Cognitive Psychology Exam 3

Kosslyn's comparing imagery and perception

Items that take up larger parts of our visual field show more details. If showed participants picture of elephant and rabbit, and asked about rabbit's whiskers, were much quicker to respond if rabbit was next to a fly instead, since the rabbit took up more of the imagery. Did a mental walk test, where they were to image that they were walking toward their mental image of an animal. Had to estimate how far away they were from the animal when they experienced "overflow," when the image filled the visual field or when its edges started to become fuzzy. Participants tended to move closer for small animals than four big ones, as in reality.

Why is the use of categories so important for our day-to-day functioning?

Categories give us a lot of information about places that we would otherwise not know much about. Example, unfamiliar cat.

Neuropsychological evidence for dissociations between imagery and perception

Dissociations:Guariglia-studied a patient whose brain damage had little effect on his ability to perceive but caused neglect on his mental images (limited to just one side). Another, R.M., had damage to occipital an dparietal lobes. Able to recognize objects and to draw accurate pictures of objects that were placed before him, but unable to draw from memory. C.K. had visual agnosia, unable to visually recognize objects, but was able to draw images from memory. When shown his own drawings, had forgotten the actual drawing experience and couldn't identify results.

What are some of the properties of a good psychological theory? How have these properties been applied to semantic network theories?

1. Explanatory power-can explain why a particular result occurred 2. Predictive power-can predict the results of a particular experiment 3. Falsifiability-should be possible to design an experiment that can potentially yield results that would be predicted by the theory, and that are NOT predicted by theory 4. Generation of experiments-stimulate a lot of research Collins and Quillian does explain and predict some results, but there are many results it can't explain. Collins and Loftus's is so flexible difficult to falsify (networks can be modified, fire engine paths can be shorter for some people than for others. No definite rules).

What is the exemplar approach to categorization? How does it differ from the prototype approach, and how might the two approaches work together?

Also determines whether an object is similar to a standard object, but instead of focusing on the "average," it involves many examples called "exemplar." A robin, sparrow, and blue jay are all exemplars of birds. Explains many of Rosch's results. Explains typicality effect (reaction times in sentence verification are faster for better category examples) by proposing that objects that are like more of the exemplars are classified faster. Exemplar good at classifying individual cases, reminds that there is diversity (Easier to classify a penguin because we can remember there are flightless birds). Also good for classifying games, since there isn't an average. Research suggests using both by using aver exemplars to find prototype, and then learning some of the exemplar information. Exemplar for small categories, prototype for larger ones.

Are the same parts of the brain activated during imagery and perception?

Amedi showed that some fMRI scans show overlap for perception and imagery, but some areas with nonvisual areas such as hearing and touch were deactivated. During imagery, their activation was increased. Suggests reason for this might be that visual mental images are more fragile than real perception, so this deactivation helps quiet down irrelevant activity that might interfere with the mental image. But also has a lot of overlap, which shows that imagery and perception share some mechanisms.

What is a connectionist network? Describe how a connectionist network learns, considering specifically how connection weights are adjusted.

Approach to creating computer models for representing concepts and their properties based on characteristics of the brain. Propose that concepts are represented by activity that is distributed across a network. Circles represent units, which are inspired by neurons in the brain, and they are connected by lines that act like axons. Units can be activated by signals from environment or other neurons. Connection weights determine how signals sent from one unit either increase or decrease activity of the next unit. Correspond what happens at a synapse that transmits signals from one neuron to another. Activation of units depends on 1. the signal that originates in the input units and 2. the connection weights throughout the network. Stimulus presented to the input units is represented by the pattern of activity that is distributed across the other units.

How did Kosslyn use the technique of mental scanning (in the boat and island experiments) to demonstrate similarities between perception and imagery?

Asked participants to memory a picture of an object, such as a boat, and then to create an image of that object in their mind and to focus on one part of the boat, such as the motor, and to press the "true" button when they found this part or the "false" button when they couldn't find it. Reasons that if imagery is spatial like perception, then it should take longer for participants to find items that are farther from the focal point. This is what happened.

Also consider how the way information is represented in a connectionist network differs from the way it is represented in a semantic network.

Both are dealing with the same concepts (like specific), and properties of the concepts are indicated in both networks by the following four relation statements "is a" (Canary is a bird), "is" (canary is yellow), "can" (A canary can fly), and "has" (a canary has wings). BUT hierarchical network indicates these properties are the network's nodes, and the connectionist network indicates these properties by activity in the property units.

How are categories represented in the brain? Describe evidence from single neuron recording in monkeys, the effects of brain damage in humans, and human brain imaging.

Brain representations are usually distributed. FFA recognizes faces, and PPA recognizes houses, but more parts of the brain are involved. Objects consist of many proerties (Cats have color, texture, form, emotional responses, motor skills, etc.) Brain has to take in all of that. Single neurons: Freedman trained monkeys to classify stimuli like "cat" and "dog." Showed them morphing into each other. Had them respond to which one was more than 505% cat o and the rest were dogs. Once learned to recognize, monkey's released a lever if judged the test stimulus to be in same category as a sample stimulus (a dog or a cat), recorded from nuerons in an area of temporal lobe called inferotemporal cortex. and prefrontal cortex. Response of IT neuron 60% of the time for cat and dog stimuli. Showed that different parts of brain respond to different stimuli. Brain damage in humans: category-specific knowledge impairment, patent has trouble recognizing objects in one category. Patients have trouble naming animals, but can name nonliving things. Had trouble telling difference between animals but also in knowing the properties of animals. Brain scanning: Differences between living and nonliving things, activate different parts of brain. Areas activated by animals also respond well to motion associated with animals. Areas activated by tools respond well to motion associated with tools. Evidence supports distributed processing. Simmons showed observers pictures of food, and activated visual cortex with food's shape and other areas associated with taste, as well as amygdala, and prefrontal cortex for how appealing it was.

Describe definition approach to categories. Why does it initially seem like a good way of thinking about categories, but then become troublesome when we consider the kinds of objects that can make up a category?

Can determine if something is a member of a category by determining whether a particular object meet the definition of the category. Can work for somethings, but not for others: Good for geometry, bad for describing chair. "Family resemblance" is better, means that definitions do not include all members of a category, for example "games" can be board games or olympic games. Determines how similar an object is to some standard representation of a category.

Is imagery just a "laboratory phenomenon," or does it occur in real life?

Effects thinking, memory, and perception. Engages others senses too, such as sight, sound, taste. People have it in dreams. Paul McCartney wrote songs that were in his head, scientist dreamed of the shape of Benzene, etc.

What additional evidence is needed to help settle the imagery debate, according to Farah?

Farah pointed out that it is difficult to rule out Pylyshyn's tactit knowledge explanation just on the basis of the results of behavioral experiments. Participants can always be influenced by past experiences. Suggests we stop looking solely at behavioral experiments, and instead investigate how the brain responds to visual imagery. Study brain damage, and use electrophysiological measurements, brain imaging.

Shepard and Metzler with mental chronometry

Had two images participants had to say if same object, but turned. Had to turn object in mind to determine if it was. Important because one of the first to apply quantitative methods to the study of imagery and to suggest that imagery and perception may share the same mechanisms.

Make a list of the important events in the history of the study of imagery in psychology, from the imageless thought debate of the 1800s to the studies of imagery that occurred early in the cognitive revolution in the 1960s.

Imageless thought debate, some psychologists supported Aristotle's idea that "thought is impossible without an image," while others thought that thinking can occur without images. Cognitive revolution, one key success was that cognitive psychologists developed ways to measure behavior that could be used to infer cognitive processes. Example, Ala Paivio linked behavior and cognition by working on memory. Showed that it was easier to remember concrete nouns like "truck" or "tree" that can be imagined, than abstract nouns like "truth" or "justice." Used paired-associate learning, where participants are presented with pairs of words, like "boat-hat" or "car-house." They are then presented during the test period with the first word of each pair. They have to remember the second word. Memory was better for pairs of concrete nouns. Proposed conceptual peg hypothesis: concrete nouns create images that other words can "hang onto."

How is the ability of young infants to form categories measured? Trace the development between two and seven months of infants' ability to categorize. What abilities are added after 7 months?

Infants can categorize even without language, starts at about 2 months. Familiarization/novelty preference procedure used to measure. Makes use of the fact that when given a choice between a familiar object and a novel one, infants will look longer at the novel object. Infants exposed to many examples in a category to familiarize (cats). Then preference test, where infants shown parts of the category that they haven't seen before, and an example from another category they don't know (New cat, and a dog). If infant looks longer at dog, has grouped cat with the others. Can have broad categories at two months. 3-4 months, begin forming basic level categories. Looks at cat longer than lion, because doesn't consider lion a cat. 6-7 months can form specific categories. If shown tabby and siamese cat, will look more at siamese cat. However, if Infants familiar with siamese cat and shown siamese and a tabby don't show preference for tabby. Categories get more fine-tuned with age.

Describe the spatial (or depictive) and propositional explanations of the mechanisms underlying imagery. How can the propositional explanation interpret the results of Kosslyn's boat and island image-scanning experiments?

Kosslyn supported spatial representation: a representation in which different parts of an image can be described as corresponding to specific locations in space. Has spatial layout, actually showing a cat under a table in a picture. Depictive representations Pylyshyn said experience imagery as spatial, but is not underlying representation. Said is epiphenomenon-something that accompanies the real mechanism but is not actually part of the mechanisms. Like lights in a computer, appear to be doing something, but if they blow out, the computer still works. Propositional representation-relationships can be represented by abstract symbols, such as an equation, or a statement such as "The cat is under the table." Said Kosslyn's boat experiment could be explained by words signifying parts of the boat, and the length of the lines indicates distance between the parts. The words in parentheses indicate the spatial relations between the parts. Participants respond as would happen if actually at the scene. They know it would take longer to get to a farther place.

Physiological explanation for connection between perception and imagery through brain deactivation

Kossyln and coworkers present transcranial magnetic stimulation to the visual area of the brain while participants were carrying out either a perception task or an imagery task. Could be asked to compare the length of stripes in pictures for perception, imagery task was to imagine stripes while answer questions, instead of looking at them. Measured reaction time, stimulation made both control and experimental group slower. Concluded that brain activation occurs in response to imagery. Not an epiphenomenon, activity in visual cortex plays a causal role in both perception and imagery.

Why were Kosslyn's experiments criticized, and how did Koslsyn answer Plyshyn's criticism with additional experiments?

Lea criticized and proposed that as participants scanned they may have encountered other interesting parts, such as the cabin, and this distraction may have increased their reaction time. Kosslyn did another experiment, asking participants to scan between two places on a map. Told to imagine an island that contained seven different locations. Scanned between every possible pair of locations, determined the relationship between reaction time and distance. Took longer to scan between greater distances. Supported his idea again, but Pylyshyn proposed imagery debate-a debate about whether imagery is based on spatial mechanisms like in perception, or is based on mechanisms related to language, propositional mechanisms.

What is the prototype approach? What experiments did Rosch do that demonstrated connections between prototypicality and behavior?

Membership in a category is determined by comparing the object to a prototype that represents the category. Prototype is "typical" member of the category. Rosch: "typical prototype is based on an average of members of a category that are commonly experienced. "Birds" might be based on some of the birds you usually see, but doesn't necessarily look exactly like a particular type of bird. Presented participants with category of "bird" or "furniture" and a list of about 50 members of each category. Participants asked to rate the extent to which members represented the category on 7 point scale, one being very good prototype. Sparrows are one, bats are later. Chair and couch one, mirror and telephone later. Rosch and Mervis showed strong relationship between family resemblance and prototypicality.

What is the evidence that solving mechanical problems can involve mental simulation? How is this evidence related to visual imagery?

Mental stimulation: operation of the mechanical system is mentally represented. Analogous with spatial representation. Gears example. Rule-based approach- applying a rule, such as "When one gear turns, the one next to it rotates in the opposite direction." Analogous with propositional representation. Two water glasses example.

What are some differences between imagery and perception? What have most psychologists concluded about the connection between imagery and perception?

Most psychologists who look at behavioral and physiological evidence believe that imagery and perception are closely related and share some, but not all, mechanisms. Perception occurs automatically, but imagery needs to be generated with some effort. Perception is stable, it continues as long as you are observing a stimulus; imagery is fragile, it can vanish without continued effort. It is harder to manipulate mental images than images that are created perceptually (Chalmers and Resiberg, asked participants to create mental images of ambiguous figures like a rabbit that can also be seen as a duck. Perceptually, easy to flip between the two, but participants who were holding a mental image had a hard time flipping. Finke showed that participants can manipulate simpler images, like turning the letter "D" around 90 degrees.).

Physiological explanation for connection between imagery and perception through brain imaging

PET and MRI, participants creating visual images or during a baseline condition in which created images.Subtracting baseline from activated response indicated which areas of brain were activated by imagery. LeeBihan demonstrated that both perception and imagery activate the visual cortex. Striate cortex activity increased both when person observed actual stimuli and imagined. In another experiment, asked participants to think about questions that involved imagery, generated a greater response of visual cortex than to nonimagery questions (Is grass greener than trees? or Is intensity of electrical current measured in amperes?" Ganis-used fMRI to measure activation under two conditions, perception and imagery. For perception, observed drawing of an object. For imagery, told to imagine a picture they had studied before when they heard a tone. Had to answer questions for both conditions "IS the object wider than it s tall?" Results show activation in three areas of the brain: frontal lobe for perception and imagery, more activation in back of the brain for perception for vision, almost complete overlap of activation caused by perception and imagery in the front of the brain, but some differences in back.

Some of the neuropsychological results demonstrate parallels between imagery and perception, and some results do not. How has Behrmann explained these contradictory results?

Paradox between imagery and perception sharing mechanisms, and having double dissociations (where after damage lose one and keep the other). Berhmann suggested that the mechanisms overlap only partially. Mechanism for perception is located at both lower and higher visual centers, and mechanism for imagery being located mainly in higher visual centers. Visual perception involves bottom-up processing, starts with light on eye and iamge focused in retina, then continues up signals to visual cortex and ten to higher visual centers. Sent to higher visual areas where perception is "assembled" and some top-down processing may be involved. Mental images are "preassemble," they do not depend on activation of cortical areas, such as visual cortex, because there is not input that needs to be processed.

How have experiments demonstrated interactions between imagery and perception?

Perky experiment: participants asked to "project" visual images of common objects onto a screen, and then to describe these images. Unbeknownst to the participants, Perky was back-projecting a very dim image of this object onto the screen. Participants descriptions of their images matched the images that Perky was projecting. None of them noticed there was an actual picture on the screen. Mistook actual image for a mental one. Modern research replicated the experiment. Farah did with letters, asked them to imagine a letter, then flashed two letters. The imagined letter was more quickly identified. Connection between perception and imagery.

Under what conditions does imagery improve memory? Describe techniques that use imagery as a tool to improve memory. What is the basic principle that underlies these techniques?

Placing images at locations: Greek poet Simonides. According to legend, presented an address at a banquet, and just after he left the roof of the hall collapsed, killing almost everyone. Bodies were so severely mutilated they couldn't be identified. But Simonides had created mental image of seating, and could identify them. Could remember by imagining a physical space, mental organization allowed him to "read out" the items. Method of loci-method in which things to be remembered are placed at different locations in a mental image of a spatial layout. Associating images with words: Pegword technique: involves imagery, but you associate them with concrete words. Create list of nouns (one-bun; two-shoe; three-tree; four-door...). Easy to remember words in order because they rhyme with the numbers. Rhyming produces a retrieval cue. Then pair each of these things to be remembered with each pegword by creating a vivid image of your item to be remembered with the object represented by the word. (Example, to remember to go to the gym, imagine an elliptical trainer inside of a shoe.)

What is the evidence for and against the Collins and Quillian model?

Positive: Hierarchical model is testable prediction for time it takes to retrieve information. Works with Sentence Verification Technique, takes longer to say canary is an animal than a bird. Spreading activation is activity that when stimulated alerts all of the associated nodes, such as canary stimulates "robin," an d"animal," makes them more readily available to remember. Influenced by priming, studied by Meyer and Schvaneveldt. Used lexical decision task. Participants were flashed two strings of letters, had to push "yes" if both were real words, and "no" if one or both weren't. Key variable was the association between the words, like "bread and wheat" and "chair and money" Reaction time was faster when words were associated. BUT, model doesn't explain typicality effect. Aso questioninst cognitive economy, because more likely to say "pig is animal" than "mammal," even though direct link between pig and mammal.

How did Collins and Lotfus modify the model to deal with criticisms of the original, and how were these modifications received by other researchers?

Proposed network where more closely associated connections have shorter lines. (Car and truck have shorter lines to vehicle than fire engine and ambulance.) Abandons hierarchical structure and bases this network on a person's experience.

What is the basic idea behind the semantic network approach? What is the goal of this approach, and how did the network created by Collins and Quillian accomplish this goal?

Proposes that concepts are arranged in networks. Ross and Quillian's goal was to create a computer model of the human memory. Network consisted of nodes connected by links. Each node represented a category or concept, and concepts are placed in the network so they are connected. Links show that these nodes are related to each other in the mind. Is a hierarchical model because levels are arranged with more specific concepts at the bottom. Stored common properties like "can fly" and "feathers" for birds at higher level node. Called "cognitive economy." Makes efficient, but problem because not all birds can fly. Added exceptions at lower nodes, like "ostrich." Shows how concepts and their properties are associated with the mind, physiological info wasn't available then.

What does it mean to say that there are different levels within a category? What arguments did Rosch present to support the idea that one of these levels is "privileged"? How has research on categorization by experts led to modifications of Rosch's ideas about which category is "basic" or "privileged."

Rosch suggests different categories, from "general" like "furniture" to specific "kitchen table." Superoridinate (global) level: furniture, basic level: table, subordinate level (specific): kitchen table. Rosch gave participants task to list as many features that would be common to all or most of the objects in the category. Example: for "table" put "has legs." Subjects listed only a few features that were common to all furniture, but almost all were common to kitchen tables. 9 features at basic level, 3 at global level. Basic lose a lot of information going to global, and doesn't gain much information going to specific, makes basic level psychological special. IN another experiment where people had to identify objects, picked basic names, like "guitar," "fish," and "pants" instead of "rock guitar," 'trout," and "jeans." Experts will say specific though, (Like bird names instead of "bird"). Categorization is learned through experience.

Physiological explanation for connection between perception and imagery through neuropsychology

See how brain damage affects imagery. M.G.S. supposed to have occipital lobe removed, before Farah had her perform mental walk talk, where she imagined walking toward animal and estimating how close she was when image overflowed. Before operation, was about 15 feet from imaginary horse, but after was 35 feet. removing part of visual cortex reduced the size of her field of view, so horse filled up more of the field. Supports that visual cortex is important for imagery. Damage to parietal lobes can cause unilateral neglect, where patient ignores objects in one half of the visual field. Asked patient to imagine standing in front of Piazza del Duomo, place he had been familiar with before brain was damaged, patient's responses showed that he neglected the left side of his mental image. Suggests that mental imagery and perception share physiological mechanisms.

What is the tacit knowledge explanation of imagery experiments? What experiment was done to counter this explanation?

Tacit knowledge explanation-participants unconsciously use knowledge about the world in making their judgements Finke and Pinker briefly presented a four-dot display, and then after a 2-second delay were presented with an arrow. Participants were told to indicate whether the arrow was pointing to any of the dots they had previously seen. Not told to use imagery, took longer to respond for greater distances between the arrow and the dot. Unlikely to show tacit knowledge because participants wouldn't have had time to memorize the distnances between the arrow and the dot. Spatial representation most supported by evidence.

Smith and prototypical objects

Used sentence verification technique to determine how rapidly people could answer questions about an object's category. Participants resond faster for objects that are high in prototypicality. Prototypical objects are named first. Prototypical objects are affected more by priming. Rosch showed that participants first heard the prime, like "green," two seconds later they saw a pair of colors side by side and indicated whether the two colors were the same or different. Colors were either same, poor representation, or different. Priming caused them to say "yes" quicker to prototype of green than lighter green. The prime will facilitate the participants' response to a stimulus if it contains some of the information needed to respond to the stimulus. Rosch showed that not all items in a category are the same.

Describe how experiments using the following physiological techniques have provided evidence of parallels between imagery and perception: d. recording from single neurons

a. recording from single neurons: Kreiman put electrodes in various areas of medial temporal lobe to determine source of seizures. Noticed that neurons responded to some objects but not others. Example, neuron fires to picture of baseball but not face, and still fires when person closes eyes and imagines a baseball, but not when imagines face. Respond to perceiving an object and imagining it.