Cognitive Psychology part 1 of Final Exam notes quizlet

What else can we predict based on this rehearsal time explanation of the primacy effect?

(See the lower half of the chart above.) Well, if we can reduce the rehearsal time of items early on the list, then we should see a reduced primacy effect. How can we reduce the available rehearsal time? One way to do it would be to use a faster presentation rate, that is, to present each word for a shorter amount of time. What happens when we do that, i.e., present the items at a faster rate? The figure below shows the results from an in-class experiment we did in an earlier semester. The experiment was with visually presented words (i.e., printed words) and used an immediate free recall task. List 1 was presented at a relatively slow, leisurely pace. Each item was shown for 3 seconds. The results are shown in the deep blue curve using diamonds. Note the robust primacy effect. List 3 (List 2 was for something else.) was presented at a faster rate. Each item was shown for a just one second. The results are shown in the chartreuse curve using squares. Note that the primacy effect is greatly reduced with the faster presentation rate, a result that is in accordance with the predictions of our hypothesis. According to our hypothesis, the recency effect shouldn't be much affected by this manipulation since it is due to recall from short-term memory and that is also the case. Recall for the last few items is roughly the same across the two rates of presentation.

Results

- Car Experts: Processing faces interfered with processing cars - Car Non-experts: Processing faces did not interfere with processing cars - Suggests that for experts same mechanism involved in processing faces and cars o But not for non-experts - Controlling factor is expertise, not whether something is a face or not The bottom line is that we use the same "expertise" module—the same resources in the brain--to recognize objects with which we are experts as to recognize faces, which we are also presumably expert in, given our lifetime of doing it. We use a different module when identifying things that we are not so expert in. You might recognize a similarity here between this conclusion and the potential reconciliation we discussed between structural description and view-based models of object recognition.

Gauthier and her colleagues compared performance of car experts and non-experts on an N-back task with a wrinkle. The wrinkles are described below. Note that, as we will see, the logic of the experiment assumes that we are all experts on recognizing faces, probably a good assumption so long as we screen people for prosopagnosia.

- The sequence of stimuli alternated composite cars with composite faceso Car—face—car—face—car, and so on - The participants task was to decide if the top part of the face/car is the same as the top part of the most recently presented face/car? o This task is an N-Back of 2, since faces and cars alternated o They used composite stimuli just to make the task a bit more difficult -- If everyone performs at 100%, we have a ceiling effect and cannot detect differences between groups - They compared this condition to performance with an N-Back of 1 (No alternation) o Car—car—car—car, and so on - Our interest is on performance with cars, where we have experts and non-experts o Everyone is assumed to be expert in face processing, so we don't expect to see differences between groups on this task

Rehersal Rate and Memory Span

-Rehersal rate is a determiner of span -So long as I can get back to and reherse an item before it is forgotten, I am within my span If I add too many items, Item A is Lost before I can get back to rehearsing it Thus the span should correlate with vocalization rate If I can manipulate rehearsal rate, I can manipulate span --If I can manipulate rehearsal rate, then I should be able to manipulate verbal short-term memory span. The next several charts test this prediction.

Multi-Sensory Integration chart

A 2 printed in black does appear as orange to WO, but that orange is perhaps not quite as vivid as if the 2 were really printed in orange.26WO and Visual Search:The ButReactionTimeNumber of DistractorsNormal: 2 among 5WO: 6 among 8WO 2 among 5Color Pop

What is a potential problem in this study

A potential problem with this study is that overall activation levels are much reduced in the After Training scans than in the Before Training scans. Although such an outcome is not terribly uncommon in the imaging literature (People may have simply been adapting to seeing faces while in the scanner.), the results would be cleaner had the overall activity level of activity for faces been more equal in the two sessions.]

What about the recency effect? Can we manipulate it like we just manipulated the primacy effect?

According to our hypothesis, yes, we can. In particular, if we can eliminate or reduce recall from short-term memory, we should reduce the recency effect. How can we eliminate or reduce recall from short-term memory? An easy way is simply to let the information in short-term memory expire. Remember that short-term memory is of limited duration. Information in it is forgotten if not rehearsed. So, if we can prevent rehearsal of the contents of short-term memory, then over time (a half minute or less), the information in short-term memory will be forgotten and the recency effect reduced.

Another way to empty out short-term memory is to delay recall for a half minute or so from the end of the list.

Again, the idea would be to allow sufficient time to elapse for the information to be forgotten from short-term memory. But we cannot simply wait for half a minute and then ask the person to recall. If they are doing nothing during that half minute, they would be free to keep rehearsing the last few items on the list and hence to maintain them in short-term memory or even (horror of horrors) commit them to long-term memory. We need to give them something to do during that half-minute retention interval that prevents them from rehearsing. A time-honored way do this, used in many psychological laboratories, is to have people count backwards by 3's from some arbitrary, large number, such as 797, during the retention interval. When we do that, we get a serial position curve like that labeled "Counting Backwards" in the figure below. Note the greatly reduced recency effect, as predicted. Our explanation is doing pretty well. (These data were collected as part of the same in-class experiment as that involving the primacy effect just discussed. Again, though, the results have been found many times in proper experiments.)

Main characterisitcs of working memory

All the information here should hopefully be evident to you by now. Short-term memory is of limited duration and limited capacity. If we are to maintain information in it for longer than that limited duration (~30 seconds), we need to rehearse it. Maintenance rehearsal—simple rote repetition—suffices to maintain information in short-term memory. Elaborative rehearsal seems to be necessary to move it into long term memory. Elaborative rehearsal involves more active engagement with the material than does rote repetition. It involves things like making visual images of the material, relating it to personal experiences, forming associations between it and already known material, and so on. We seem to have a preference to encode material in short-term store in a phonetic or verbal form, i.e., as words. We'll discuss that more when we discuss working memory. For now, I'll just add that that conclusion may be limited to adults living in a literate society (We are all presumably adults and able to read, so it would apply to us.). We are, though, when necessary or convenient, perfectly capable of encoding material in short-term store using a different modality, for example, as visual images.

example of the n-back task

An N-Back task consists of a series of trials. On each trial, a single stimulus (often just a digit) is presented. The participant needs to decide if that stimulus is the same as that stimulus presented N trials ago. For instance, with N = 1, the participant must decide if the stimulus on the current trial is the same as that on the previous trial. This is a relatively easy task. With N = 2, participants must decide if the stimulus on the current trial is the same as the stimulus 2 trials ago. This task is harder. With N = 3, the task is to decide if the current stimulus is the same as that presented three trials ago. N-Back with an N of 3 is extremely difficult.

object recognition involves____ processing

Analytical

Audio-Visual Speech

Audio-visual speech shows the senses working co-operatively together, rather than one dominating the other. Here are a couple of links to the McGurk effect (The first one is the better one).

Baddeley compared performance in each of these three conditions with performance in a condition where there was no other task that had to be performed during the short retention interval. Baddeley found

Baddeley found that both the tapping task and the letter generation task interfered with remembering the chess positions. That is, compared to a "do-nothing" condition, people remembered the chess positions less well in these conditions. Interference by the tapping task is consistent with the memory involving visual working memory. Interference by the letter generation task indicates that the central executive is also involved in remembering the chess positions. In contrast, he found no interference from the counting task, suggesting that verbal working memory is not involved in remembering the chess positions. Like Brooks', these findings are also consistent with the hypothesis that the visual and verbal working memory stores are separate from one another. If they were the same store, then we would have expected both the tapping task and the reading/counting task to have interfered with memory for the chess positions.

What happens when we take reading out of the equation? Do we still have this bias towards phonological coding?

Brandimonte and Gerbino used reversible figures such as the one in the next chart to find out. Note that this figure can be seen as either a duck (the beak is pointing to the right, the small ellipse is its eye, and it has a funny indentation at the back of its head) or a rabbit (that funny indentation on the left is its mouth, the ellipse is still the eye, and the long ears are pointing out towards the right). Brandimonte and Gerbino first familiarized their participants with this type of figure and made sure that they could reverse them. They then showed them a figure like the one above very briefly (1/2 second). During the presentation, ½ the participants repeated "la la la..." out loud (This is another example of articulatory suppression.); the other half were silent. After the figure had been removed, participants were asked what they had seen. They were then asked whether they could reverse the figure in their "mind's eye" and, if so, what they saw when the figure reversed.

What would counting do

Counting while reading presumably engages the phonological loop and should thus prevent rehearsal of information in the phonological store. But if people are storing the information in the visual store (and the visual store is separate from the phonological store), this manipulation should have no effect (since they are not using the phonological store).

Performance on the Composite Task examined

Finally, let's examine performance in the composite task. As a reminder, for most people, in the composite task, faces are processed faster in the offset condition than in the together or aligned condition, whereas for common objects, there is no difference in processing times in the offset and aligned conditions. As it turns out, car experts (but not non-experts) process cars (but not other objects) faster in the offset condition than the aligned condition, greeble experts (but not non-experts) process greebles faster in the offset condition, and ziggerin experts process ziggerins faster in the offset condition. So, the pattern of performance in the composite task condition seems not to be unique to faces but to be special to our areas of expertise.

Duration of the Phonological Store (Peterson & Peterson Experiment)

First, what is the duration of information in verbal short-term memory? Obviously, if allowed to rehearse information, we could maintain it in short-term memory for our lifetime, so long as we were content with doing nothing but rehearsing that information. But how long does information last in verbal short-term memory in the absence of rehearsal? Lloyd and Margaret Peterson set out to find out. On each trial in their experiment, a participant was given a trigram (a meaningless sequence of three consonants) to remember. They had to remember the trigram over a short retention interval, which varied from trial to trial from 3 to 6, 9, 12, 15, or 18 seconds. To prevent rehearsal during that retention interval, participants had to count backwards by 3's from an arbitrary 3-digit number (e.g., 724) during the length of the retention interval. In this experiment, at least, the duration of verbal short-term memory does not seem to be much more than 15 or 18 seconds. The estimate of the exact duration does change a bit depending on the precise experimental procedure used and the materials, but the modal result is on the order of 15 to 30 seconds, with almost all estimates falling in the range of 2 to 45 seconds.

Result

Generally, that is what happened. Thus, it does appear that WO really is experiencing the colors that he claims to experience

H.M

H. M. is perhaps the most famous and extensively studied patient in cognitive psychology. As a probable result of head injuries suffered in a bicycling accident when he was a teenager, H. M. suffered from severe epilepsy. Over time, the epilepsy became worse, and prevented H. M. from leading anything like a normal life. He was suffering seizures multiple times a day. In 1953, when H. M. was in his mid-20's, he underwent surgery that was intended to control his epilepsy. The surgery involved removing the hippocampus, a bilateral structure in the brain tucked in underneath the temporal lobe. The hippocampus from both the left and right hemispheres was removed. The surgery was successful in controlling H. M.'s epilepsy. The side effects, however, were devastating, so devastating that the surgeon who performed the operation refused to do any further similar surgeries. After the surgery, H. M. had intact short-term memory—his memory span was well within the normal range. His IQ tested slightly above average, as it had before surgery. He could remember events that occurred prior to his surgery. He had very little memory, though, for events that happened after his surgery. Later in life, he had a vague idea of an American president being assassinated (that would have been John F. Kennedy in 1963). Brenda Milner, a psychologist from McGill University in Montreal, traveled many weekends from Montreal to Connecticut (where H. M. lived) to study H.M. After many years, he did seem to show some recognition of Dr. Milner. But if you met him, left the room for 30 minutes, and then returned, he would have no recognition of you. At dinner, he would not remember what he had had for lunch, or even whether he had lunch. He knew who won the 1952 World Series (his surgery was in 1953; and it was probably the Yankees), but not the 1954 World Series or any World Series thereafter. He could learn new motor and perceptual skills. For example, he showed normal learning of a pursuit tracking task. (Such tasks are a long-time favorite of psychologists who study motor learning. In it, the participant needs to track with a joy stick or similar device a cursor as it moves about a screen.) Computers became common during H. M.'s lifetime, and he certainly would have been capable of learning how to use a mouse. But he could not remember events that happened after his surgery for more than about 30 seconds (the duration of short-term memory). (As an aside, H. M. died just a few years ago, in November 2008 or 2009, at the age of 78. By his agreement made prior to his death, immediately after his death, his brain was removed, frozen, and then shipped to the University of California at San Diego (UCSD). There, in a 53 hour long procedure, it was sliced into very thin sections, several tens of thousands of them. Each section can be made into a slide, and from these slides a 3-dimenisonal model of H. M.'s brain can be made. The procedure was carried out live on a webcam, so that people from. around the world could watch the procedure via the Internet. Such a three-dimensional model can then be used by researchers to determine more precisely the exact areas of H. M.'s brain that were removed in the surgery, thus allowing fine-tuning of the conclusions made based on the research that he participated in.)

Baddeley's Study on Chess Positions

He conducted an experiment where participants were briefly shown a chessboard with various pieces on it. After a brief retention interval, people had to recall the positions. Presumably, this task engages visual working memory—that seems to be the most straightforward way of remembering the locations of the pieces. Some involvement of the executive controller would also be expected, as people need to decide what locations to refresh from time to time, and so on. He had people perform one of three tasks during that retention interval. Counting while reading presumably engages the phonological loop and should thus prevent rehearsal of information in the phonological store. But if people are storing the information in the visual store (and the visual store is separate from the phonological store), this manipulation should have no effect (since they are not using the phonological store). A second group of people tapped a series of keys in a pre-learned visual pattern during the retention interval. Presumably, this task requires use of the visual store (to hold the pattern in working memory). Hence, we might expect it to interfere with remembering the locations of the chess pieces since this remembering is competing with the tapping task for visual working memory resources. A third group had to produce a stream of random letters. This task taps into verbal working memory (they had to speak) but also heavily into the central executive (coming up with and executing a strategy for choosing random letters). Hence, assuming that remembering chess positions also requires central executive involvement, we would expect this task to also interfere with remembering chess positions, since both tasks would be competing for central executive resources.

Brooks interference study

He had people make responses in each task in two different ways, by either saying "yes" or "no" or by pointing to the appropriate character in an array of "Y's" and "N's." Presumably, saying yes or no engages our phonological working memory (i.e., the articulatory loop) and hence would be interfered with by the phonological memory task, but not the visual memory task. Conversely, pointing presumably engages visual working memory and hence would be interfered with by the visual memory task, but not by the phonological memory task.

Why is the face inversion effect—the fact that inverting a face disrupts its processing more so than does inverting an object—taken as an indication of holistic processing.

Holistic recognition is like the image-based models we talked about in connection with object recognition. Or like the templates we talked about in connection with letter recognition. We match an image of the face as a whole against stored, photographic-like representations of faces, looking for a match. If we don't find a match, then we have to mentally rotate the image until we do. Rotation takes time; hence when I have to rotate an image to recognize it, my time to recognize it increases. Through the course of my lifetime, I see objects in many orientations. Dogs lying on their back, chairs that have been knocked over and turned upside down, even cars that have flipped over. So, I have images of those objects in an inverted orientation. When then shown a picture of such an object in a laboratory experiment, no mental rotation is necessary in order to identify it. (Admittedly, trees are a bit of a problem here.) A particular face, on the other hand, I would rarely see inverted (unless I happened to have observed that person walking on their hands). Hence, when I see an inverted version of the face, I have to rotate it in order to identify it, which slows me down, resulting in the face inversion effect.

This translation is the responsibility of the articulatory control process. If this translation is somehow prevented, perhaps we can overcome the bias to store things phonologically. How can we prevent this translation? We should be able to do so by keeping the articulatory control process busy. How can we keep it busy? Since it's responsible also for talking in general, we can keep it busy by keeping it talking. So let's have it simply repeat (aloud) a single word over and over and over while a printed list of words is also presented. This technique is known as articulatory suppression. In that case, we should not be able to translate the printed words into a speech code. To remember them, we will be forced to rely on visual short-term memory—the sketchpad. In that case, the phonological similarity effect should disappear, assuming we have correctly sorted out all these components to working memory. The following two charts summarize these ideas.

I am too tired for this

In reality, the situation is a bit more complicated than that, since both tasks no doubt also use central executive resources. Hence, some interference might be expected between visual and phonological tasks even if the stores are separate because of capacity limitations in the central executive. So, if we compare performance on a visual task without and with a concurrent phonological task, we are likely to see some interference (i.e., worse performance) from the concurrent phonological task. However, that may be due to central executive capacity being exceeded. So, generally what we do is compare performance on a visual task when it is done concurrently with another visual task to the case where it is done concurrently with another phonological task. If performance is worse in the visual-visual case than in the visual-phonological case, then we typically interpret this result as indicating greater interference in the visual-visual case than in the visual-phonological case. And presumably that interference was greater because the visual-visual case involved just one store whereas the visual-phonological case involved two separate stores.

I am too tired to understand this

Word Length Effect

I can say more short words per unit of time than I can say long words, right? Short words, after all, take less time to say. Consequently, my span for short words should be longer than my span for longer words. And in fact it seems to be so, a result known as the word length effect. The graph below plots the proportion of words immediately correctly recalled from 7-word lists as a function of the number of syllables in each word in the list. As the number of syllables increases—as the word length increases—the proportion of words recalled decreases.Earlier, I said that the statement that we can talk about verbal short-term memory capacity as being 7 +/- 2 chunks needs to be qualified a bit. It was this result that I had in mind. A word should be a single chunk. In that case, we should be able to immediately recall as many 5-syllable words as we can 1-syllable words. According to the above graph, we clearly cannot. Hence the qualification.

Composite Task

Idea that better performance in the offset than in the aligned condition indicates holistic processing. Why does that indicate holistic processing? The idea here is very similar to that behind the reasoning for the part-whole effect. Things that we process holistically we are simply incapable of breaking down into component parts, or at least it is difficult to do so. In the aligned condition, this inability causes us to see the face as a single entity. It is hard to distinguish the two separate faces in the composite. In the offset condition, the experimenter has given us some assistance by breaking the face into its relevant component parts for us.

ventriloquist effect.

If we hear a human voice and see moving lips, we will attribute the voice as coming from the moving lips. So, if I can talk without moving my lips, and instead move the lips of a dummy while I'm talking, then people will perceive the speech as coming from the dummy.We frequently experience this when we go to a movie. We see a person speaking on the screen. But the sound of their voice doesn't come from the screen; it comes from speakers usually off to either side of the screen. But we don't hear the voice coming from one side or the other; we hear it as coming from the screen, where we see the person talking. Our perception might change if we closed our eyes and listened to someone on the screen talking. (Because many theatres now use stereo, rather than monoaural, speakers, this effect may not be so apparent today. It certainly was when I was a child, before the widespread use of stereo sound.) The ventriloquist effect demonstrates how vision tends to dominate audition—we tend to believe the visual stimulus over the auditory stimulus.

Part-Whole Effect experiment explained

In phase 1 of such an experiment, participants first learn to name a number of different sketches of faces (One face, for instance, might be named "Larry;" another named "Mary," and so on. ) and to name a number of different objects, all of the same general class (e.g., "Larry's house," "Mary's house," and so on). After learning to names these faces/objects, they are given a two-alternative forced choice test (2AFC). Two of the faces or two of the objects that they have seen are shown and they are asked to choose which one is Larry's face, or which one is Larry's house.

Overcoming the Phonological Bias

In retrospect, given the nature of the stimuli used in those experiments, that conclusion (that we have a bias for phonological coding) may be a bit premature. We have a tendency to automatically read print. Reading involves transforming the printed stimulus into a phonological code. Perhaps these experiments simply reflect our automatic reading response to print.

subtraction technique steps

In the subtraction technique, the participant performs a control task (e.g., holding an object in the hand) that is intended to be exactly like an experimental task of interest except that the control task lacks the processing that we are interested in. We measure the areas of the brain that light up, i.e., are active, when that control task is performed. Then, in the stimulation or experimental condition, the participant performs a task that includes the process of interest, but in other respects is just like the control task. In our example, that task might be manipulating an object in the hand. We measure the activity in the brain during the experimental task. We then subtract the control activity from the experimental activity. What is left is the brain activity presumably associated with the process of interest. In our example, both the control task and the experimental task involve grasping an object. So, areas of the brain that work when we grasp light up. In addition, the experimental task requires that we actively manipulate the object. When we subtract the control activity from the experimental activity, we are subtracting out the activity common to both tasks. That would be the activity associated with grasping. The activity that is left over is then the activity associated with the process unique to the experimental task, i.e., to manipulating the object.

More

Information can be maintained in short-term store by rehearsing it. In the case of verbal materials, simple rote rehearsal—repeating (aloud or subvocally) the information over and over to ourselves. Without rehearsal, the information is lost, i.e., forgotten from short-term store. To get to long-term store, information must also (in general) go through short-term store, and rehearsal is also necessary to transfer it to long-term store. But not just any type of rehearsal will do. Rote repetition (also known as maintenance rehearsal) does not seem very effective for transferring information from short-term to long-term store. We'll look at some of the evidence for that claim in a later unit. To transfer information to long-term store, we must perform elaborative rehearsal on it—associating it to things we already know, forming visual images of it, turning it into a ditty or into a clever joke, that sort of thing. These rehearsal processes are carried out in short-term store, as are a number of other processes. Short-term store is where we make decisions, where we decide what to attend to, what to retrieve from long-term memory, what retrieval cues to use to help us retrieve things from long-term memory, how to encode incoming information for purposes of either working on it or transferring it to long-term store, where we actually carry out that encoding, where we parse (divide up into its meaningful units) sentences that we are hearing, and a whole lot more. Collectively, all these things are referred to as control processes. They are called control processes because they are believed to be to a large extent under control of the individual. The memory stores themselves are thought of as fixed, immutable structures, part of our biological make-up, just as a stomach, intestines, a heart, a pancreas and so on are all part of our biological make-up. We don't control them, they don't form as a result of learning and experience, they simply come as part of the package of being human.

Clive Wearing's situation

Is worse than H. M.'s. Clive Wearing suffered extensive damage to the temporal lobes (both left and right side) as a result of viral encephalitis. His ability to read music and play the piano remained intact. His short-term memory seemed normal. But he could remember no events from prior to his illness (Remember that H. M. did have memory for events prior to his surgery.) and (like H. M.) could not form memories of new events occurring after his illness. Again, we have a dissociation between short- and long-term memory, in the same direction as that shown by H. M. Clive Wearing lives in basically a 30 second window. He has little memory of things that happened more than about 30 seconds ago. Unlike H. M., he cannot even remember events that happened prior to his illness. Clive Wearing has intact short-term memory but severely impaired long-term memory.

phonological similarity effect

It is harder to remember a list of phonologically similar items than of phonologically dissimilar items, regardless of whether the items are presented as spoken words or visually printed words. The similar sounding items interfere with one another, making everything one big blur. The effect occurs for visually presented words, i.e., printed words, because we have a tendency to automatically read such words, translating them into a speech code and placing them in the phonological store.

Part-Whole Effect

It might be a bit easier to see why the part-whole effect, the finding that a whole face is easier to recognize than part of a face but a whole object is not typically easier to recognize than a part of an object, is seen to indicate holistic as opposed to analytic processing. Holistic processing says that we have difficulty selectively attending to parts of faces (or anything else that we may be processing holistically), and that processing of a face as a unit is obligatory. We have to treat the whole face as a single entity that is not decomposable into lower level entities. To say it a bit differently, we cannot process one part of a face without simultaneously processing all the other faces. Under such assumptions, it should be fairly evident why recognizing a part of the entity would be difficult.

Visual-Verbal interference (Lee Brooks)

John ran to the store to buy some oranges memorize sentence and indicates whether each word is a noun by; Task 1: Saying "Yes" or "No" Task 2: Pointing to "Y" or "N"<-- this is the easier (spatial task) a). a verbal stimulus F visualize "F" and indicate whether each corner is "Outside" by: Task 3: Saying "Yes" or "No" <-- Easier (Verbal task) Task 4: Pointing to "Y" or "N" b). Spatial stimulus

serial position curve

Literally thousands of experiments have found the U-shaped serial position curve that is evident in the figure. The good recall at the beginning of the list, we call the primacy effect (primacy for first—we remember the words we are first exposed to well). The good recall at the end of the list, we call the recency effect (as in recent—we remember well that which we have seen/heard/experienced most recently). As mentioned, the primacy and recency effects have been found in many, many, many different experiments. In fact, it can be a challenge to conduct an experiment and not find them (But we can manipulate them, making them larger or smaller—see below). A finding so ubiquitous deserves explanation. And many years ago, it was results like these that helped influence early memory researchers to think in terms of separate and distinct short-term and long-term memory systems rather than in terms of a single, monolithic memory system.

Typical result for the experiment

Look first at the left-hand side of the table, the Silent condition (i.e., no articulatory suppression). There is a phonological similarity effect for both visual presentation and auditory presentation (note that it is much larger with auditory presentation than with visual presentation, suggesting some visual coding with the visual presentation condition). Now look at the right hand side of the table, the condition with articulatory suppression. For visual presentation, there is no phonological similarity effect. (And overall recall is reduced, suggesting that phonological coding is more efficient than visual coding.) For auditory presentation, as we predicted, there is still a phonological similarity effect, reflecting the fact that we are still encoding these words phonologically. However, compared to auditory presentation in the silent condition, recall is much lower, reflecting the fact that people were unable to rehearse in the articulatory suppression condition.

Components of Working Memory

Most researchers would agree that working memory can be divided into several components. Suppose that you have been browsing Avers' pizza menu on your laptop and now decide to call them to place your order. You note their number and commit it to working memory. In what form do you store or encode the number? Most people, if they introspect about this question, would say that they store it verbally or phonologically. They remember the words, "three three three five five five five." Note that there is no logical reason why they could not have remembered the number in a different format. They could have formed a mental image of it as printed in white chalk on a blackboard, or as written with an orange felt-tip marker on a whiteboard, or as written in the sky by a sky-writing plane. Generally, though, people simply remember it as a sequence of words and maintain it in working memory by rehearsing it by silently repeating it to themselves until they finally find the phone. Let's go back to our example of the GPS system telling us to turn right in 400 feet. We can also ask people how they store and maintain that instruction in their working memory. Here we get more of a split opinion. Some people report that they store that information too in a verbal form—they remember the words of the verbal instruction. Others will say that they convert the verbal information into a visual image—they picture the road stretching out in front of them for another 400 feet and then making a right turn—and maintain that visual image in working memory.

alterations of the McGerck Effect

Now suppose we see the face of the person who is talking, but the sound has been turned off. Again, we just have to identify what word they are saying. Though a handful of people do pretty well on this task (getting about 25% correct), most of us do lousy. We only get 1-2% correct. Now suppose that we let our observer both see the face of the person doing the talking and hear his/her voice. Gradually we turn up the noise. What happens in this case is that as we turn up the noise the visual information becomes quite useful—the same information that only allowed us to get 1 or 2 percent correct in the absence of hearing, now lets us do much better than would be expected on the basis of how well we did with auditory presentation alone and how well we did with visual presentation alone. Suppose that we are at a noise level where with just auditory presentation we can identify 60% of the words. Suppose that with just visual presentation, we can identify 5% of the words. Then, when we have both auditory and visual presentation, we should be able to identify.60 + (1 - .60) * .05 = .62of the words. (If you don't understand how I arrived at the algebra, see below. If you still don't follow it, please ask in class.) In fact, we understand more like .70 of the words both seen and heard. This phenomenon is known as superaddivity. It's important because it is one of the few known cases where neither sense dominates the other, but instead the two work together to determine the percept and result in a more accurate percept than what we would get with either alone.Here's the logic of how we calculated the above number. With auditory alone, we are going to get 60% (.6) correct. That leaves 40% (1 - .6 = .4). Of that 40%, we will get .05 correct via our visual capabilities: .4 * .05. Our total percent correct is then the sum of those two: .6 +.4*.05 = .6 +.02 = .62Let's turn finally to a case of multi-sensory integration where one sense evokes a percept in another sense, even in the absence of any stimulation of that second sense.

When looking at houses

Now the activity has moved even closer to the midline, to the parahippocampal place area of PPA.

A Bias for Phonological Coding

Okay, we have separate verbal and visual working memories. Some information is naturally encoded in one; other information is naturally encoded in the other. For yet other information, we might have a choice of whether to encode it visually or linguistically.

working memory has two different components.

One, that we will refer to as verbal or phonological working memory, is used to hold information in a verbal or phonological form (The term phonological refers to sounds, and specifically to the sound of words.). Information in this phonological store is in the form of spoken words. The second we will refer to as visual working memory, or as the visual spatial sketchpad. Here, we store (and manipulate) information in a visual form, as a visual image.

Wo and Visual Search summary

Orange 2 and Green 5, blue 8 and 6 • 5's popped out among 2's (when both printed inblack)• 8's did not pop out among 6's (when bothprinted in black) We have to temper that conclusion just a bit when we consider the data in more detail. The figure below shows reaction time as a function of the number of distractors. Consider first the top line, in which all numbers were printed in black. For normals (non-synesthetes) searching for 2's among 5's or 6's among 8's, and for WO searching for 6's among 8's (both of which presumably appear blue to him), reaction time increased greatly as the number of distractors increased. Now consider the bottom line, corresponding to normals searching for an orange 2 in a background of green 5's. Here there was no effect of the number of distractors; the 2 simply popped out. Now consider the middle line, WO searching for a black 2 (which should appear orange to him) among black 5's (which should appear green). Because the digits presumably are perceived in different colors by WO, the 2's should pop-out and there should be no effect of the number of distractors. To a large extent, that prediction is borne out: as the number of distracting 5's is increased, WO's reaction times are much faster than when he is searching for 6's among 8's. Still, though, unlike the case for non-synesthetes using letters that really are printed in different colors, there is a small increase in reaction time as the number of distracting 5's increases. The usual interpretation given to this pattern of findings is that, yes, it does seem that WO really experiences the color that he claims to experience when seeing a particular digit, but the perception of that color is not quite as vivid as perceiving the actual, physically present color.

Risk factors

Perhaps the statement that the more "risk factors" one has, the more likely one is to be a synesthete also merits some elaboration. If one is an artist one is more likely to be a synesthete than if one is an engineer. If one is female, then one is more likely to be a synesthete than if one is a male. If one is both a female and an artist (2 risk factors) one is more likely to be synesthetic than if one is a female engineer (one risk factor) or a male artist (one risk factor).

What has limited capacity when we talk about attention

Phonological store and the visual-spatial sketchpad that have limited capacity and the central executive has limited capacity

What was the purpose of having half the participants say "la la la..." while viewing the stimulus?

Presumably, this keeps the phonological loop busy. The phonological loop is presumably involved in the translation of a visual stimulus into a phonological code—i.e., naming the picture. Consequently, if it is busy, then we cannot encode the figure phonologically; even if we have a bias towards such an encoding, the conditions imposed by the experiment force us to preserve the visual encoding. In that case, more people ought to be able to reverse the figure in their memory in the "la la la" condition than in the silent condition. The chart below summarizes these points.

results of the whole-part experiment

Proportions correct in the isolated part condition and in the whole object conditions are plotted separately for faces and houses. For faces, it makes quite a difference whether the participant is asked to identify the whole face or just a part of it —they are much worse when identifying just a part than when identifying the whole face. For houses, it does not make a difference whether the participant must identify the whole house or just a part —proportions correct are about the same for those two cases for houses.

Capacity of the Phonological Store (Digit Span)

Second, what about the capacity of verbal short-term memory? Capacity is most frequently measured by using the digit span task. (There is nothing particularly sacred about the digit span task. But by consistently using the same task across different studies, we can compare the results of those studies with one another.) In the digit span task, on trial 1, the person is presented a single digit and asked to immediately recall it. Most people succeed at this tough task, so we make things tougher on trial 2. Now we present them 2 digits and ask them to immediately recall those two digits. Most people can do this too, so on trial 3, we add a third digit. We keep going until the person is unable to perfectly recall all the digits. We then define span as the longest string of digits that they were able to immediately recall without error. Measured in this way, the capacity of verbal short-term memory of a young, healthy adult is 7 +/- 2 items, i.e., it is from 5 to 9 items. Okay, so our capacity is 7 +/- 2 digits when measured by the digit span task. More generally, we can say that the capacity of verbal short-term memory is 7 +/- 2 items. But what is an item? Consider the digits 8123325947. That's a lot of digits (10) and beyond most people's capacity. But break it apart and we have 812 332 5947. Most of us now recognize that as a Bloomington phone number (Please don't dial it; I have no idea whose it is.) and when viewed in that way it becomes much easier to remember all 10 digits. Basically, we need only to remember the last four digits plus the digit that comes after the 33. What we are doing here is a process called "chunking," or combining multiple lower level items into a single higher level item, or chunk. The three digits 8 1 2 get combined into the single item "Bloomington Area Code." Now, instead of three things to remember, we have just one. When we talk about the capacity of verbal short-term memory, we really need to talk about chunks—the capacity is 7 +/- 2 chunks (There are some minor qualifications to this statement but we can take it as being basically correct.)

We (the modal model) have been arguing that the short- and long-term memory stores are distinct from one another. How do we know that long-term memory and short-term memory are really distinct and different memory stores?What evidence is there that they are in fact two different systems?

Several types of data have been advanced to support that view. We are only going to discuss two, the serial position curve and some neurophysiological evidence. The evidence for separate long-term and short-term stores from the serial position curve that has been used to support the long-term/short-term memory distinction is the evidence that we reviewed earlier here. The distinction gives us a rather simple and elegant explanation for the ubiquitous occurrence of primacy and recency effects. But does that explanation really demand that the stores be physically independent or separate stores? Not really. One could take the view that short-term memory is simply that portion of long-term memory currently activated and still buy into our explanation of the serial position curve. There is though also some neurophysiological evidence for a separation of short- and long-term memory, reviewed in the following section. (Note that this is an example of converging evidence. The neurophysiological evidence converges with the evidence from the serial position curve to support the hypothesis of separate short and long-term memory stores.)

How can we prevent rehearsal?

Simply by having the person do something else that prevents rehearsal. One way is to switch from a free recall task to a serial recall task. Recall that in serial recall, the person must begin at the start of the list and recall items in order, from first to last. So, the strategy of recalling the last items doesn't work here. And while the person is recalling the items from the beginning of the list, they cannot be rehearsing—the two behaviors are incompatible with one another. By the time the person finally gets to the end of the list and is ready to recall it, information will have been forgotten and the recency effect will be reduced. Does that in fact happen? Flip back to the figure showing the serial position curve briefly. The left hand panel shows the serial position curve for serial recall. The right hand panel shows the serial position curve for free recall. Note that the recency effect is indeed greatly reduced in the left hand panel (serial recall) as compared to the right hand panel (free recall).

K.F

So far, we have seen damaged long-term memory with intact short-term memory. To get a double dissociation between short- and long-term memory, we would have to find patients with impaired short-term memory and intact long-term memory. Patient K. F. seems to be such a patient. She has severely impaired short-term memory. Given a list of digits to remember, she can remember an average of 2. (Normal is in the range 5 - 9.) Given a list of words to recall, she shows no recency effect (Remember our explanation of the recency effect.). Her long-term memory though seems okay. She shows a primacy effect in free recall. To be sure, it does take her longer to learn a list than it would take you or I, but that is to be expected given her diminished short-term memory abilities—information needs to go through short-term memory in order to get to long-term memory. Her diminished short-term memory ability means information goes through it more slowly. But once something does get into her long-term memory, it is in there solidly—she forgets it no more rapidly than you or I would. Note that her damage is to the frontal lobe. K. F., then, is a patient with seemingly intact long-term memory but impaired short-term memory. There are some other patients with damage to similar areas of the brain that show a similar deficit. The chart below summarizes the situation with K. F.

Inversion Effects with Experts

So, activity in the fusiform gyrus seems not to be special to faces but rather to expertise. What about the effects we interpreted as evidence for holisitic versus analytical processing in face recognition (try to recall those three effects)? Are they also specific to faces or might they to apply to other areas of expertise and not just to faces? Let's start with the face inversion effect. Experts also show an inversion effect for things that they are expert in. Dog experts, for example, show a very large increase in reaction time in recognizing dogs that are shown upside down instead of right side up, an increase that is much larger than the increase they show for other objects. Similarly for car experts and bird experts. Apparently, the inversion effect is not specific to faces but is associated with expertise.

Speaking Rate and the Capacity of Phonological Working Memory

So, chunking can increase the capacity of our verbal short-term memory. What else can? It turns out that how fast we can talk also affects our capacity.Consider the chart below.

An Expertise Module?

So, the previous studies suggest that we do not have a face module but perhaps an expertise module. We are now going to look at a rather clever use of the N-back task to get some converging evidence that the brain module we use for processing faces is the same module that we use for processing objects in which we have expertise. Put another way, the question concerns whether face recognition and recognition of objects in which we are expert share mental resources. The logic of this experiment is quite similar to the interference studies that we looked at in connection with visual and verbal working memory, or will look at when we talk about working memory. If two tasks share mental resources, they should interfere with one another. If they do not share mental resources, then they should not interfere.

Frequency of Synesthesia

Some characteristics of synesthesia and synesthetes are indicated in the list below. Regarding the frequency, although some authors claim a frequency as high as 1 in 100, the figure 1 in 1000 appears to be more accurate (that's why it is circled). The 1 in 100 figure almost certainly includes a lot of associators (see below). Most of the remainder of the slide is hopefully self-explanatory, but perhaps the last bullet, concerning cause, needs a bit of elaboration, provided below the list.• Frequency: extremely rare: 1 in 1000 to 1 in 100• Much more frequent in females than in males• Many synesthetes have an artistic or musical bent- The more" risk factors" (female, artist) one has, the more likely one is to be synesthetic• Most synesthetes consider synesthesia a gift, not a disorder• Associated with higher levels of neural activity between the two areas of the brain involved with the individual sensations - Increased neural connections between different sensory areas of the brain?- Decreased inhibition of connections between different sensory areas of the brain?

Now let's consider a third task.

Suppose that I ask you how many windows are in the house or apartment that you currently live in. When answering that question, most people report relying almost exclusively on visual information. They picture the layout of their apartment, walk into the first room, form a visual image of that room, count the windows in that room by inspecting the image, walk into the second room, form a visual image of that room, count the windows in it, and so on.

How did Brandimonte and Gerbino experiment go

Tested three different groups of participants: 6 year olds, 10 year olds, and adults. Their thought was that perhaps any bias towards phonological encoding needs time to develop and in particular is a consequence of becoming more competent with spoken language. In that case, adults might show a bias towards phonological coding whereas children still developing their language skills would not, as they have not yet had time to develop that bias.

What is the articulatory loop

The articulatory loop refers to the normally sub-vocal rehearsal that we do in order to maintain information in the phonological store, i.e., silently repeating the information over and over to ourselves.

What about the primacy effect?

The basic idea here is that we will have more opportunity to perform elaborative rehearsal (the type of rehearsal that helps transfer information from short-term memory to long-term memory, and hence helps us remember things over the longer term) on items presented early on the list than for items presented in the middle of the list or towards the end of the list. Why is this? It's simply because when the first couple of items are presented, we have no other items to rehearse but those couple. But when we are in the middle of the list, and an item is presented, we have not only that item to rehearse but all the previous ones as well. So, in the middle of the list, our rehearsal time is spread out across more items, giving us less time to focus on the new ones. Let's make the simplifying assumption that people rehearse each item that they can currently remember equally often. Each row in the table above shows the proportion of time rehearsing the first, second, third, and fourth items when the first items is presented (row 1), when the second items is presented (row 2), when the third item is presented (row 3), and when the fourth item is presented (row 4). The last row, labeled "Total," shows the total units of time spent rehearsing each item after the first four items have been presented. What is rehearsed when the first item is presented (row1)? Since the first item is the only item known so far, only that item is rehearsed and it gets a full unit of rehearsal time. What happens when the second item is presented (row 2)? Well, now we have 2 items to rehearse. Under our simplifying assumption, we'll devote half the time to rehearsing item 1 and half the time rehearsing item 2. Each item gets ½ of a unit of rehearsal time. Now we present the third item (row 3). At this point, we can give each item only 1/3 of a unit of rehearsal time. And when we present the fourth item (row 4), each item gets only ¼ of a unit. By summing the values in each column, we get the total amount of rehearsal for each item shown in the last row. As can be seen, items presented earlier in the list get more rehearsal time; hence (assuming that it was elaborative rehearsal—see later in the course), they are more likely to make it into long-term memory; hence, they are more likely to be recalled. That is, items earlier in the list are more likely to be recalled than items more towards the middle of the list, i.e., a primacy effect should occur.

A Note on the Capacity of the Visual Store

The capacity of visual working memory is estimated to be about 3 to 4 items. We should note, though, that defining a visual item is even more slippery than defining a verbal item.In addition, there is increasing and recent evidence that in visual store we can trade off quality for quantity. An analogy to a photograph may be in order here. We can store either a larger part of a scene at relatively low resolution, or a smaller part of the scene at higher resolution.

results of the grid experiment

The dependent variable is percent correct. The solid chartreuse line with squares corresponds to memory for the identity of the letters and is plotted as a function of whether the intervening task was visual (odd-grid out) or verbal (addition problems). The dotted blue line with diamonds corresponds to memory for the position of the letters, and is again plotted as a function of whether the intervening task was visual (odd-grid out) or verbal (addition problems). Memory for position information was much worse with the visual intervening task than with the verbal intervening task. With the visual intervening task interferes visual memory performance is worse than verbal memory. Conversely, memory for identity information was worse (though the effect is smaller) after a verbal intervening task than after a visual intervening task. With the verbal intervening task verbal memory s worse than visual memory. Again, the pattern of results is consistent with the hypothesis that we have separate verbal and visual working memories, as opposed to one memory capable of encoding information in both modalities. The interference is selective—verbal interferes with verbal but not visual; and visual interferes with visual but not with verbal.

What were the errors in assuming that adults have a bias to store things phonologically

The gist of this evidence was that errors in short-term memory tasks tended to be phonological in nature rather than visual in nature. For instance, given a list of letters to hold in short-term memory, people had more difficulty immediately recalling them when all the letters were acoustically or phonologically similar (compare the two lists under the first bullet on the chart—the first results in many more errors in immediate recall than the second). This effect occurs regardless of whether the letters are presented auditorally (spoken) or visually (in print). No similar effect occurs for a list of visually similar letters compared to a list of visually dissimilar letters—there is no difference in how easy or hard it is to recall such lists. That is true regardless of whether the letters are spoken or printed. Similarly, no such effect occurs when comparing lists of semantically similar words to lists of semantically dissimilar words. Similarly, when people make an error recalling a list of letters or words, they tend to recall a letter or word acoustically similar to the intended word; they do not recall a letter visually similar to the target. See the example under the second bullet. Again, this effect occurs regardless of whether the list is presented auditorally or visually. Visually similar letters, however, are not confused with one another, even when the letters are presented visually.

results of Brooks experiment

The graph below shows Brooks results. The dependent variable is response time—how long it took participants to complete the task (indicating nouns vs. non-nouns, or inside vs. outside corners). The X-axis shows the memory task, either the sentence task or the block-letter task. The line with solid figures corresponds to the case where participants responded by pointing; the line with open circles to the case where they responded by speaking. Though reduced in magnitude, the opposite pattern occurred when participants spoke their response—they were slower on the sentence memory task than the visual memory task. Hence, a visual responding mode interferes more with visual memory than it does with verbal memory. And a phonological responding mode interferes more with verbal memory than it does with visual memory. This pattern of results is that expected if in fact the phonological and visual working memory stores are separate from one another.

What were the results to the la la las condition

The gray bars correspond to the "la la la" condition; the blue bars to the silent condition. The plot is the number of participants able to reverse the image, shown separately for each age group. Look first at the adults. Reversals are twice as common in the "la la la" condition than in the silent condition. Apparently, in the silent condition, people did in fact have a bias to encode the stimulus phonologically—basically, they named it and remembered the name. In the "la la la" condition, since the phonological loop was busy, such recoding was blocked and participants were forced to retain the visual coding. Since the visual coding was still available to them, they were able to reverse the figure in memory. Basically the same pattern of results occurred for 10 year olds. A different pattern, however, emerged for the 6 year olds. The children were equally able to reverse the figure in the silent condition as in the "la la la" condition. Apparently, these young children had no such bias to recode the picture into a word and were quite happy to remember the visual coding. Since they were remembering visually, they were able to reverse it. (It is true that in the "la la la" condition, fewer 6 year olds were able to reverse the figure than were adults, but the same would be true if we left the figure visible to the 6 year olds and asked them to reverse it. This is a hard task for them. What it important here is that the same number of 6 year olds could reverse the figure in the silent condition as in the "la la la" condition.) So, literate adults and even 10 year olds have a bias towards encoding stimuli that could go either way into a phonological code, even when the stimulus is presented visually. Six year olds, who are still developing their language skills, demonstrate no such bias.

Rubber hand experiment proves

The illusion is another case where vision dominates another sense, in this case the sense of proprioception (though to be sure in this case vision needs a bit of an assist from our haptic senses). Our sense of proprioception is our sense of where our body parts are. Our haptic sense is our sense of touch. Another example of the dominance of vision: take wine snobs who pride themselves on being able to judge different wines. Now add red food dye to a white wine so that it appears as a red wine. Our wine snobs will describe it as a red, not a white. So much of their superior knowledge. In this example, vision is dominating the senses of olfaction and taste.Then there's the dog feces experiment that really involves fudge, not dog feces. Experimenters offered fudge to participants that had been molded into the shape of dog feces. Despite the fact that the fudge clearly smelled like fudge and not like dog feces, there were very few takers. In this case, vision is dominating olfaction.

Synesthesia

The list below defines synesthesia and lists some of the forms of synesthesia. Interpret the -> symbol as "leads to" and the chart should be self-explanatory.• Input in one sensory modality (or sub-modality) causes a perception in another modality (or sub-modality) - Graphemic synesthesia: a printed letter is seen as having a particular color - Days of week/months --> color perception - "Blue tones" - Musical chord --> color perception - Musical chord --> taste perception - Musical chord --> taste and color perception

What does the mirror box show

The mirror box shows the phantom limb operating in a normal way via a visual illusion. That illusion is enough to often trick the brain into thinking everything is okay with the limb, and the phantom pain will disappear. The use of the mirror box can be considered another example of the visual system dominating the haptic sense. So, although phantom limb itself is an exception to our generalization, at least one treatment for it takes advantage of our generalization.

The Modal Model of Memory.

The model is more of a framework than a model. A model can be proven wrong or receive so much empirical support that we eventually accept is as a good approximation to a true description of the world. Models make predictions concerning what will or will not happen under various laboratory conditions. We can conduct experiments to determine if the model's predictions hold or not and based on the correctness of the model's predictions, we come to reject or accept the model as accurately reflecting reality. A framework is more a way of thinking about a problem area, a way of guiding research. It doesn't necessarily make specific empirical predictions and hence cannot be shown to be right or wrong. It does, though, suggest what questions are worth asking. It can be more or less useful, depending on how well it helps us formulate hypotheses, develop methods of investigating phenomena of interest, apply our knowledge to real-world problems, and so on.

Separating the Verbal and Visual Stores; A question that makes sense to ask is whether there really are separate visual and phonological stores. Or is there perhaps a single store that is capable of storing and manipulating both visual and phonological information?

The most common way that cognitive psychologists have addressed this question is through interference studies. The logic of the interference study is as follows: Suppose that I give a person a task to do that is relatively demanding on the visual store—it requires that person to use a significant chunk of that store's available capacity. Now, I add a second task to the mix, but that task is a phonological task. Suppose the phonological task uses the same store as the visual task, but just encodes the information in a different format. In that case, adding the phonological task will cause the capacity of that single store to be exceeded and performance on the visual task will suffer—the two tasks will interfere with one another. But suppose on the other hand that the phonological task uses a different component of working memory, with its own capacity. In that case adding the phonological task will not cause interference with the visual task. Similarly, if we begin with a phonological task and add a visual task, if both rely on the same store, then we should see interference between the two tasks. But if in reality they actually rely on the same store, then we should see interference. Basically, the argument is that if visual processing interferes with phonological processing and phonological processing interferes with visual processing, then there is a single store capable of handling both phonological and visual information. But if visual processing does not interfere with phonological processing and phonological processing does not interfere with visual processing, then the implication is that the two stores are separate from one another, each with its own capacity. And of course, under both hypotheses (common store or separate stores) phonological processing should interfere with phonological processing and visual processing should interfere with visual processing—if they do not, then we are not using a very demanding task.

The isolated part condition

The participant sees only a part of two different faces (e.g., the eyes) or a part of two different houses (e.g., the door) and must now choose which of the two are Larry's eyes, or which of the two is the door to Larry's house.

whole object condition

The participant sees two whole faces or two whole houses and must choose, for instance, which is Larry or which is Larry's house.

What happens as time btn. Rehearsals increases?Decreases?

The plot here shows the strength of the short-term memory trace as a function of time. The stronger the trace, the more likely are we to recall the corresponding item. If the trace strength gets too low, then we are not going to be able to recall the corresponding item. An item goes into verbal short-term memory. It starts off with maximum possible strength. Over time, if we do not rehearse the item, its trace declines. After a period of time, though, suppose that we do in fact rehearse that item (Indicated by the R's along the horizontal axis in the figure.). That rehearsal revives the trace strength all the way back up to its maximum. It begins to decline again. If I get back to it in time and rehearse it again, the strength goes back to the maximum. This cycle continues until for whatever reason I stop rehearsing that item, so that its strength declines until it gets low enough that I am no longer able to recall it. To rehearse an item, I have to implicitly recall it—I rehearse the first item, the second item, and the third item. Now, I want to start over and rehearse the first item again. To do that, I have to be able to recall the first item. What then happens as the interval of time between rehearsals of an item increases? The longer that interval the weaker the trace becomes. The weaker the trace, the less likely I will be able to recall it in order to give it another rehearsal. In other words, the longer the time between rehearsals, the more likely I am to forget the time. By similar reasoning, the less time between rehearsals, the less likely I am to forget the item. Perhaps now you can begin to see why the faster I can talk, the larger is my memory span, i.e., the more items I can hold in verbal short-term memory. Rehearsing is basically talking. Suppose that I am trying to hold a bunch of items in verbal short-term memory. I start saying the bunch to myself. When I get to the end of the bunch, I start over again at the beginning. But if the bunch is too long, by the time I get through the bunch, the trace of the first item is too weak for me to be able to recall and rehearse it. I've forgotten it. If I can talk really fast, though, then I can get through the bunch more quickly and get back to the first item before its trace becomes too weak. If a trace lasts for 1 second and I can say four items in that 1 second, then I will be able to hold four items in my verbal short-term store. But if I can say 6 items in that 1 second, then I will be able to hold 6 items in my verbal short-term store. The faster I can talk, the larger is my memory span, predicts this analysis. Measure how long it takes you to recite the alphabet to yourself. The less time it takes, the longer your verbal short-term memory span tends to be.

how can we account for the shape of the serial position curve, i.e., the primacy and recency effect?

The recency effect is probably the easiest to explain so let's begin with that. The basic idea is that the recency effect reflects recall from the short-term memory store. The short-term memory store is of limited capacity and limited duration. That is, it can remember only a few items of information (a few words in our experiment, approximately 5) and if those items are not rehearsed, they will quickly be forgotten from the short-term store. You can think of short-term memory as containing that information that you are conscious of at any given moment. It is the here and now of experience. When presented with a list of words and then asked to immediately recall them, a sensible strategy would be to first recall those words still in short-term memory—if you recall them now, you'll get them right; if you wait and try to recall them later, they will have slipped out of short-term memory and unless you managed to get them into long-term memory, you won't be able to recall them. What words will be in short-term memory when we get to the end of the list? Since we constantly update short-term memory with new information coming in, the words that will be in short-term memory will be the ones most recently seen or heard, that is, they will be the ones from the end of the list. So, we will do well on the last items of the list—we will show a recency effect-- because we can recall them from short-term memory.

Results of the overall situations

The table above summarizes the overall situation. Putting together Clive Wearing and H. M., on the one hand, with K. F. on the other hand, we have a double dissociation. Wearing and H. M. show a long-term memory deficit but with intact short-term memory. K. F. shows a short-term memory deficit but with intact long-term memory. This pattern of results suggests that in fact short-term memory and long-term memory are implemented in the brain as separate and independent systems. This neurophysiological evidence is perhaps the best evidence we have for that claim. We could spend half a semester in a seminar course discussing and reading about this debate, but we are not going to do that. We are going to tentatively accept the reality of the distinction and move one. And even if we are hesitant to accept that the two memory structures are implemented in separate systems in the brain, the distinction is certainly useful for organizing. a ton of research that has been done in memory. At least at a descriptive level, the distinction has been and continues to be very useful.

subtraction technique

The technique used in brain imaging in which baseline activity is subtracted from the activity generated by a specific task. The result is the activity due only to the task that is being studied.

results from an immediate free recall experiment.

The three different curves show the results for lists of three different lengths (where length simply refers to the number of words in the list). The data plotted are the percent recalled as a function serial position. Serial position refers to whether the word occurred in the first position of the list, the second position, the third position, etc. You can see that the curves for all three list lengths have a distinctive shape—they are all U-shaped. Recall is very good at the beginning of the list and at the end of the list; it is relatively poor in the middle of the list.

Brooks experiment expalined

The visual memory task required people to hold in memory an image of a block letter F. They then traveled around that image in their mind's eye, indicating whether each corner of the F was an inside or outside corner. Again, they made their response in one of the two ways described above. In this case, assuming that the two stores were separate, we would expect performance to be worse when participants pointed versus when they spoke their response, since both the memory task and that mode of responding (pointing) require resources from visual memory.

two conditions in this 2AFC test

The whole object condition The isolated part condition.

Is there an intermediate store?

There is also some evidence, much of it from the neuropsychological world, that besides short-term (or working) and long-term memory there is what might be considered a third memory store, which I will call here "intermediate memory." It seems to function as a temporary storage place, somewhere to hold the memory while it is being integrated into our long-term memory, a process often referred to as consolidation. Until consolidation occurs, memory for an event can be very fragile. Sleep may play a role in consolidation, in particular deep or non-REM sleep. Finally, there is some neuropsychological evidence that the hippocampus may be the storage "location" for this intermediate memory—disruption of the hippocampus during the consolidation period can disrupt our long-term memory for the event. This would explain H. M.'s deficit (as well as that of numerous people with deficits similar to H. M.'s). That is, we temporarily hold a memory in the hippocampus, but for it to become permanent, it needs to be transferred to other long-term store in other areas of the brain. That process of transfer is what we call consolidation. See the summary below.

three areas where we obtain different empirical results for face recognition compared to what we normally get for recognition of other objects.

These are - the inverted faces effect - the part-whole effect and performance on a composite task.

Holistic Processing

These three effects: —the inverted faces effect -- the part-whole effect, and performance in the composite task (better performance for faces in the offset than in the aligned condition) —have led a number of researchers to propose that recognizing faces is fundamentally different from recognizing other types of objects. In particular, the idea is that object recognition involves analytical processing, whereas face recognition involves holistic processing.

What is the nice thing about the grid experiment

They can examine memory of item information (verbal information) separately from memory of position information (spatial information) for the same stimulus. Do participants recall the right set of letters (item information), regardless of where in the grid they are? Do participants put a letter in the correct spatial positions (position information), regardless of whether that letter is the correct letter or not? Hence, with the same recall task, they are measuring both phonological short-term memory and visual short-term memory.

The Grid Experiment den Heyer and Barrett explained

They looked at the question of whether the form of concurrent activity affected the specific type of information that people fail to recall about a common stimulus. On each trial of their experiment, participants saw a 3 x 3 grid. Four of the nine squares were populated with a letter—see the top figure. Participants had to remember both which letters were shown (verbal information) and in what square on the grid they were shown (visual-spatial information) across a short retention interval (approximately 30 seconds). During the retention interval, on some trials participants solved addition problems, as shown on the left side of the middle set of figures. Presumably this activity engaged the phonological store. On the other half of the trials, they had to perform an odd-grid out task, as illustrated on the right side of the middle row of figures. They were shown 3 grids and had to select the one that was different from the other two (the right-most grid in the example), a task that presumably engaged visual working memory. They then had to recall the original grid.

What Kind of Expertise Involves Holistic Processing? This study asks the question of whether any old expertise (or just simple familiarity) is sufficient to engage the "expertise module," or a particular form, behind simple familiarity, is required to engage that module

This study used Ziggerins, another mythical set of objects coming out of the Gauthier lab. The experiment went as followed: Participants in Wong et al.'s experiment received extensive training with Ziggerins. Different participants received different kinds of training. Individuation training involved learning to identify (i.e., name) individual Ziggerins from several different classes (Hence, a unique response was made to each Ziggerin. This is like learning to identify an individual dog as "Fido," or "Pooch," and so on.). Class training involved learning to name a Ziggerin's class (Hence, the same response was made to all Ziggerins in the same class. This is like learning to identify a Labrador, a German Shepherd, and so on amongst dogs.). Participants were then given a composite task using these Ziggerins as stimuli. The individuation training group showed better performance in the offset condition than in the together condition. That is, they showed the pattern of performance characteristic of faces and of experts. We might expect this since they had received extensive training with the Ziggerins. This pattern indicates holistic or configural processing. The class training group showed no difference in performance between the offset and together conditions. That is, they showed a pattern of performance characteristic of non-experts, despite their extensive practice with Ziggerins. Hence, not just any kind of experience will do for developing holistic processing. Learning to identify an object's general class does not. Learning to identify it as an individual will.

notes

This time, we'll start with the primacy effect. Our explanation claims that we rehearse items occurring earlier in the list more frequently than we rehearse items occurring in the middle or end of the list. Do we? How can we measure number of rehearsals? Rundus (1971) addressed this question by asking his participants to rehearse aloud as they went through the list. He used an immediate free recall task. His results are plotted in the figure above. The closed circles show proportion correct as a function of serial position on the free recall task. We see that he obtained the classical serial position curve, showing both a primacy and recency effect. The open circles plot the number of times an item was rehearsed as a function of its serial position. As can be seen, and as predicted by our explanation, items at the beginning of the list were indeed rehearsed more often than items in the middle (or end) of the list. And they were recalled better. In fact, for items at the beginning and middle of the list, the curve showing number of rehearsals quite closely tracks the curve showing proportion recalled. (For this analysis, we don't really care about the items at the end of the list, since we have a different explanation, namely recall rom short-term memory, to explain the recency effect, an explanation that is entirely independent of rehearsal time.)

events in a composite task

Time flows from the bottom of the slide to the top. First a fixation point is shown, followed by a photograph of an object. (Now we are using Ziggerins, another class of objects invented by Isabelle Gauthier.) Then a mask is used to wipe out after images of the object and is followed by a cue. (Here the bracket at the top of the mask indicates that the judgment is to be on the top part of the stimulus.) Finally, a test figure is shown, with the two halves either together (left hand figures) or offset from one another (right hand figures). In this figure, two different trials are shown: (a) and (b). (a) is a trial from the aligned condition; (b) is a trial from the offset condition.

Are Faces special? experiment explained

To get at this question, they invented these greeble figures, examples of which are shown here. (Greebles were meant to be non-face like, but they sure remind me of faces.) At the beginning of the experiment, they scanned the brains of participants while they viewed greebles or faces. They found much greater activity in the fusiform gyrus when the participants were looking at faces than when they were looking at greebles (See the left hand pair of bars in the graph.). They then gave the participants extensive training (multiple days) recognizing and naming greebles. The intent of this training was to make the people experts at identifying greebles. After the training, they again scanned their participants' brains. Now they found about the same amount of activation when viewing greebles as when viewing faces, a result predicted by the hypothesis that the fusiform gyrus is involved in the visual processing of things we are expert at identifying, and inconsistent with the hypothesis that the fusiform gyrus is devoted specifically to face recognition.

Notes

To summarize, the shape of the serial position curve can be reasonably explained by hypothesizing that the primacy effect is due to recall from long-term memory and the recency effect is due to recall from short-term memory. Implicit in that explanation is the assumption that short- and long-term memory are in fact separate and distinct things. It was experiments and distinctions like the ones we have just gone through that led early memory researchers to 1) formulate the distinction between short and long-term memory and to 2) formulate the modal model of memory, first written down by Atkinson and Shiffrin (1968) and depicted in the figure above. The modal model divides memory into different structural stores, with various processes operating within those stores. The three stores are sensory memory (or sensory registers), short-term memory, and long-term memory. We have multiple sensory memories, one per major sense. The idea is that we store for an extremely brief period of time (less than a second) a copy of the image impinging on our senses in order to give the brain a chance to process it. There is no interpretation of the data in that image, it's just a raw copy. Because there is no interpretation of what the image is, there is no information about meaning in the sensory store. These days, the sensory stores are considered more as perseverance of the neuronal signal in the sensory systems than as a memory store. As such, they are more appropriate for study in a Sensation and Perception course, and we are not going to spend much time on them in this class. The other two stores are the short-term store and the long-term store. As already mentioned, short-term store is of limited duration (less than 30 seconds) and limited capacity (7 +/- 2 items in the best of all possible worlds). It can be thought of as the information that we are currently conscious of. The long-term store is our hard drive. It contains everything we know. And it contains it on a more or less permanent basis. When we need to work with information, we need to retrieve it from long-term store and get it into short-term store. We do this by generating various retrieval cues (not necessarily consciously and often supplied by the environment) that we use to locate the information in long-term store. We are not always successful, but this does not mean that we have forgotten the information, in the sense that it is gone from long-term store. We may simply be using inappropriate retrieval cues, something we will be discussing in the next unit. A later retrieval attempt, using different cues, may result in success.

Where does the information that's in working memory come from?

Typically, it comes from one of two sources: the environment or from long-term memory. My GPS system telling me to turn right in 400 feet is an example of the former. I'm going to hold that piece of information in working memory so that I don't forget to make my turn. As an example of the latter, information coming from long-term memory, suppose that I am adding 29 and 42 in my head. I retrieve from long-term memory several things to make this happen and hold and use them in working memory, including a) the rules of addition, b) the fact that the sum of 9 + 2 is 11, and c) the fact that the sum of 2 + 4 is 6.

Visual Dominance

Visual dominance is the idea that when our senses receive conflicting evidence, as humans we have a tendency to believe the visual (Other species might show other senses to be dominant.). Suppose we did the following experiment, illustrated in the above diagram. We set up a row of lights, with a loudspeaker in front of each. We place an observer at the center of the row. Now we simultaneously light up the third light from the right while playing a sound over the loudspeaker in front of the central light. We ask our observer where the sound came from. Almost invariably, the observer will point to the loudspeaker in front of the light that was lit up. This phenomenon is known as visual capture. They do not point to the loudspeaker where the sound actually came from, and they do not point to the loudspeaker halfway between the light and the right loudspeaker. They point to the loudspeaker by the light that was lit. It is as if the visual stimulus completely dominates the auditory one.

What did Wo see

WO saw (or at least claimed to see) 2's as orange, 5's as green, and both 6's and 8's as the same shade of blue. For most people, 5's should be very hard to find in a background of 2's since they share a lot of features. Similarly, 8's should be hard to find in a background of 6's because they too share a lot of features. WO was asked to search for a (black) 5 in a background of black 2's (and vice-versa) and for a black 8 in a background of black 6's. What if WO really sees the colors that he claims to see? Well, the 2's should appear orange and the 5's as green and pop-out should occur—he should have no trouble finding 5's amongst 2's. But both 8's and 6's appear as blue, so no pop-out would occur.

Wo's experiment

WO's performance was more error prone and much slower on incongruent than congruent trials, suggesting that he really was perceiving the colors that he claimed to perceive.The experiment is summarized here:• Took advantage of WO's word synesthesia• Task: Name real color of ink a word is printed in• Congruent trials: real ink matched WO's synesthetic color- moose death• Incongruent trials: real ink mismatched WO's synesthetic color- moose death• Performance disrupted on incongruent trials

What else do we need with storing information

We also need a way of maintaining it in our working memory until we no longer need it—I need to remember that phone number until I find my phone and can dial it. I need to maintain that image of my apartment until I get the windows counted. We do this through by rehearsing the information—i.e., via maintenance rehearsal. In the case of the phonological store, maintenance rehearsal takes the form of sub-vocal speech—we repeat the information to ourself. We refer to this as the articulatory loop.

Can we influence the control process

We can discover them, invent them, be taught them, modify them based on experience. And the ones we use at any given time we choose based on our strategic purposes. If we are trying to remember the number to Aver's pizza long enough to dial it we are going to choose a different control process than if we are trying to remember it forever and forever. And if we are trying to multiply two two-digit numbers in our head we are going to use another set of control processes all together. Hopefully, from this discussion, it will be apparent that short-term memory is really much more than a store. It's where we perform a lot of cognitive activity. For these reasons, it is more frequently referred to as Working Memory these days. Some of these points are summarized in the chart below.

Holistic processing

We can think of holistic processing as an inability to selectively attend to or selectively ignore individuals parts of a face. Perhaps a better way of saying it is that we cannot process one part of a face without processing every other part of the face. I cannot process the nose without also processing the eyes. I cannot "see" the nose without also "seeing" the eyes. Holistic processing is also meant to imply that we process the image as a whole, we do not break it down into component parts, such as geons or features. In this sense, holistic processing is what templates do for letter recognition and what image-based models do for object recognition.

Do Synesthetes Really Perceive Those Colors?

We should be skeptical about synesthetes claims concerning their perceptual experiences. How do we know that a synesthete really experiences red when seeing the numeral 2 (Such a person would be termed a Projector), and isn't simply thinking of red rather than directly experiencing it (i.e., has a strong association between the numeral 2 and the color red; Such a person would be called an Associator.), or isn't even simply faking the whole thing? The next several paragraphs talk about ways people have gone about determining whether that perceptual experience is real, particularly for graphemic synesthetes, the most extensively studied.The initial studies that laid the ground work of addressing the question of whether synesthetes really perceive those colors that they claim to perceive involved a synesthete identified as WO. WO is a graphemic synesthete who experiences letters and numbers as having a particular color, regardless of the color of print the letter or number is presented in. Sekuler and colleagues at Vanderbilt University have done some extensive testing with WO to determine if his synesthetic perceptions are real.WO was presented with 100 of the words for which he claimed to see an associated color, and simply named the color that he saw. In a re-test a month later, he was given the same task. His responses on the second test were in agreement with his responses on the first test 97% of the time. (In contrast, the responses of non-synesthetic normals agreed with their earlier responses only 43% of the time.) It seems unlikely that WO could remember over a month's period the particular color he named as associated with a particular word. On the other hand, if he is really seeing the color, then a high degree of consistency from test to re-test would be expected, because he doesn't have to remember the color, he simply has to perceive it on both test and re-test.Before concluding, though, that WO's color perception is real, we might want some converging evidence. After all, there are people with phenomenal memories for certain things (witness Dustin Hoffman's character in Rainman) and perhaps WO's "gift" is not synesthesia, but an extremely good memory. He simply remembered all his responses to those 100 words from the first experimental session. Or perhaps he doesn't really see the color when shown the word; he just has such a strong association between the word and color (i.e., he's an associator, not a projector) that whenever he is shown the word, he thinks of the color. (Some people, like myself, question whether associators are true synesthetes.) That too would lead to a high degree of agreement between the two tests.To test WO further, a version of the Stroop task was used. Recall that the Stroop task requires naming the color of ink a word is printed in. If that word is a color name different than the color of the ink, then interference results and responding is labored and slow. Here, I've printed the word GREEN in red ink. In this case, if this were a stimulus in a Stroop task, we would need to say "red," since the word is printed in red, but since the word is the word green, we get interference which creates disfluencies when we try to pronounce the word "red."WO was given a task where he had to name the color of ink a word was printed in. Some of the words were words for which WO simultaneously experienced a synesthetic color as well as the real color. The word "moose," for example, WO experienced as pink. The word "death" he experienced as green. So, in the Stroop task, some trials involved words that were printed in a color congruent with WO's claimed synesthetic experience (Moose was printed in pink and death in green) and some trials were presented in a color incongruent with WO's claimed synesthetic experience (moose was printed in green and death in pink).

Neurological Evidence

We'll turn now to the neurophysiological evidence for a distinction between long-term and short-term memory systems. This evidence is a bit more convincing for separate stores. The evidence takes the form of a double dissociation between short-term memory and long-term memory deficits in brain-damaged individuals. As explained in Unit 02, a double dissociation is usually interpreted as evidence of separate and independent systems.

Multi-Sensory Integration

We've been talking primarily about visual perception, but, as you know, we have five senses (Actually, it's more, but who's counting?). We don't use just one, we use them all to figure out what is going on in our world. In this unit, we are going to be talking a bit about phenomena that involve integrating information across multiple senses. We will start by talking about visual dominance. The idea here is that although we do have five senses, vision tends to be our dominant sense. Put vision in conflict with one of our other senses and we will tend to believe our eyes. We will then talk about audio-visual speech, an exception to that general rule. In audio-visual speech, we integrate information from both our sense of hearing and our sense of vision to extract more than the sum of the information that we get from each of those senses alone. Finally, we will talk about synesthesia, a condition where stimulation in one sense can evoke quite vivid and real sensations in another sense, even when there is no physical stimulus for that second sense.

Part-Whole Effect in Experts

What about the part-whole effect? Again, experts show the part-whole effect in their area of expertise. Car experts show a part-whole effect for cars. Non-car-experts do not show such an effect. People trained in the laboratory to become Greeble and Ziggerin (We'll meet some Ziggerins in a bit) experts show a part-whole effect when judging Greebles and Ziggerins. Non-greeble and non-ziggerin experts do not show such an effect. Molecular biologists show a part-whole effect for photographs of microscopic slides of biological cells. These results suggest that to the extent that the part-whole effect indicates holistic processing, then we do holistic processing for things that we are expert in, that the part-whole effect and holistic processing are not unique or special to faces. So, the part-whole effect seems to be indicative more of processing in an area of expertise than to be unique to faces.

Just to summarize the results again

With faces, participants are faster and more accurate at making judgments in the offset condition than in the together condition. With non-face objects, no such effect occurs—people are about equally fast in the offset and aligned condition.

What did Martha Farah has observed

a double dissociation between face recognition and object recognition. Some patients, with damage to the fusiform gyrus, show a deficit in face recognition but not object recognition. Other patients, with damage to other areas of the temporal lobe, show a deficit in object recognition, but not in face recognition. This result suggests to researchers that face recognition and object recognition are implemented as independent systems in the brain. Some researchers would express this by saying that we have different brain modules for face and object recognition.

When looking at faces there is

a lot of activity in the FFA, or Fusiform Face Area, another name for the fusifrom gyrus.

the at least two stores of working memory

a phonological store (for holding and manipulating verbal or linguistic information; phonological refers to the sounds used in speech) and what has been called the visualspatial sketchpad, for holding and manipulating visual information. (There may be other stores as well, roughly corresponding to the other sensory modalities; the visual and phonological, however, are the most widely studied.)

What does each store include

a rehearsal mechanism, referred to as the visual scribe in the case of the visual spatial sketchpad and the articulatory loop in the case of the phonological store.

Difference in Subtraction Technique

activity due to active manipulation

What is the visual scribe

allows us to maintain information in the sketchpad by basically refreshing it, just as the picture on your television screen is refreshed 32 times per second. You can think of the scribe as redrawing the visual image in your head before it fades away.

Row of ziggerins suggest a

class and thus they all look quite similar

What is the rehearsal for visual working memory

essentially refreshes the picture we have in our "mind's eye." We sometimes use the term the Visual Scribe to refer to the mechanism by which we do this.

Serial Position Effect

et's start by considering a simple experiment. In this experiment, participants are shown (or hear; it doesn't matter.) a list of words, one word at a time. After the last word in the list is presented, they are asked to recall all of the words in the list. They might be allowed to recall the words in any order, in which we case we refer to the task as free recall, or they might be required to recall the words in the same order as they were presented, in which case we refer to the task as serial recall. Since participants were asked to recall the words immediately (or nearly immediately) after the presentation of the items, we often describe these tasks as immediate free recall or immediate serial recall. If we introduced a significant delay, say an hour, or a day, or a week, and so on, between presentation of the to-be-remembered items and the recall of them, we would refer to the task as delayed free recall or delayed serial recall. You should remember these terms. Before talking about the results of this experiment, we should take an aside that we are going to have to take sometime, so we might just as well take it now. The aside is into the concept of ecological validity, another term that you will want to remember. In brief, an experiment or study or whatever is ecologically valid to the extent that it involves participants doing things that they would do in the real world. The above experiment uses a method that is very, very common in memory research—participants memorize a list of words. There have iterally been thousands of experiments done and published using this technique. But, it should be fairly obvious that such an experiment is not very ecologically valid. In the real world, people do not spend their time memorizing list of words (especially if we overlook the occasional memorization of a grocery list). Why then have so many experiments been done using this technique? (Before answering that question, we might point out that it is perhaps possible to build an argument that memorizing list of words does in fact possess at least a degree of ecological validity. True, people do not memorize list of words in the real world, but they do memorize things. And the things that they do memorize, for example, facts learned in school courses, are often verbal in nature—that is they can be stated in words. So, the laboratory experiments, like much of the real world, are at least using verbal materials.)

fMRI imaging studies, using the subtraction technique, have found enhanced activity in the

fusiform gyrus when people are looking at pictures of faces versus looking at a variety of different types of control pictures.

What is synesthesia related to

higher levels of neural activity between the two areas of the brain responsible for the individual perceptions involved. For example, synesthesia involving perception of a color when hearing a particular musical note is associated with increased activity between the area involved in auditory perception (particularly frequency perception) and the area involved with color perception. What is not known and is still being debated is why this increased activity occurs. It could be there is simply a greater number of neural connections between the two involved areas. Or it could be that synesthetes have the same number of connections between these regions as other people, but that there is less inhibition operating on these connections.

face recognition involves _______ processing.

holistic

part-whole effect

identifying a face based on seeing just a part of it is much more difficult than identifying an object based on seeing just a part of it.

The consideration with listing the states alphabetically and multiplying two big numbers suggest

in addition to a visual and verbal working memory, we also need a boss, some sort of decision maker, who decides how we are going to go about performing the various tasks that we are using working memory to do, what information is going to be stored visually, what information is going to be stored verbally, what information to keep and what information to throw away. We call that boss the central executive.

Prosopagnosia

is a condition where people have difficulty or even an inability to recognize faces, even of people very familiar to them, such as their spouse. That difficulty is typically not also associated with a difficulty recognizing objects. At least some cases of prosopagnosia are caused by or at least correlated with damage to the temporal lobe and in particular damage to the fusiform gyrus. That observation is also consistent with the hypothesis that the fusiform gyrus is a face recognition module. It is the inability to recognize faces

Short-Term memory

is simply a place to temporarily bold information Besides is is not just a place for storing information, it is also a place for manipulating information, for working on information. It's where we think and reason (to the extent that we ever do in fact think or reason). Hence, we generally now use the more general term "working memory." In addition, we also are concerned as much with visual working memory—the short-term storage and manipulation of visual images—as we are with verbal working memory—the short-term storage and manipulation of linguistic information. Some researchers do try to distinguish short-term memory from working memory. For our purposes, we will treat these terms synonymously and use them interchangeably.

How many individual items can we hold in our verbal short-term memory

it depends on how well we can organize them into chunks. If we are good chunkers we can hold many more individual pieces of information than if we are lousy chunkers. As another example of chunking, consider the following sequence of letters: • FB ITW AC IAIB M That string of letters is hard to immediately recall, because we see no organization to it and are unable to chunk it into higher level units. Now consider the following string: • FBI TWA CIA IBM The second string directly provides the chunks. Instead of having to remember 12 letters (well over our 7 +/-2 capacity), we have to remember only 4 well known acronyms (well within our 7 +/- 2 capacity), at least for those old of us to remember what TWA was (It was an airline. It stood for Trans World Airways. Those of us who had the pleasure of flying it overseas fondly remember it as Try Walking Across.) When we chunk like this, holding all 12 letters in short-term memory is not much of a problem, because we are really holding only 4 items, not 12.The rate and rhythm that we recite something to ourselves can also help us chunk arbitrary information. Saying "8 1 2 3 3 2 5 9 4 7" doesn't help us chunk much. Saying "8 1 2 <pause> 3 3 2 <pause> 5 9 4 7" naturally leads us to chunk it into two three digit numbers and one four digit number, making it easier to remember.

What do these three findings show

it is evidence, that although objects are processed or recognized analytically, faces are processed holistically, a distinction quite similar to that between structural description theories of object recognition and image-based models of object recognition. That difference has in turn led people to claim in the literature that processing of faces is special, it is unique to faces.

Difficulty in Subtraction Techniques

it is not always obvious what appropriate control task

Analytical processing

means that we break the object down into its component parts in order to recognize it. This is what the feature net model of letter recognition does—it breaks letters down into their features (veritical lines, oblique lines, and so on). This is what structural description theories, including Biederman's Recognition by Components, do for objects —they break them down into their component parts, their geons, and recognize them by determining which geons they contain.

Why have so many experiments been done involving memorizing list of words?

more than likely has to do with convenience and do-ability. What people do memorize in the real world is things that they learn in school as well as episodes or experiences from their own lives (They may not actually actively go about memorizing such experiences, but they certainly do remember them.). Mimicking such situations in the laboratory would be prohibitively expensive. It is relatively easy to bring 100 people in the laboratory and have them each memorize a list or two of words. It's not so easy to bring 100 people in the laboratory and give them a whole course in organic chemistry. Or to give them a lifetime of experiences in one area of life or the other. And then test their memory on organic chemistry or on those life experiences. Words, on the other hand, are easy to use. Psychologists know a lot about words—how long each one is, approximately how frequently each gets used in the language, and so on—so it is relatively easy to control for different aspects of the words used in our experiments. And the recall of words is easy to score—the person either recalls the word or does not. When recalling real life experiences, it can be a lot more difficult to determine when recall is correct and when it is not. People will no doubt leave out some details when recalling real events from their life, but is that because they don't remember them, or is it because they simply judge them as unimportant? So, we use words out of convenience. At the same time, though, we don't want to waste our time studying something if the results do not apply to the real world. So, every once in awhile, we need to stop, take a deep breath, and ask ourselves whether the principles that we are learning from laboratory experiments do in fact apply to the real world.

Stimulation in Subtraction Technique

participant manipulates object

what is the control in the Subtraction Technique

participants holds object in hand

As you can see, for the sentence task,

participants were a bit slower when speaking than when pointing.

Here, if the verbal and visual stores are separate, we would expect

performance to be worse when participants had to speak versus when they had to point, since both that mode of responding (saying "yes" or "no") and the memory task are drawing on resources in phonological working memory.

Inverted Faces Effect

reaction times increase much more for inverted as compared to upright faces than for inverted as compared to upright objects; It says that turning a face upside will result in a greater increase in difficulty for the face than turning an object upside down.

Some of the control conditions have included pictures of

scrambled faces, of everyday objects, of hands, of more complete bodies but without the face being seen. None of these control conditions led to increased activity in the fusiform gyrus. This finding has led a number of researchers to hypothesize that the fusiform gyrus implements a face recognition module in human beings, that the fusiform gyrus is specialized for the recognition of faces. They often refer to the fusiform gyrus as the fusiform face area of FFA.

Having experienced damage to the fusiform gyrus what happens

show a deficit in face recognition but not object recognition.

Having experienced damage to the other areas of the temporal lobe

show a deficit in object recognition, but not in face recognition.

claims require extraordinary evidence

so we should look for even more evidence that synesthetes really do experience the colors they claim to experience. We next describe another common method that, along with variants (see the Ramachandran video from the links above), has been used to test the reality of graphemic synesthetes' color perception. This technique takes advantage of the phenomenon known as pop-out, described below.First, suppose that we had someone (a non-synesthete) search for the letter Q in a background of a bunch of O's. That is a relatively difficult task and people are relatively slow at it. Furthermore, as we increase the number of O's in the background, their reaction times become longer. Now suppose we print those O's in blue and the target Q in a different color, like red. Now, the Q just pops right out among the O's. Finding the Q becomes trivially simple. In addition, reaction times for finding the Q no longer increase as we increase the number of O's in the background.WO saw (or at least claimed to see) 2's as orange, 5's as green, and both 6's and 8's as the same shade of blue. For most people, 5's should be very hard to find in a background of 2's since they share a lot of features. Similarly, 8's should be hard to find in a background of 6's because they too share a lot of features. WO was asked to search for a (black) 5 in a background of black 2's (and vice-versa) and for a black 8 in a background of black 6's. What if WO really sees the colors that he claims to see? Well, the 2's should appear orange and the 5's as green and pop-out should occur—he should have no trouble finding 5's amongst 2's. But both 8's and 6's appear as blue, so no pop-out would occur.

As you can see, for the sentence task

speaking response was slower than pointing

Working memory (WM) refers to

temporary store for information that we currently need active in order to achieve some immediate goal or complete some immediate cognitive task. We used to use the term short-term memory, or STM.

at least when we are dealing with letters and words, phonologically similar items are confusable in short-term memory; visually similar items are not. This general result indicates

that people tend to encode those items phonologically in short-term memory, i.e., that we have a bias towards phonologically encoding, even when the presentation is visual. In the case of a visual presentation, we are not simply copying the stimulus over into our visual short-term memory, but instead we first take the trouble to convert it to a phonological code and then store it in our verbal short-term memory.

Capacity limits how much we can attend to at any given time; how is it alleviated

the fact that we have different capacities for different modalities; hence we may be able to attend to a fairly large amount of information without suffering a performance decrement, provided that information is spread across different modalities. Even in this case, though, given that just about every interesting cognitive task also requires the resources of the central executive, we are likely to meet up with the capacity limitation of the central executive

Though reduced in magnitude, participants

the opposite pattern occurred participants spoke their response—they were slower on the sentence memory task than the visual memory task. Hence, a visual responding mode interferes more with visual memory than it does with verbal memory. And a phonological responding mode interferes more with verbal memory than it does with visual memory. This pattern of results is that expected if in fact the phonological and visual working memory stores are separate from one another.

In the case of the visual memory task

the opposite pattern occurred. Participants were much slower when pointing than when speaking. Overall, when pointing, participants were much slower on the visual task than on the verbal task.

articulatory loop

the part of the phonological loop involved in the active refreshing of information in the phonological store by repeating information back to ourselves

What is a common way of analyzing PET and fMRI data is

the subtraction technique

Performance on the Composite Task

the two faces aligned or unaligned For faces, we are actually better at doing this task in the offset condition than in the aligned condition. For objects, we are about equally good at doing this in the two conditions. Again, for faces, performance, as measured by either reaction time or percent correct, is better in the offset condition than in the aligned condition. For objects performance is about the same in the two conditions. Even for Greebles It is like when he showed those faces of tom cruise and that chef Ramsey; it was easier to identify them sperately when their faces were not aligned rather than when they were

McGurk Effect

the visual stimulus is of a person saying /ga/. If vision dominated audition in this case, then we would perceive the syllable /ga/ being spoken. But we rarely if ever do. The auditory stimulus is a /ba/. If audition dominated, then we would hear /ba/. In fact, some people (about half) do hear /ba/. The other half of the people, though, hear /da/. The syllable /da/ is acoustically and visually half-way between a /ba/ and a /ga/. It is as if we blend the visual information together with the auditory information, and come up with something half-way between, the syllable /da/. We may rely on visual cues for understanding speech more than is commonly thought, especially in noisy environments. Suppose that we have to identify spoken words, without seeing the person who is speaking. So long as there is no noise, we are very good at this, near 100%. As we add noise, though, and increase its volume, performance gets worse and worse.

Why does the primacy effect occur

we have more opportunity to rehearse, and consequently commit to long-term memory, items that occur at the beginning of the list as opposed to items that occur later in the list. The recency effect occurs because we recall the last few items in the list directly from short-term memory, without ever having to even get them into ong-term memory. Because short-term memory is of limited capacity, we can only get the last few items this way, and not the entire list.

What is the central executive

weeding out useless information and placing information where it needs to be

working memory is equal to conciousness?

what we are conscious of is what's in working memory, and what's in working memory is what we are conscious of. I would agree that what we are conscious of is what's in working memory, but would not agree that we are conscious of everything that is in working memory. We can hold information in working memory, perhaps even use it, without being conscious of it.

central executive

which has a number of not entirely well-defined responsibilities. The executive determines what information is going to be held in the sketchpad and what information is going to be held in the phonological store. If information in one needs to be translated into the format of the other (e.g., when we create a mental map based on driving directions given verbally), the central executive is responsible for that translation. The executive is also responsible for determining when information is needed from long-term memory, retrieving that information, and placing it in the proper store. It is responsible for determining how any (non-automatic) cognitive task is going to be approached. In the case of endogenous attention, it determines what we allocate our attention to. When we are doing multiple things at once, the central executive decides when to switch attention from one mental activity to another. It also is responsible for inhibiting responding to irrelevant information, a function that recently has begun receiving a lot of attention from neuroscientists.

Neurological Evidence for a Face Module

with the front part of the brain at the lower part of the figure and the back part at the upper part of the figure. Buried deep within the temporal lobe is a bilateral structure known as the fusiform gyrus. It is close to but not quite at the midline.

Articulatory Suppression

• "the, the, the,..." • Prevents encoding visual stimuli into speech-code - Should eliminate effect of phonological similarity for visually presented items • Prevents or greatly reduces opportunity for(verbal) rehearsal - Should harm recall of auditory stimuli

Patient H. M.

• Bilateral brain surgery to control severe epilepsy • Much of medial temporal region of brain removed, including the hippocampus • STM remained intact - Could recall series of numbers just presented - Could converse normally - Recency, but not primacy, in free recall • Long term memory for events prior to surgery was intact • Performed well on standard IQ tests • Could learn new perceptual and motor skills • But could not form new long term memories - Everyday he spoke with his doctors and everyday it was like making new friends - Everyday he re-discovered current events To summarize, in H. M., we have a patient with intact short-term memory but impaired long-term memory, at least for new events. So far, that is a single-dissociation.

How Do We Know the Phonological Store and Visualspatial Sketchpad are Separate?

• Couldn't they be the same store, with just two ways of representing the information? • Interference Studies • Does visual processing interfere with verbal processing(in STM)? • Does verbal processing interfere with visual processing? • If yes --> common store - i. e., verbal should interfere as much with visual as it does with other verbal tasks • If no --> separate store - Visual should interfere with visual and verbal with

Working Memory and Chess MemoryResults

• Counting did not interfere with remembering chess positions • Tapping and generating random letters did interfere • Suggests separate verbal and visual stores - And that central executive also

Modal Model

• Distinguishes structure from processes - Emphasis was initially on structure • Structural components - Fixed, universal • Processing components - Frequently under control of person. • Hence the name: control processes - Can be learned, discovered, invented, and so on - Selection of control processes determined by strategic interests of person • E.g., elaborative vs. maintenance rehearsal • To get to LTM, information must go through STM - (This does not seem to always be the case)

Encoding Ambiguous Figure

• If figure "translated" to a name (either duck or rabbit) then it should not be possible to reverse the figure in working memory • If figure coded visually, then it should be possible to reverse the figure in working memory • During brief presentation of reversible figure, ½participants were silent, ½ said "la, la, la..." • Tested children and adults - Children tend to be more visual and less ve

Summary

• It can take several days for a memory to become "permanent" - Longer than the duration of working memory - Consolidation - Possible role of sleep • During deep sleep (i.e., non-REM [Rapid Eye Movement] sleep, memories are transferred from hippocampus to distributed neo-cortical networks • Hippocampus may be storage

Characteristics of STM

• Limited capacity • Limited duration • Rehearsal (maintenance) required to retain information • Preferred coding is phonetic - At least for adults - Rehearsal (elaborative) required to copy information into LTM

Phonological Similarity

• List of items that are phonologically similar harder to recall • When we make errors, they sound like (but don't look like) their targets • Effect occurs for visually presented as well as auditorily presented items - Because visually presented items are xlated to speech-code • Responsibility of articulatory control process • Can we prevent xlating visual items into speech-code? - by keeping busy the articulatory control process - How can we do that? • Repeating single word over and over continuously - Shouldn't that eliminate phonological similarity effect for visually presented items?

The chart below provides some background on the N-back task, a task commonly used to study executive control.; some background on N-Back

• Participant shown a sequence of stimuli- E.g., a sequence of digits • On each trial, participant answers question: Is current stimulus same as stimulus from n trials previously • N = 1: Easy • N = 2: Difficult but doable• N = 3: pure torture

Manipulating the Serial Position Curve

• Preventing the use of STM - Should reduce recency effect - Delay recall while preventing rehearsal• Counting backwards by 3's • Reducing probability of transfer to LTM - Reducing opportunity for elaborative rehearsal - Should reduce primacy effect

Patient K. F.

• Severely impaired STM - Digit span of 2 - Normal is 5 - 9 - No recency effect in free-recall • Normal long-term memory - Even for information learned post-injury - Primacy effect in free-recall

Clive Wearing

• Viral encephalitis • Remembered little of his previous life • Was able to form no new memories • Short-term memory intact - Could, for example, carry on a conversation