PSC100Y unit2
Gestalt Psychology
"Gestalt" is a German word that means "figure" in the sense of a unified or meaningful whole • Studied the rules that our perceptual systems use to group together the different elements of a sensory input to make meaningful objects. • They believed that the perception of an object is more than just the perception of the individual parts.
Karl Popper
"In so far as a scientific statement speaks about reality, it must be falsifiable; and in so far as it is not falsifiable, it does not speak about reality." • This series of experiments shows an important thing about how people do research on the mind and brain. • One experiment doesn't usually prove something like the FFA being selective for faces. • There may be some alternative theory that could explain the results. • Attempts at falsifying a theory are essential in science.
Parvizi et al. Research Participant
"The subject was a 45-year-old man implanted with intracranial electrodes to localize the source of medication-resistant seizures... Standard presurgical evaluation revealed normal intellectual abilities and visuospatial functioning without any psychiatric comorbidities or visual field abnormalities."
Color Consistency
* mechanisms that partially cancel the effects of different light sources • This is because your visual system has color constancy mechanisms that adjust your perception of color. They work by using the distribution of wavelengths across the whole scene to make a guess about the wavelengths present in the source of light. They then factor out the lighting in your perception of the objects in the scene. • The light that reaches your eyes from a scene will depend not only on the objects in the scene, but also on what wavelengths are present in the light source. • Differences in the light source between two scenes create enormous differences in the wavelengths of light that hit your eyes, but the two scenes look only a little bit different to you.
Humans cannot Detect _________ light.
*Humans cannot detect infrared light -many snakes have specialized receptors on their heads that pick up infrared energy. - hard to imagine what it's like to perceive infrared energy with them.
Laws of Physics
*Inverse square law: Intensity α 1 ÷ distance2 -The laws of physics play a big role in our perception, because they govern how energy is influenced by the environment. • Many laws of optics are used to create computergenerated images. • For example, when an object blocks a source of light, it creates a shadow according to the geometry shown above. • However, due to the way that light scatters, the edge of a shadow is slightly blurry; that's how we tell the difference between a shadow and an object
Stages of Form Perception Parsing
*breaking the complex scene up into individual objects. • Finding edges • Filling in behind occluders • Grouping elements together
Retinal Size of the orange is temporary and non-diagnostic:
*retinal size can change from moment to moment without influencing our interpretation of what the object is. -Things look bigger when they are closer to the retina. -Things look smaller when they are far way from the retina
Almost all of the energy in a fluorescent light bulb is in the _________________, which is why it is more efficient than an incandescent light bulb.
*visible wavelengths
• Objects may emit energy, or they may influence environmental energy
- An object may generate a sound - An object may differentially reflect certain wavelengths of light
Locomotion- Dilation
- Dilation & Contraction: we used dilation of a scene away from a point to tell that we are heading to the point and we use contraction around a point to tell that we are heading away from it. • Motion serves as a cue to tell you where you are headed. • The point you are headed toward remains stable as the rest of the world dilates around it. • If heading for the end of a landing strip, motion will dilate out away from that spot. • When heading away from a point, the visual scene will appear to contract around that point • Every bit of the scene will move inward, toward this point.
Gestalt Psychology Example
- If we had a pile of spikes it would mean nothing to us, however if we arrange theses spikes in a specific order wee can see a that the borders of the spikes make the illusion that there is a 3 dimensional sphere • This is an illusion that usually gives us the right answer.
Example of filling in Occluders
-If ink spilled on a word occluding parts of that word it is easier to read with the ink on top of the word than if the parts were the kink were spilled were taken away. -Our visual system doesn't just strip away the occluding ink; it actually fills in missing information, allowing us to perceive the partially occluded object better. - The visual system uses assumptions about what things are likely to happen in the world and what things are unlikely to happen. CAN SOLVE WITH BOYES'S THEOREM • Hypothesis: big thick line that extends behind the ink. • Evidence: the fact that the edge on one side of the ink lines of with the other edge. -These edges are collinear • What we are computing: the probability that a big thick continuous line is present given the fact that these two edges are collinear. • P(E|H): the probability that the two edges would be collinear if we had a single thick, continuous line behind the spilled ink. • P(E): the probability that these two edges would be collinear just by chance, even if they weren't part of the same continuous line.
Distance & Motion Parallax
-Motion Parallax: how we determine which things are close to us and which are far away. Motion Parallax: things that are closer to you appear to move faster than things that are farther away
Choose the option that gives good examples of both "figure-ground segmentation" and "looming."
. Figure-ground: an X-shaped group of dots moving among a set of stationary dots; Looming: a basketball about to hit you directly in the face
Which of the following is a particularly important contribution of Biederman's Recognition-by-Components Theory to the problem of object recognition? From almost any 3-dimensional viewpoint, we can determine the parts (geons) of an object by means of template matching.
. From almost any 3-dimensional viewpoint, we can use non-accidental properties to determine the parts (geons) of an object.
When J.J. Gibson was training pilots during World War II, he realized that:
. The incoming scene dilates around the location that the plane is heading toward
You watched a video clip of Dr. Josef Parvizi conducting an experiment with a drug-resistant epilepsy patient in a hospital. The clip showed that face perception can be disrupted by _______________________________ an area of the ____________________________, which provides evidence that this part of the brain plays ________________________________ in face perception.
. stimulating; fusiform gyrus; a causal role
Heuristics
A sequence of operations that usually leads to the correct answer but sometimes fails
Which of the following would be an example of the use of top-down knowledge to influence perception?
A. Being able to recognize spoken words more easily if they are spoken in meaningful sentences rather than in random sequences
Key Issues in Cognitive Psychology Differences among healthy individuals, across development, and in disorders
Acquired Prosopagnosia - Caused by brain lesion or damage Congenital Prosopaggnosia- Born with
Prosopagnosia: Specific to Faces?
Almost all faces have the same structural description - the same basic parts in the same categorically defined locations. We cannot use this kind of representation to discriminate among faces Instead, we use the metric properties of faces to identify different people - the exact sizes and shapes of the parts and relative distances among them.
Algorithms
An algorithm is a sequence of operations that is guaranteed to reach a correct solution.
Faces of different people typically:
Are members of the same basic-level category Have the same structural description Contain the same parts in the same categorical spatial relationships
What was Magritte's point in putting the text "this is not a pipe" on a painting of a pipe?
Because it is paint on canvas, not an actual pipe
Recognition-By-Components Example
Brick: 3 parallel edges Inner Y vertex 3 outer arrow vertices • These properties can be seen from almost any viewpoint. Cylinder: 2 parallel edges 2 tangent Y vertices 2 parallel curves
Bayes's Theorem is useful in perception because it allows us to:
Combine previous knowledge with incoming sensory evidence to make inferences about the sensory input
Kanwisher Study Part 4
Compare hands with faces (wearing hats to hid hair) to rule out possibility that FFA responds to hair, to any biological stimuli, or to objects that have the same structural description and differ only in metric properties Hypothesis: FFA responce when all stimuli have the same structural description,but differ in metrc properties (false) FFA Stimulus Face > hands
According to structural description theories of object recognition, we recognize an object by:
Comparing a structural description of the sensory input with a structural description stored in memory.
Lesions of Area MT lead patients to complain about:
D. Having difficulty telling where they are heading when they are walking
Key Issues in Cognitive Psychology Hardware (cognitive neuroscience)
Edge detection in the retina Brian areas involved in face perception
What does "Thatcherizing" tell us about face perception?
Face perception is impaired when the face is inverted.
Car experts exhibit a larger N170 ERP response to both cars and faces than to other objects, whereas people who are not experts show a larger N170 response only for faces and not for cars. This finding supports the hypothesis that:
Faces are processed by a general system for expert, metric-based, subordinate-level categorization
Kanwisher Study Part 1
Find a possible face-selective region by comparing fMRI response to faces vs non-face objects Faces > Objects Fusiform Face Area (FFA)- Stronger responses to face than objects They found that there was greater activity for faces than for objects in the FFA. • This was not enough to show that the FFA is actually face specific. • There may be differences in low-level visual features such as texture or curvature that differ between a typical face and object.
The appearance of an object also depends on the properties of the object
For example, the color of an object depends on which wavelengths of light it absorbs and which wavelengths it reflects.
Which of the following is an example of "filling in" while parsing a visual scene?
Forming a perception of a complete fan behind the bubbles in "Fan Blowing Bubbles".
Isabel Gauthier Results
Fusiform gyrus was strongly activated by faces in all three subjects. • However, there was no activation for the greebles. • After training, all three subjects showed activation in the face area for the greebles. • Greeble experts show a large N170 curve for both greebles and faces, whereas non-experts show a large N170 for faces but not greebles. • Researchers concluded that face perception is not based on a face-specific processing system. • Relies on a more general system for expert, metric-based, subordinate-level categorization
Template-based theories of object recognition require that the visual system rotates the sensory input to match the orientation of the template. This assumption:
Has been confirmed by experiments showing that reaction times increase the more a to-be-recognized object is tilted away from upright
The visual system often uses heuristics rather than algorithms because:
Heuristics are typically faster than algorithms
mental rotation Shepard and Metzer
Less Mental rotation= Faster reaction time more mental rotation= slower reaction time
Gestalt Principles: Pragnanz
Literally, "pregnant" (with meaning) Often translated as "good figure" Pragnanz says that we should choose the simplest possible interpretation of the sensory input, which would be two simple shapes.
Nancy Kanwisher
MIT argues there really is specific parrot the gyrus for frossessing faces
J.J. Gibson:
Motion in particular provides a lot of information for vissusal perception i.e. Locomotion- Dilation
Figure-Ground Segmentation Example
Moving dots among stationary dots this segregates the four ground from the background
Motion parallax helps us determine the distances of objects because:
Nearby objects appear to move across the field of view faster than distant objects
Conditional Probabilities
P(A|B) Probability that A is true given that B is true Example 1 Probability that a person is > 6 feet tall for adult males and females: P(over 6' | adult male) > P(over 6' | adult female) Example 2 Probability that a circular object is present in the world given the image shown below P(circular object | sensory input) • The sensory input may be consistent with multiple interpretations, so prior knowledge is important to determine the most likely interpretation
Humans have color constancy mechanisms that allow us to _____.
Perceive the color of an object in the same way independent of the wavelengths present in the light source
Key Issues in Cognitive Psychology Limits on human abilities (e.g., speed, capacity)
Pigmentation in photo-receptors are able to absorb some but not all wavelengths (eg we cant see inferred wavelengths)
Gestalt Principles: Good Continuation
Preferred Interpretation |+\ Nonpreferred Interpretation >+< we tend to group things together if they fallow a straight line or a smooth curve involves similarity
Stable and Diagnostic
Properties are relatively constant over time an can be used to identify an object
Temporary and Not Diagnostic
Properties very over time and don't tell use the identity of an object
Which of the following hypotheses did Dr. Farah's research on face and eyeglass perception in a prosopagnosia patient help to disprove?
Prosopagnosia reflects a broad impairment in the ability to use metric properties to categorize items at the subordinate level.
Biederman's Recognition-by-Components Theory attempts to explain how we:
Recognize objects at the "basic" level (e.g., determining whether an object is a guitar or a briefcase)
Which of the following is the clearest example of research that examines the nature of mental representations?
Research that attempts to determine whether object recognition is based on templates or structural descriptions
Kanwisher Study Part 5
Rule out attentional differences by using "1- back task" (look for repetitions) for hands and faces in this expirement subjects pressed a button if they saw a stimulus twice Faces > hand (1-back)
Kanwisher Study Part 2
Rule out low-level differences between faces and nonface stimuli by comparing scrambled faces with intact faces Intact Faces> Scrambled faces scrambled faces have the same low level visual features as the intact faces FFA= more activity for the intact faces
Kanwisher Study Part 3
Rule out possibility that FFA responds when viewing different examples of the same category (houses) FAA response= Faces> Houses
Gestalt Principles: Similarity
Similar shapes are grouped together, causing perception of vertical groups things that are alike should be grouped together
Single-unit studies as well as intracranial and scalp ERP studies have revealed what about face perception?
Some neurons and cortical areas respond preferentially to face stimuli.
Color Blind- Red and gree Inhibited
Some people (usually men) have only 2 types of cones, and their perception of color is different.
Stimulation of the FFA
Stimulation --> Fusion Gyrus causal role With electric stimulation, patient will see face infront of them change
Stages of Form Perception Theories of Recognition
Structural Description Theories Template Theories (View-Based Theories)
Kanwisher Study
Subjects viewed sequence of images, including faces and some common objects subjects passively watched the images with no task
Key Issues in Cognitive Psychology Representations (format, persistence)
Template Theories Structural Description Theories
Key Issues in Cognitive Psychology Architecture of the human mind
The EPIC Model -• This model shows a single visual processing system; however, research on perception has shown that this can be subdivided even further. Kanwisher Part 1- • Kanwisher proposed that the FFA is specialized for face perception, while others have proposed that it is specialized for expert, metricbased, subordinate-level categorization. • This is an example of how cognitive psychologists are trying to understand the general organization of the visual system.
The Kanwisher et al. study that identified the fusiform face area (FFA) included an experiment in which subjects passively viewed sequences of faces and sequences of hands. Larger responses were observed for faces than for hands in this study. This result rules out the hypothesis that:
The FFA is equally responsive to any biological stimulus The FFA is equally responsive whenever subjects see a sequence of stimuli that are from the same basic-level category The FFA is equally responsive whenever subjects see a sequence of stimuli that differ only in metric properties and have the same structural description
The properties of the environment impact our perception of the objects in that environment. A clear example of this is:
The object we are trying to see might be occluded by some other part of the environment (e.g., a table might partially occlude a person sitting on the other side of it)
During a recent game of Frisbee, you took a photograph that captures an image consisting of a sky blue background and what looks like a fuzzy white tilted oval (the Frisbee). If you then show the picture to someone else who was playing Frisbee with you, that person is likely to perceive the object as being round rather than oval (which is actually correct because Frisbees are round). Bayes's Theorem would use which of the following options to explain why this person perceived the object as being round rather than oval?
The person had been playing Frisbee, and Frisbees are round, so the prior probability of the object being round was fairly high.
Bayes's Theorem computes P(H|E), which is:
The probability that a perceptual hypothesis (H) is true given the sensory evidence (E)
Imagine that you are looking at an object that casts an oval image on your retina, but the object might actually be a circle viewed from an angle. If we applied Bayes's Theorem to this example to determine if we were seeing a circle, P(H|E) would correspond to __________________________.
The probability that the object is a circle given the pattern of sensory input.
The term "parsing" refers to:
The process of dividing the continuous visual input into discrete objects
Both prosopagnosia studies and fMRI studies find evidence that:
The right hemisphere plays a stronger role than the left hemisphere in face perception
Figure-ground segregation can be difficult in static scenes because:
The visual input is just a 2-dimensional array of brightness and color values, with nothing to indicate borders between the foreground and background
The wavelengths of light that reach the eyes, and underlie the perception of the color of an object, are determined by:
The wavelengths that are present in the light source The ability of the object to reflect specific wavelengths
Boyels Theorem: Prior Probability
Theorem doesn't just weigh the current evidence, but also considers the prior probability that something is likely to be true Gives us a formal way of thinking about how we can combine top-down knowledge with bottom- up sensory information to draw inferences about what we are sensing among our many modalities
Achromatopsia
There is a true form of color blindness called "achromatopsia" that can result from brain damage. People with this disorder see the world like a black-and-white photograph.
Internal Factors
Things inside our brains that influence how we construct a representation of the world based on energy that hits our receptors
The Simultaneous Contrast Illusion Lightness
This illusion is the result of a heuristic in which the visual system uses the region surrounding an object to make a judgment about the lightness of the object—that is, how well the object reflects light—and the overall amount of light coming from the light source. • Although the number of photons hitting your eyes from the A is much larger in sunlight than in candlelight, you don't perceive this as a change in the lightness of the A • Your perception is that the white A is still white, but that the intensity of the source of illumination has changed.
The Kanwisher et al. study that identified the fusiform face area (FFA) included an experiment in which subjects performed a task in which they had to press a button if the same stimulus happened twice in a row. With this task, they observed greater FFA activity for face stimuli than for hand stimuli. The main purpose of having subjects perform this button-press task was:
To rule out the possibility that previous differences in FFA activity between faces and non-face stimuli was a result of subjects paying greater attention to the faces
The Kanwisher et al. study that identified the fusiform face area (FFA) included an experiment in which the FFA response to houses was compared to the FFA response to faces. They found that the FFA gave a larger response to faces than to houses. What was the purpose of this comparison of faces with houses?
To test whether the FFA is not specifically responsive to faces, but instead responds whenever multiple instances of the same category are presented
Energy Sources
Visible light is electromagnetic energy in a particular range of wavelengths
Humans can discriminate thousands of different colors. We can do this because:
We can use the relative activation of red-, green-, and blue-sensitive cones to differentiate between different colors
When we are asked to determine whether a rotated complex 3-dimensional object (Object A) is the same as another 3-dimensional complex object (Object B), it takes longer to make the comparison when the two objects are presented at different rotations. The amount of time increases more and more as the difference in the 3-D rotations of the two objects become greater. This finding has been used as evidence that that:
We mentally rotate one object so that it can be compared with the other object.
Plugging in Some Numbers
Which is more likely, H1 (sphere) or H2 (disk)? P(E|H1) = .8 (80% chance that a sphere would produce this image) P(E|H2) = .2 (20% chance that a flat disk would produce this image) P(H1) = P(H2) = .5 (Assume that, without any stimulus, we have equal reason to predict a sphere or a flat disk) P(E) = 1 (Not relevant for this example)
Isabel Gauthier and other researchers have proposed that prosopagnosia is a disorder in expert metric-based subordinate level categorization, which may apply to other stimuli in addition to faces. If you were an expert in categorizing Greebles and then experienced severe carbon monoxide poisoning, leading to damage on the inferior surface of the brain near the occipital-temporal junction, which of the following might happen to you (according to this theory)?
You might no longer be able to categorize faces or Greebles, but you could still tell the difference between a car and a truck.
3-D Structure from Motion
_Structure from Motion: motion allows us to determine the 3D structure of an object. • Putting this picture in motion gives us a sense of three dimensions, allowing us to easily perceive that it is a dinosaur.
Avoiding collisions:
avoiding collisions by noticing looming
• We have no ____________ about objects in the world
direct information
3-dimensional structure:
figuring out the 3D shape of an object despite the fact that each eye has just a 2D image
Stages of Form Perception Surface interpretation:
figuring out the surface properties of each object (color, lightness, 3D-shape)
Stages of Form Perception Parsing Surface interpretation:
figuring out the surface properties of an object, such as its color, texture, and 3D shape. Recognition
Figure-ground segmentation:
figuring out what's the foreground and what's the background - Figure-Ground Segmentation: pulling the foreground and background apart because they are moving differently from each other
Locomotion:
figuring out where we're heading when we're walking or driving
Associative Prosopagnosia
higher-order perceptual problem; people with this disorder would know that these are faces and could figure out that the one in the middle is a different person than the others, but wouldn't be able to recognize them by their names.
External Factors
influence the energy that is transduced by receptors
Stages of Form Perception Recognition
link lower level perceptual properties with the objects stored in
Apperceptive Prosopagnosia:
low-level perceptual problem; someone with this disorder would report that these images look like jumbles of features and wouldn't be sure if they were faces and wouldn't be able to determine which two images show the face of the same person.
What did J.J. Gibson notice when he was helping train Navy pilots to land planes on aircraft carriers during World War II?
motion can be used to help pilots determine the point towards which they're heading
Distance perception:
telling how far away we are from an object
Stages of Form Perception Recognition
the process of linking a sensory input to a representation in memory
An example of a stable and diagnostic property of an pencil might be:
the 3-dimensional shape of the pencil
The example a cloud of stationary dots and a few moving dots was a demonstration of:
the use of motion in figure-ground segmentation
The laws of physics play a big role in perception because they determine how energy is transmitted over time and space and influenced by the objects in the environment. Which of the following aspects of perception are influenced by the laws of physics?
the way that the 3-dimensional shape of an object leads to a particular pattern of light on the retina the way that the intensity of a sound depends on the distance between the source of the sound and the listener the way that a sound will bounce off the walls of a room, leading to echoes
Key Issues in Cognitive Psychology Real-world applications
understand how speach perception works, help us make hearing aids and optimal voice detection systems
The human retina contains three kinds of cones, which are sensitive to red, green, and blue wavelengths of light. How can we see the color pink?
we see the color pink as a result of the relative activations of the red, green, and blue cones
How dose perception work
when you perceive an object, like a book, you are taking little bits of energy that comes into your brain and actively constructing a representation of that object there is no little little book in your head. all you have is a pattern of light falling on the back of your brain eyes, and your brain figurs out the dimensions and color of the object. this is an interpretation that your brain imposes on the incoming sensory information
Area MT/V5
• Area MT (also known as Area V5) is crucial for motion processing. • A lesion of area MT in one hemisphere does not have a large impact as long as MT in the other hemisphere is healthy. • Those with lesions in both hemispheres have difficulty with walking around a room and recognizing objects.
Charlie Gross Experiment
• Charlie Gross did single unit recordings from inferotemporal (IT) cortex of the rhesus monkey -Interior cortex - V1 • Recordings show that the cell strongly responded to various hands and a bit to an oven mitt, but not to other stimuli. This cell gives a big response to monkey and human faces, but if you scramble the face, the cell doesn't respond very much.
Prosopagnosia: Specific to Faces?
• Dr. Farah had subjects do a face recognition task, and the prosopagnosia patient performed much worse than control subjects. • Prosopagnosia patients were actually unimpaired at the eye glasses task that required subordinate-level categorization. • Concluded that prosopagnosia is a specific problem with face perception and not a general problem with using metric properties to perform subordinate-level categorization. • Face perception is also something that involves enormous expertise. • Perhaps prosopagnosia involves a problem with subordinate-level categorization for categories of objects that we are highly experienced at perceiving. Expert, metric-based subordinate-level categorization _ ex) can tell different models of cars apart
External factors that influence the energy that is transduced by receptors:
• Energy sources, such as lights • Object properties, such as reflectance • Environment properties, such as walls • Laws of physics, which determine how energy changes over time and space • Receptor properties, such as sensitivity to particular colors • Algorithms and heuristics, which process the raw sensory data into a more useable form • Experience and knowledge, which influence how we classify sensory information
Holistic Perception of Faces
• Faces are perceived and stored holistically, meaning we do not store individual parts. • We also store metric relationships among parts and other overall parts. • It is difficult to perceive the face when it is inverted • Inversion doesn't have much of an impact on the identification of most other kinds of objects.
Perception 4 - Face Perception
• Faces gave the same basic parts in the same categorical positions, so they have the same geon structural description. • Faces differ in metric properties: the exact sizes and shapes of the parts and the relative distances among them.
Structural-Description Theories
• In structural-description theories, objects are represented abstractly as parts and relations between parts. • Recognition consists of forming a structural description of the input and comparing it with structural descriptions in memory. the visual system would form a structural description of the sensory input of that letter which would be compared with all of our structural descriptions in memory
Perception 3 - Object Recognition
• Is a printed round object a circle or a sphere? • Our visual system tests these hypotheses and chooses the one that best fits with both the sensory evidence and our prior experience of the world. - Bayes's theorem: P(H|E) -Can help us answer this question
Allison et al expirement
• It is also possible to record face-specific ERP responses from the surface of the cortex in humans. • This is usually done in people who have epilepsy that does not respond to anti-seizure medication. • The goal is to figure out exactly where the seizure is coming from so that that part of the brain can be removed. • In this study, they recorded activity while subjects looked at various stimuli such as cars, faces, scenes, words, and numbers. • The dark line shows the ERP elicited from the cortical surface to faces, which contains a very large response around 200 ms after onset. • This response was seen for faces, but not for other stimuli. Each dot on the brain represents one electrode in which a face-specific response was obtained. (can see dots in back of head) Right hemisphere shows more face-related activity than the left hemisphere. • The face-specific ERP component is called N170 because it is a negative-going voltage that peaks around 170 ms after onset of stimulus. • It is larger for faces than most other stimuli.
Subordinate-Level Categorization
• It is not designed to explain how we differentiate between two members of the same category. • Categorizing faces is subordinate-level categorization. • We do this using the metric properties of the faces.
Prosopagnosia
• Lesions of the occipital-temporal junction can lead to difficulty perceiving faces. • This disorder is called prosopagnosia.
Looming
• Looming: a pattern of motion that occurs when something is headed straight for your face. - Looming: how motion tells us if something is heading right towards us and if we need to take evasive action. • Retinal image size expands rapidly as the object gets very close to your face
Isabel Gauthier
• Many researchers are interested in the more general question of whether face perception relies on a set of face-specific processing mechanisms or instead relies on a more general system for expert metric-based subordinate-level categorization. • It's difficult to rigorously test these theories with naturally occurring stimuli such as faces and cars. • There are many differences between these types of stimuli, and people who become car experts may not be typical. • So Isabel developed an approach in which she trained people to become experts with artificial stimuli called greebles. • Greebles have same basic parts, but differ in metric properties. • Subjects were trained over a period of several days to identify greebles and group them into related families, which required them to make subordinate-level category decisions.
Mental rotation
• Our brain also mentally adjusts the size and rotation of the image so that it can be matched against the template.
Presurgery fMRI results from left inferior occipital-temporal region
• Performed an fMRI session to find regions of the fusiform gyrus that showed greater activity for faces than for other objects. • Face-sensitive regions are shown here as orang-ish blobs.
Receptor Properties pigment epithelium
• Photoreceptors need to be next to the pigment epithelium, an opaque substance that has to be at the very back of the eyes. • There are pigments in the photoreceptors that capture photons of light, leading to a change in the release of neurotransmitters from the photoreceptors. • After a pigment molecule captures a photon of light, it can't work again until it sends a molecule called retinal to the pigment epithelium to be regenerated.
Basic-Level Categorization
• Recognition by Components theory explained how we do basiclevel categorization
Recognition-By-Components Basic Level Categorization
• Recognition-By-Components theory is designed to do basiclevel categorization, where we figure out what basic category an object belongs to • It is not designed to explain how we can tell the difference between two members of the same category, such as a PRS custom 22 and '56 strat guitars.
Two Types of Object Properties We can make a distinction between two main types of object properties:
• Some are stable and diagnostic (the color of an orange, the parts of a blender) • Others are temporary and not diagnostic (the size of the orange's image on your retina, the outline of a blender)
Recognition-By-Components
• Structural descriptions work well for simple objects, such as T's, but it becomes more difficult for 3D objects. • Irv Biederman's theory, Recognition by Components, provides a solution for the problem of figuring out shapes of the parts of 3D objects, given that the shape of the retinal image depends on the 3D viewpoint from which we view an object. Common objects can be broken down into a set of simple 3D parts, that Biederman called geons. • 20 different geons, such as bricks, pyramids, and cones. • Real world objects can be defined by these geons and the relationships between them. • Some objects use the same geons, but have different spatial relationships, such as a cup or pail. • Some properties of an object are stable and diagnostic, whereas others depend on the particular viewing conditions and are not diagnostic. • The non-diagnostic features are termed "accdiental properties," and the diagnostic features are termed "non-accdiental properties."
Template Theory Example
• Template Theory: we have a picture-like representation of each object that we know in memory. • These picture-like memory representations are called templates. • We compare the incoming sensory input with our templates to see which template is best matched by the sensory input. • We compare our sensory input with the template in memory to see which fits best. Not all things look exactly the same (for example the letter A can be written in many different fonts, however we can still recognize it
Template Example Rotation & Scaling
• Templates still work even if they do not match perfectly. • As long as one template matches well enough, and matches better than any other template, we can recognize the object. • We also can have multiple templates of a given object, such as that of the letter A. • Our brain also mentally adjusts the size and rotation of the image so that it can be matched against the template.
fusiform gyrus
• The most consistent activity was in the fusiform gyrus, shown here in yellow. • Face-related brain activity can also be seen in fMRI studies...
Shape Constancy
• The outline changes from a rectangle to trapezoid, but we do not perceive the shape of the door as changing. • Shape Constancy: The ability of the visual system to see an object as having the same 3D shape in the face of changes in your viewpoint that influence the 2D image that hits the eyes.
Surface Interpretation
• The perception of an object's color is a result of our red, green, and blue photoreceptors. • There are also color constancy mechanisms we use to factor out the color of the light source from the reflectance properties of the surface of the object. • Using motion cues to figure out the 3D form of the exterior of an object is also an example of Surface Interpretation.
Parsing Finding Edges
• The process of finding edges is accomplished by early stages of visual cortex, which apply a variety of complex algorithms. • The result is something like the image above to the right of the original.
Receptor Properties Cone Types
• There are three kinds of cones, each of which is sensitive to a different range of wavelengths. We have red cones, green cones, and blue cones. • Every one of the millions of different colors you perceive is coded by your retina in terms of specific amounts of red, green, and blue
• This energy is picked up by ___________ in our eyes, ears, skin, nose, and mouth that ______________ various forms of energy into electrical signals that can be used by neurons
• This energy is picked up by receptor cells in our eyes, ears, skin, nose, and mouth that transduce various forms of energy into electrical signals that can be used by neurons - Photoreceptors, hair cells, mechanoreceptors, thermoreceptors, etc.
Heuristics The Ponzo Illusion
• This illusion happens because the railroad tracks are getting closer and closer as we go higher up in the picture, and the stones are getting smaller and smaller. • Your visual system uses these factors to guess that the upper part of the scene is farther away than the lower part. • This sort of heuristic usually works. • For example, it tells us that all of the poles along the side of the train track (circled in orange) are actually the same size, even though some cast a larger image on your retina than others.
Lesion Site for Prosopagnosia
• This image shows bilateral damage in a patient with prosopagnosia. -Bi lateral damage
Color Constancy
• This is analogous to color constancy; the ability to see an object as having the same reflectance properties in the face of changes in lighting that influence the specific pattern of wavelengths that hits the eyes. • In both color and shape constancy, the visual system is combining multiple sources of information to make a best guess about what is really going on in the world.
Bayes's Theorem
• This theorem tells us the probability that some hypothesis is true given some evidence. • The hypothesis might be that this middle letter is an H, and the evidence would be the shape of the middle letter. • The theorem states that the probability of the hypothesis being true given the evidence is equal to the probability that it would be true even without that evidence (based on other letters in the word) multiplied by the probability that this pattern of evidence would occur if the hypothesis is true (probability we'd get the slightly odd shaped "H "if it really is an "H"), and then divided by the probability that we would get this pattern even if it wasn't an "H."
Parvizi et al. experiment
• Used brain stimulation to provide evidence that a region in the fusiform gyrus plays a causal role in face perceptio
Parsing Filling in Behind Occluders
• We actually do more than simply eliminate occluding objects; we also make a guess about what lies behind the occluders, filling in the missing information. • We don't have conscious experience of the information we fill in, but research has demonstrated that we unconsciously represent information we infer lies behind occluders.
Structural Description by Theories VS Template theory
• We don't yet know which type of theory is correct. • One possibility is that the brain uses both kinds of mechanisms because each has its strengths and weaknesses.
Precept
• We then take this energy and construct a percept, a mental representation of the objects in the world
Acquired versus Congenital Prosopagnosia
• When prosopagnosia is a result of brain damage, it is called acquired prosopagnosia. • Those that are born with it have congenital prosopagnosia. -People with prsopagnosia quickly learn to recognize people from other cuse, such as glasses or voice
Thatcherization
• When the face is upside down, it is difficult to see the transformation. •Turing the mouth and eyes controverting them compared to the rest of the face -Hard to tell when face is upside down -Easy to see when face is right side up
Variations in Evidence
• Which is more likely, H1 or H2? - H1 = sphere - H2 = flat disk with fancy shading • This pattern of shading is likely for a sphere - P(E|H) is high for H1 • This pattern of shading is unlikely for a flat disk - P(E|H) is low for H2 P(H|E) = P(H) x P(E|H)/P(E)
Receptor Properties Photoreceptors
• receptors that transduce the energy from the environment, turning it into electrical signals. • The photoreceptors are in the retina at the back of the eye • Interestingly, the light has to go all the way through the eye and through these other cells to get to the photoreceptors. This causes some distortion and scattering of the light, which impairs our vision slightly.
Lightness Constancy:
•Lightness Constancy: we see the lightness of an object as being the same (constant) regardless of the intensity of the light source. • The is different from Color Constancy, in which we see the color as being the same even when the pattern of wavelengths of light changes.
Properties of the Environment
•When we're trying to perceive one object, a key property of the environment is the other objects that are present in the environment. • Occlude- the object we're trying to see is partly blocked by another object that is in front of it Auditory example- In the auditory modality, the walls of a room are a property of the environment that has a huge impact on the sound.
