Cognitive Psychology:
Attention, Recall, & Knowledge Representation

By: Dr. Susan Siegfried, Clinical Psychology

In this article we will take a brief look at how attention and knowledge representation work in the cognitive process to aid recall.

A major element in cognition is attention. Researchers have found that attention is a very limited resource. It is thought of as being single minded. That is, there is only enough room for one task at a time. Think of it in terms of energy. You have only a fixed amount so if you attention were run on a fixed amount of energy it would have to stay within the confines of that energy source. While attention may be thought of as single minded but it could handle two small tasks at the same time or one large task on that limited amount of energy. You practice tasks until they become automatic, for example walking or talking. Once those things are automatic you can use that limited energy source for attention on other things. All information gets into sensory memory but only that which is attended to will remain in long term storage.

If you were to hear two things at one time (one in each ear) and have to remember only what was coming in one ear most people are able to attend to the important message and tune the other out. Psychologists have discovered that very little of what is heard in the other ear is processed at all. We have all experienced this, think of when you are at a party and you tune into one conversation. You become unaware of any other conversations going on even though there are many people in the room talking. Messages from both ears get into your sensory memory but only one set of sensory coding is processed. Only that which is attended to is processed so only that which is attended to is stored in memory.

Recall

The more processes we can practice the more space, or energy to use the above analogy, we have in our attention to use on tasks. It is believed that if practiced enough a task does not need any attention space. So you can walk and chew gum at the same time. The way this automatically happens is that we learn patterns so no matter in what format this writer presents the letter A to you your brain will recognize it because you have practiced reading most of your life. You match your learned pattern to the new A and you can still read it without thinking about it. You recognize the pattern so because of this recognition you can think about the concept of pattern recognition in recall.

Another theory of automatic recall is feature analysis. This theory says you must do more than matching. In this model, stimuli are thought of a combination of element features. So it considers angles, curves and vertical and horizontal lines. Features are similar to templates analysis, however, since features are similar to matching whole structures it reduces error and looking at combinations of things allows for specification of relationships.

These two ways of recalling of practiced items work for speech in that we match phonemes as you match structures. You match tastes and smell as you match structures. So using this principle we can match complicated gestalts. A gestalt is the grouping of items into a whole. So with our letter A we rarely see it alone it is usually in a word such as Apple or similar or taste. For us to understand the A in different places our brain segments the letters and recognizes the whole and the individual patterns. If you have ever shared a meal with a trained chef you know that that person has so practiced taste that they can tell you every ingredient in a complicated dish and come very close the exact amount of each ingredient used. That is the Gestalt Principle of Organization. This may sound simple but to see how this actually works in the brain it does not seem so simple. Below you see a diagram of how researchers think this looks.

Depiction of the Gestalt Principle of Organization (Anderson 2009)

Subjects are able to apply context into a visual scenes, speech, smell, taste, sound and touch. We have looked at attention, sensory memory, stored memory (practice) and recall in pieces but remember cognition needs to be/should be like a computer program, that is, it should be precise specification of the behavior, be offered in terms sufficiently abstract to provide a conceptually trackable framework for understanding the phenomenon. Put into information processing format the cognition process would look like the diagram below.

The Cognition Process and the future of human evolution

Click to Enlarge

Depiction of the Cognition Process (Anderson 2009)

Perception-Based Knowledge Representations

The storage of perception-based knowledge representations appear to be processed in the neural regions close to where the original perceptions were experienced. This would be in the parts of the body where your five senses are located. There are two types of perception based knowledge representations. They are images that encode the spatial structure of items and linear orderings that encode the sequence.  When asked to perform a mental transformation of an image such as rotating something 180 degrees, people imagine the image turning. When comparing two mental objects, people engage in a process similar to that of comparing two physically present objects. There appear to be two different types of mental imagery. One preserves visual detail and the other not tied to the visual modality, encodes spatial relationships. Both spatial images and linear orderings have hierarchical organization in which sub images or sub lists can occur as elements in larger images or lists. People have more rapid access to the first and last elements of a linear ordering. They also tend to search linear orderings from beginning to end. They can more rapidly judge the order of elements in a linear ordering and  how far the elements are apart.

Propositional Representation and Knowledge Based Storage

It is well known that meaningful information is easier to remember than nonsense information. So knowledge based information is more likely to be stored. Representations that do not preserve the exact perceptual structures are what go into long term memory. The meaning of pictures and linguistic communication are encoded. Propositional analysis only stores the meaning of an event. The unimportant details that we tend not to remember are not stored. Memories (propositions) are layered into propositional networks for faster and easier retrieval. The sentence: Obama gave a beautiful Cadillac to Kim Jong-un, who is the leader of North Korea. would be stored for recall as seen below.

Click to Enlarge

Depicted above: Propositional Representation and Knowledge Based Storage (Anderson, 2009)

Schemas

Propositions are a useful way to represent meaning that is more abstract. A schema is a collection of memory representations that are practiced and then applied to a new situation to make sense of the new situation.  An example is that we all have a schema of what we expect a house to be like.  When we are in a strange place we match what we see to our schema of an office, a store or a house. The one that matches best informs us of where we are. The irony of this is that our schema informs us of what we remember about the new house. So later people are found to remember the things they expected to see in a house from their schema even though they were not in the new setting.  They also remembered what they expected to see from their schema but did not remember what was not in their schema of a house that was in the new house but not what they did not expect to see. Now we understand why memories are not perfect and no two people ever remember exactly the same thing.

A natural inference from this chapter is that the units that make up schemas are propositions, however, this is not always the case. For instance, some of the components that make up a house include spatial information about object shape as well as propositional information about function of a house. Similarly, the major part of an event schema is information about the order of sub events. Spatial information is encoded by images and order information by linear orderings Thus, we see that schemas encode and the co-occurrence relations amount propositions, spatial images and linear orderings.

Sources

Anderson, J.R. (2009). Cognitive psychology and its implications. New York, NY: Worth Publishers.

Eysenck, M.W. (1990). Cognitive psychology: An international review. West Sussex, England: John Wiley & Sons, Ltd.

Chomsky, N. A. (1959), A Review of Skinner’s Verbal Behavior. New York, Meyer

Neisser, U. (1967). Cognitive psychology. Englewood Cliffs, NJ: Prentice Hall.