lynch has proposed that structural modifications of receptor sites on dendritic spines underlies memory formation. Presumably, neural activity allows an enzyme, calpain to expose more receptor surfaces for neurotransmitter activity. Calpain's breakdown of the cellular coating allows dendrites to change shape and spread out resulting in new neural connections.

   5. Anatomical Basis of Memory Formation

      At least two brain structures - the medial temporal lobe and the mediodorsal thalamus appear to be important areas for the storage of information. Information is initially processed in the cortex and is then sent down to the medial temporal lobe and associated structures (e.g., hippocampus) which project to the mediodorsal thalamus.

      1. Medial Temporal Lobe

         The case of H.M. reveals the importance of an intact medial temporal lobe including the hippocampus. To control epilepsy a bilateral temporal lobectomy was performed which also removed the hippocampi. Following the operation, H.M. developed anterograde amnesia, the inability to form new memories. This amnesia affected new declarative memories but not new procedural memories. The specific cause of H.M.'s amnesia appears to be due to the bilateral removal of the hippocampus. Other clinical cases support this view.

      2. Mediodorsal Thalamus

         Damage to or the destruction of the mediodorsal thalamus also results in anterograde amnesia for declarative memories but not as severe as that produced by destruction of the hippocampus. The memory deficit called Korsakoff's syndrome is also associated with damage to the mediodorsal thalamus.

   6. Mnemonics

      People with exceptional memories often use mnemonic techniques to enable themselves to recall impressive amounts of recently learned information. Two commonly used mnemonics are the method of loci and the pegword method. In the method of loci the "to-be-remembered" information is associated with familiar locations using visual imagery. The pegword method is effective for remembering the items on a list; each item is associated with an already memorized "pegword."