Human Memory
It is good to have an end to journey towards; but it is the journey that matters, in the end.
Ursula K. Le Gui
http://www.wnyc.org/shows/radiolab/episodes/2007/06/08
*
Stage theory of memory
Assumes humans have 3-stage
Memory
Process by which information is:
Recording
Encoding
Stored in the brain
Storage
Later retrieved
Retrieval
Eventually (possibly) forgotten
*
Our memory is the process by which information is retained for later use. The basic process by which information is processed follows this format: information is acquired and encoded, which leads to storage in the brain, which leads to the possibility of later retrieval (though as you know at test time, is not a guarantee), and the possibility of eventually forgetting the information.
Today, cognitive psychologists like to compare the human mind to a computer and memory to an information-processing system. I think you can appreciate the analogy. Your PC acquires (or receives) input from a keyboard or a mouse; it converts the symbols into a special numeric code; it saves the information on a hard drive, CD, or disk; it then retrieves the data from the disk to be displayed on a screen or sends it to a printer. If the computer crashes, if there’s not enough space on the disk, if the file was deleted, or if you enter the wrong retrieval command, the information becomes inaccessible, or ‘forgotten’.
Three-System Approach to Memory
*
Three-System Approach to Memory
Three types of memory
Sensory memory
Only an instant
Short-term memory (STM)
15-25 seconds
Long-term memory (LTM)
Can hold vast quantities of information and relatively permanent
*
Using the computer as a model, memory researchers seek to trace the flow of information as it is mental processed. In this Three-System Approach to Memory, a stimulus that registers on our senses can be remembered only if it 1. Draws attention, which brings it into consciousness; 2. Is encoded, or transferred to storage sites in the brain, and 3. Is retrieved for use at a later time.
Within this information-processing memory approach, three types of memory have been distinguished: sensory, short-term and long-term. Sensory memory stores all stimuli that register on the senses, holding literal copies for a brief moment ranging from a fraction of a second to four seconds usually less. Sensations that do not draw attention tend to vanish, but those we ‘notice’ are transferred to short-term memory , another temporary storage system that can hold seven or so items of information for about 20 seconds, less than 1 minute. Although STM fades quickly, information can be held for a longer period of time through repetition and rehearsa or chunkingl. When people talk about attention span, they are referring to short-term memory.
Finally, long-term memory is a somewhat permanent storage system that can hold vast quantities of information for many years. Science writer Isaac Asimov once estimated that LTM takes in a quadrillion separate bits of information in the course of a lifetime. Mathematician John Griffith estimated that, from birth to death, the average person stores five hundred times more information than the Encyclopedia Britannica. When people talk about memory, long-term memory is typically what they have in mind.
We’ll talk about each of these in a little more detail later on.
Information-Processing Model of Memory
Short-term
memory
Stimulus
Sensory
memory
Long-term
memory
Attention
Encoding
Retrieval
Forgetting
Forgetting
Forgetting
*
Many events register in sensory memory. Those that are noticed are briefly stored in short-term memory; those that are encoded are transferred to a more permanent facility. As shown forgetting may be caused by failures of attention, encoding, or retrieval.
Note, however, that this is only a model and does NOT mean that the brain has three separate storage bins. This is only one view of how memory works. There is a radically different view. Most computers process instructions in fixed sequence, one linear step at a time. In contrast, the human brain performs multiple operations simultaneously, ‘in parallel’. Thus, some cognitive psychologists have rejected the information-processing model in favor of parallel-processing models in which knowledge is represented in a web-like network of connections among thousands of interacting ‘processing units’ all active at once.
The two main questions we’ll be asking ourselves throughout this chapter are: How are memories stored? And to what extend are our memories of the past faithful to reality?
Sensory Memory
Two types
Iconic memory
Visual
Lasts less than a sec
Sperling’s tests (1960s)
Echoic memory
Auditory
Fades within 2-3 sec
*
Take a flashlight into a dark room, turn it on, shine it on a wall, and wave it quickly in a circular motion. What do you see? If you twirl it fast enough, the light will appear to leave a glowing trail, and you’ll see a continuous circle. The reason: Even though the light illuminates only one point in the circle at a time, your visual system stores a ‘snapshot’ of watch point as you watch the next point. The visual image is called an icon, and the snapshot it stores is called iconic memory.
People typically don’t realize that a fleeting mental trace lingers after a stimulus is removed from view. Nor did cognitive psychologists realize it until George Sperling’s ingenious series of experiments.
Sperling’s Experiment
Presented matrix of letters for 1/20 seconds
Report as many letters as possible
Subjects recalled only half of the letters
Was this because subjects didn’t have enough time to view entire matrix?
No
How did Sperling know this?
*
Sperling instructed subjects to stare at the center of a blank screen. Then he flashed an array of the letters for 1/20 of a second and asked subjects to name as many of the letters as possible. Try it for yourself. You’ll probably recall about a a handful of letters. In fact, Sperling found that no matter how large the array was, subjects could name only four or five items. Why? One possibility is that people can register just so much visual input in a single glance – that twelve letters is too much to see in so little time. A second possibility is that all letters registered by the image faded before subjects could report them all. Indeed, many subjects insisted that they were able to ‘see’ the whole array but then forgot some of the letters before they could name them.
Did the information that was lost leave a momentary trace, as subjects had claimed, or did it never register in the first place? To test these alternative hypotheses, Sperling devised the ‘partial-report technique’. Instead of asking subjects to list all the letters, he asked them to name only one row in each array – a row that was not determined until after the array was shown. In this procedure, each presentation was immediately followed by a tone signaling which letters to name: A high-pitched tone indicated the top line; a medium pitch, the middle line; a low pitch, the bottom line.
If the tone was presented very soon the participants would recall most of the letters in the indicated row. But if the delay was more than one quarter of a second, the participants recalled an average of just over one letter per row, indicating how quickly information is lost in the sensory register.
Sperling’s Iconic Memory Experiment
*
Sperling’s Iconic Memory Experiment
*
Sperling’s Iconic Memory Experiment
*
Sperling’s Iconic Memory Experiment
*
Sperling’s Experiment
Recall was almost perfect
Memory for images fades after 1/4 seconds or so, making report of entire display hard to do
http://highered.mcgraw-hill.com/olcweb/cgi/pluginpop.cgi?it=dcr::640::480::/sites/dl/free/0073370207/25025/ICONIC.dcr::Iconic Memory
http://www.mhhe.com/feldmaness8e
*
If the saw the entire array, subjects should have been able to report all the letters in a prompted row correctly – regardless of which row was prompted. Sperling was right: subjects correctly recalled 3 letters per row. In other words, all 9 letters, not 4 or 5, were instantly registered in consciousness before fading, held briefly in iconic memory.
To determine how long this type of memory lasts, Sperling next varied the time between the letters and the tone that signaled the row to be recalled. He found that the visual image started to fade as the interval was increased to ¼ of a second. Since this study, researchers have found when it comes to pictures of objects or scenes, words, sentences, and other visual stimuli briefly presented, people form ‘fleeting memories’ that last for just a fraction of a second.
Not an afterimage because Sperling showed he could present the letters to one eye and influence the memory by presenting a bright flash to the other eye. This would not have worked if the visual information was stored on the retina.
Sensory Memory
Why do we need sensory memory?
*
A similar phenomenon exists for auditory stimuli. The next time you listen to the radio, notice after you turn it off how an ‘echo’ of the sound seems to reverberate inside your head. This auditory sensory register is called echoic memory. Just how much auditory input is stored in echoic memory? In a study modeled after Sperling’s, Christopher Darwin and others (1972) put headphones on subjects and all at once played three sets of spoken letters – in the right ear, in the left ear, and in both ears at once. Subjects then received a visual signal indicating which set to report. Using this study and others, researchers have found that echoic memory holds only a few items but lasts for two or three seconds, and perhaps even longer, before activation in the auditory cortex fades.
Whether a sensory memory system stores information for one-third of a second or for three seconds, you might wonder: What’s the point of having a ‘memory’ that is so quick to decay? To answer this question, try to imagine what your perceptions of the world would be like without sensory memories. Without the visual icon, for instance, you would lose track of what you see with every blink of the eye – as if you were viewing the world through a series of snapshots rather than on a continuous film. Similarly, it would be hard to understand spoken language without the persistent traces of echoic memory. Speech would be heard as a series of staccato sounds rather than as connected words and phrases. In fact, we have other sensory memories as well – for touch, smell, and taste stimuli.
Short-term Memory
Conscious processing of information
Attention is the key
Limits what info comes under the spotlight of short-term memory at any given time
Limited capacity
Can hold 7 ± 2 items for about 20 seconds
AKA working memory
Working or
Short-term
Memory
Sensory
Input
Sensory
Memory
Attention
*
Think about what your environment is like as you walk from class to class on campus. You’re seeing people, trees, buildings, trash. You’re hearing multitudes of conversations, the sounds of cars as they drive past, the sounds of leaves as they fall. You’re smelling the car exhaust, the perfume of the girl next to you, the flowers that are blooming, and a pungent trash can that you walk past. More stimuli is probably reaching your sensors than you can think or write about, but most never reach your consciousness and are immediately ‘forgotten’. The key is attention. As we talked about earlier, sensations that do not capture our attention quickly tend to evaporate, whereas those we notice are transferred to short-term memory – a somewhat more lasting but limited storage facility. As we saw in the ‘Sensation and Perception’ chapter, people are selective in their perceptions and can instantly direct their attention to stimuli that are interesting, adaptive, or important.
From the sensory register, the brain encodes information – that is, it converts it into a form that can be stored in short-term memory. A stimulus may be encoded in different ways. After you read a sentence from a book, you might recall a picture of the letters and their placement on the page (visual encoding), the sounds of the words themselves (acoustic encoding), or the meaning of the sentence as a whole (semantic encoding). Research shows that people typically encode this type of information in acoustic terms. Thus, when subjects are presented with a string of letters and immediately asked to recall them, the make more ‘sound-alike’ errors than ‘look-alike’ errors. For example, subjects mis-recall an ‘F’ as an ‘S’ or ‘X’, but not as an ‘E’ or ‘B’. Subjects are also more likely to confuse words that sounds alike (man, can) than words that are similar in meaning (big, huge) – further indicating that we tend to encode verbal information in acoustic terms rather than in semantic terms.
Memorize the following list of numbers:
1 8 1 2 1 9 4 1 1 7 7 6 1 4 9 2 2 0 0 1
*
Write down the numbers in order.
*
Now, try again…
1812 1941 1776 1492 2001
*
Demonstration
Attend to the words in the green box as they flash on the screen. When the last word disappears, write down as many words as you can recall.
CAT
BREAD
DOOR
HAT
TABLE
FOOT
DOG
SON
SNOW
BUS
END
*
Cat, Bread, Door, Hat, Table, Foot, Dog, Son, Snow, Bus
Accuracy of recall for a single group of three consonants declines rapidly when subjects are prevented from rehearsing by being asked to count backwards
*
Test your memory
Do you remember what a penny looks like?
*
A
Long-term memory – Encoding
Rehearsal
Elaborative rehearsal
Levels of processing
Semantic (meaning) is more effective than visual or acoustic processing
Self-referent effect
By viewing new info as relevant to the self, we consider that info more fully and are better able to recall it
*
Do you remember your fourth birthday, the name of your first-grade teacher, or the smell of floor wax in the corridors of your elementary school? Can you describe a dream that you had last night or recite the words of the national anthem? To answer these questions, you would have to retrieve information from the mental warehouse of long-term memory. Like the hard drive on a computer, long-term memory is a relatively enduring storage system that has the capacity to retain vast amounts of information for long periods of time. We’ll examine long-term memories of the recent and remote past – how they are encoded, stored, retrieved, forgotten, and even reconstructed in the course of a lifetime.
Information can be kept alive in short-term working memory by rote repetition or maintenance rehearsal. But to transfer something into long-term memory, you would find it much more effective to use elaborative rehearsal – a strategy that involves thinking about the material in a more meaningful way and associating it with other knowledge that is already in long-term memory. The more deeply you process something, the more likely you are to recall it at a later time.
To demonstrate this process, Craik & Tulving (1975) showed a subject a list of words, one at a time, and for each asked them for 1) a simple visual judgment that required no thought about the words themselves (Is the word printed in capital letters?); 2) an acoustic judgment that required subjects to at least pronounce the letters as words (Does the word rhyme with smell?); or 3) a more complex semantic judgment that compelled subjects to think about the meaning of the words (Does the word fit the sentence ‘I saw a blank in the pond’?). Subjects did not realize that their memory would be tested later. Yet words that were processed at a ‘deep’ level, in terms of meaning, were more easily recognized than those processed at a ‘shallow’ level.
Does making complex semantic judgments, compared to simple visual judgments, activate different regions of the brain? Is it possible to see physical traces of deep processing? Using functional MRI technology, researchers devised a study similar to the Craik & Tulving study where subjects were shown stimulus words on a computer and were instructed to determine whether the words were concrete or abstract (a semantic judgment) or simply whether they were printed in uppercase or lowercase letters (a visual judgment). As in past research, subjects later recalled more words for which they made semantic rather than visual judgments. In addition, however, the brain-imaging measures showed that processing the words in semantic terms triggered more activity in a part of the frontal cortex of the language-dominant left hemisphere.
