CHAPTER 5: SHORT-TERM AND WORKING MEMORY

Working memory is like your mind’s workspace— a limited capacity system for storage and processing of information. While early researchers focused on the storage function of working memory, contemporary scientists stress the processing and manipulation functions of the system. In fact, the name for the system was changed from “short-term memory” to “working memory” to emphasize the processing function of this stage of memory.

When people talk about memory, they are describing the mind’s ability to encode, store, and retrieve information. Our ability to remember is what allows us to learn from our experiences. How does memory function? In the process of answering this question, many different models of memory have evolved. Distinctions are drawn between working memory and long-term memory based on the period of time information is accessible after it is first encountered. Sensory Memory has the smallest time span for accessibility of information. With Short Term and Working Memory, information is accessible seconds to minutes after it is first encountered. Long Term Memory has an accessibility period from minutes to years to decades. This chapter will focus on short-term and working memory— different terms for overlapping concepts.

SHORT-TERM MEMORY

In the middle of the 20th century, many scientists were interested in short-term memory (STM), or how humans can hold small amounts of information actively in their minds for a short period of time. In 1968, Richard Atkinson and Richard proposed a model of memory referred to as the Modal Model of Memory (Figure 1). In this model, information first enters sensory memory, which is a highly transient storage space for information that recently entered your sensory system. This information is high quality but fades away very quickly. Have you ever had the experience of hearing you weren’t really paying attention, then repeating it immediately in your head, and then being able to understand it? For about three seconds, you can play back the auditory information that you just heard. You can even hear it in the speaker’s original tone of voice! You use your sensory memory to do this. someone say something when

Figure 1. Atkinson and Shiffrin’s Modal Model of Memory. Image from Wikimedia Commons.

How does short-term memory work? If we look up a phone number in the phone book and hold it in mind long enough for dialing the number, it is stored in the Short Term Memory.

According to George Miller (1959), the capacity of short-term memory is five to nine pieces of information (The magical number seven, plus or minus two). That’s why I could read a 7- digit phone number to you and have you repeat it back, but if I read you my 16-digit credit card number and asked you to repeat it to me, it would probably feel impossible. If we can remember about 7 pieces of information, what counts as a “piece of information,” also referred to as a chunk? A chunk is a meaningful unit of information. All of the following can be chunks: single digits or letters, whole words, or even sentences. An example of chunking information is the following.

Try to remember the following digits:

Now try to remember the same digits, but group them differently:

With this strategy you chunked eight pieces of information (eight digits) to three pieces to remember them as a date on the calendar. You could chunk the information even more efficiently if you recognize 12-25 as a single unit, the date of Christmas. The process of chunking is the process of combining smaller units of information into larger, meaningful units of information. The term “meaningful” is subjective— a meaningful chunk for you might not be a meaningful chunk for me. For example, 4 1 3 3 might not make a meaningful chunk for everyone, but if you’re a football buff, you might chunk it as 41–33: the final score of the Super Bowl for the 2017 season.

A famous experiment concerned with chunking was conducted by Chase and Simon (1973) with novices and experts in chess playing. When asked to remember certain arrangements of chess pieces on a board, the experts performed significantly better that the novices. However, if the pieces were arranged randomly, i.e. not corresponding to possible game situations, both the experts and the novices performed equally poorly. The reason is that expert chess players spend hours studying chess games and memorizing board configurations. When trying to remember the layout of a chess board, the experienced chess players do not try to remember single positions of the figures in the correct game situation, but whole chunks of figures from their memory. In random board configurations this strategy cannot work, which shows that chunking (as done by experienced chess players) enhances the performance only in specific memory tasks.

FROM SHORT-TERM MEMORY TO BAFFELEY’S WORKING MEMORY MODEL

Although the Modal Model of short-term memory proposed by Atkinson and Shiffrin explained aspects of human memory under certain conditions, the model turned out to be limited in its explanatory power. Baddeley and Hitch (1974) drew attention to problems with the model. For example, Baddeley and Hitch emphasized that we can not only hold information actively in our minds, we can also do things with that information. Whereas the Modal Model conceived of short-term memory as a mostly passive storage space, Baddeley and Hitch emphasized our ability to process and manipulate information. They even set aside the term “short-term memory” in favor of “working memory,” in order to emphasize the processing and manipulation functions of the system. Scientists don’t use the term “short- term memory” very often now, and tend to prefer to use “working memory.” The modern understanding of the working memory system includes all of the things that early scientists called “short-term memory,” but also includes other capabilities as described in the next section.

WORKING MEMORY
According to Baddeley, working memory is not only capable of storage, but also of the manipulation of incoming information. Baddeley and Hitch’s 1974 model consists of three parts: two storages spaces called the phonological loop and the visuospatial sketch pad, and a control unit call the central executive.

We will consider each part in turn: The Phonological Loop is responsible for auditory and verbal information, such as phone numbers, people’s names, or general conversation. One source of evidence that we have a special storage space for auditory information is the phonological similarity effect. Read the following list of words, then look away and try to repeat it to yourself:

And now try this one:

If you are like the participants in Conrad’s 1964 study, your performance was worse on the second list than it was in the first. The reason the second list was harder is because when you read the words on the page, you translate them into an acoustic form. Because the words in the second list sound alike, you are more likely to confuse them as you repeat them in your phonological loop.

How much information can the phonological loop hold? Researchers have found that the magical number seven plus or minus two does not explain all of the available data. While Miller’s magical number is approximately accurate when English-speaking participants remember digits or letters, it doesn’t hold when the length of the words is manipulated.

To demonstrate this to yourself, try to remember the following list of words:

And now try this one:

Both lists are seven words long, and yet people are much worse at a list like the second one (Baddeley, Thomson, & Buchanan, 1975). This is called the word-length effect: lists of short words are recalled better than lists of long words. If you think back to the research on short- term memory, this result is surprising! According to Miller’s magical number, these lists should be remembered equally well because they contain the same number of items. Findings like the word length effect led researchers to conclude that the capacity of the phonological loop should not be measured in number of items, but in amount of time instead: the phonological loop can hold about two seconds of auditory information, which it can replay over and again through an active articulatory process. Imagine you have two seconds of tape — you could fit a lot more short words on it than long words!

The next component of working memory is the visuospatial sketch pad, which handles visual and spatial information. Like the phonological loop, the visuospatial sketchpad is primarily a storage space. What evidence do we have that visual and spatial information is stored independently from auditory and verbal information? Look at the block letter F to the right. In an example experiment, you would be instructed to memorize the letter, and then, starting with the starred corner and then traveling up and around the letter, indicate whether each corner is an outside corner (like the starred corner) or an inside corner (like the fifth corner as you travel around the shape). In one condition, participants would verbally say “outside” or “inside” for each corner. In the other condition, participants would point at the words “outside” and “inside” displayed in front of them for each corner. Brooks (1968) conducted an experiment similar to this one, and found that participants were much better at the task when they could verbally indicate the type of corner than when they had to point. Why? We have two storage spaces in working memory and each of them is limited in capacity. Mentally traveling around the block letter and judging the corners is a spatial task, and so it puts a load on the visuospatial sketchpad. If you add pointing on top of that, participants’ visuospatial sketchpads get overloaded and they struggle to do both simultaneously. If instead you allow participants to respond verbally, they are distributing the response aspect of the task onto the phonological loop. That way, neither storage system becomes overloaded.

We have seen that the phonological loop and the visuospatial sketch pad deal with different kinds of information, which nonetheless have to interact in order to do certain tasks. The component that connects these two systems is the central executive. The central executive coordinates the activity of both the phonological loop and the visuospatial sketch pad. In fact, most of the “working” part of working memory is done by the central executive.

The functions of the central executive can be broken down into three categories: shifting, updating, and inhibition (Miyake et al., 2000). Shifting refers to engaging and disengaging from tasks, such as switching your attention back and forth between watching television and doing the dishes. Updating refers to monitoring information that is incoming into working memory, and making room for it by replacing old information in working memory. Inhibition refers to the deliberate inhibition of responses, such as when the ticket taker says “enjoy your movie!” and you stop yourself from saying, “you too!”

THE EPISODIC BUFFER

Science is an ongoing process, and so despite the usefulness of Baddeley and Hitch’s working memory model, it was updated in 2000 to add another component: the episodic buffer (Baddeley, 2000). The episodic buffer is a limited capacity, temporary storage system that is controlled by the central executive and integrates information from a variety of sources including long-term memory. The episodic buffer was added to help account for human performance in complex cognition— for example, we can remember many more words when we remember a meaningful sentence than when we remember a random word list. The episodic buffer was added to help account for this and other phenomena in which working memory performance seemingly requires additional storage as well as interfacing with long- term memory.

CHAPTER 5 LICENSE & ATTRIBUTION

Source: Multiple authors. Memory. In Cognitive Psychology and Cognitive Neuroscience. Wikibooks. Retrieved from https://en.wikibooks.org/wiki/ Cognitive_Psychology_and_Cognitive_Neuroscience

Wikibooks are licensed under the Creative Commons Attribution-ShareAlike License.

Cognitive Psychology and Cognitive Neuroscience is licensed under the GNU Free Documentation License.

Condensed from original version. American spellings used. Content added or changed to reflect American perspective and references. Context and transitions added throughout. Substantially edited, adapted, and (in some parts) rewritten for clarity and course relevance. Chapter introduction added. Content added including transition from STM to WM approach, description of episodic buffer, description and evidence for working memory components, addition of episodic buffer.

Cover photo by Matt Briney on Unsplash.