Models of Memory

The importance of memory skills in academic learning cannot be overestimated (Liddell & Rasmussen, 2005). Research has linked memory deficits among children who are learning disabled with reading problems, language problems, difficulty in spelling, and other areas (Bender, 2002; Swanson, 1994). Finally, our own school experiences tell us that memory skills are used in many required tasks in the classroom.

Historically, memory has been differentiated into two levels: short-term memory and long-term memory (Swanson, 1994). Short-term memory represents storage of a limited amount of information (six to eight bits) for a limited amount of time (usually less than 15 seconds). Long-term memory has been defined as memory of a longer duration.

More recently, the term working memory has been used to describe a refinement and extension of short-term memory skills (Sprenger, 2002; Sousa, 2005, 2006; Swanson, 1994). Working memory represents the ability of a student to hold a small amount of information in short-term memory while working with that information and integrating it with other information. Swanson (1994) compared students with LD and students without on a number of short-term memory and working memory tasks and indicated that working memory was more influential in reading skill than short-term memory for both groups. In other words, short-term retention of isolated facts is less important than the skill of short-term retention in combination with the need to integrate that information with previous knowledge.

Memory has also been conceptualized as including three relatively distinct processes. Encoding refers to translating a sensory input into some representational form for storage. When a student picks one key word to help remember an important phrase, that is an encoding process. Storage refers to the durability of the memory, and retrieval refers to the process of recovering an encoded representation of a stimulus from memory (Torgesen, 1984).

Sousa (2006) presented a more recent model of memory processing that captures the multiple aspects of memory. This model represents our most recent understanding of how memory works in human beings. As indicated, information enters the brain from the environment through the senses (five arrows on left). This information immediately passes through the sensory register, represented as the side view of a venetian blind. If, based on past experience, the information seems important, the stimuli will pass through the sensory register. Alternatively, if the stimuli are deemed unimportant, they will be ignored and not stored in memory. Sousa (2006) postulated that all activities of the sensory register are unconscious, suggesting that many stimuli that are temporarily registered in the brain are eventually eliminated without any conscious thought.

If the stimuli pass through the sensory register, this will be noted for a brief period of time (usually 3 to 7 seconds) in short-term memory (Sousa, 2006). Short-term memory is represented by a clipboard and serves as merely a continuation of storage beyond the sensory register. In Sousa's model, short-term memory is an unconscious process.

Working memory, in contrast, is a conscious process in which a person considers a stimulus (or new knowledge) in terms of attributing sense and meaning to it, based on previous knowledge (Sousa, 2006). For example, in the classroom in order for something to be learned (i.e., stored in long-term memory), the new knowledge must make sense to the student. This knowledge must be understandable in the terms and context the student has already mastered. Further, the new knowledge must have meaning. It must in some sense answer a question for the student or fill a gap in understanding that the student recognizes; the student must want to learn it. The worktable in the model represents the individual's conscious efforts to place the stimuli within the context of something that is already known.

In one sense, this working memory construct represents the constructivist perspective (see Chapter 1) in a nutshell, since all stimuli to be learned must be based on meaningful relationships that are constructed between the new knowledge and other knowledge that was learned previously (Sousa, 2006). If sense and meaning can be associated with the new knowledge, the likelihood of long-term memory storage is great. Otherwise, the new knowledge falls out of the system. Thus, the attachment of sense and meaning is the critical aspect of learning. In Sousa's model (2006), both encoding and retrieval would be functions of working memory. Instructional suggestions based on this emerging brain-based memory research deal predominantly with assisting a student to attach sense and meaning to new knowledge.

Note that long-term storage is represented in the model by filing cabinets (Sousa, 2006). This suggests a regulated system by which knowledge is associated with other relevant knowledge and is immediately available for retrieval. Also, long-term memory storage is part of the individual's cognitive system, self-concept, and past experiences and impacts future learning, as represented by the arrows from long-term storage to the sensory register and short-term memory.

Another conceptualization of memory involves memory of specific types of information, and these are referred to as explicit memory and implicit memory (Sprenger, 2003; Sousa, 2006). For example, explicit memory is memory of details, facts, and events (e.g., memory of one's name or locations of friends' homes). Explicit memory is the memory system teachers attempt to impact when helping students make connections between a known fact or principle and a new fact that has yet to be mastered. This memory system involves processing in the hippocampus and cerebrum (Sousa, 2006). Explicit memory may be further divided into two types. First, semantic memory involves general factual memory and memory connections (e.g., knowing where the state of Oregon is on the map or how to read a clock). Because semantic memory is tied to connections between facts, strategies such as mnemonic instruction function well to enhance semantic memory. Second, episodic memory involves memory that may be based on location and circumstance. For example, when asked to recall what you did at 1:00 p.m. yesterday, you first try to recall where you were and then what additional actions you engaged in (Sprenger, 2003).

Implicit memory describes memory for nonfactual information (e.g., how to hit a baseball) and is processed in various areas of the brain, depending upon the type of memory. That is, the implicit memory of how to hit a baseball may be processed by the motor cortex and cerebellum, whereas other types of implicit memory may be processed elsewhere. Researchers today further catagorize implicit memory into various areas, as noted below (Sousa, 2005).

Procedural memory—involves learning of motor skills (e.g., hitting a ball or driving a car)

Perceptual register memory—involves the structure and form of words and objects that may be prompted by prior experience; also involves our ability to complete fragments of words

Associative learning—involves memories that result from Pavlovian conditioning (e.g., when a conditioned stimulus and an unconditioned stimulus are paired together)

Nonassociative learning—involves the brain's tendency to briefly remember and then "screen out" information (e.g., noises of the city)