Perception

Question 1

What are the Gestalt principles of organization?

Answer 1

A set of principles on how the visual system organises visual information into patterns and shapes to create larger objects (p.163, pdf)

Question 2

Apperceptive agnosia

Answer 2

A type of visual agnosia where patients are unable to recognise simple shapes (like squares or triangles) or to draw shapes that are shown.

It is believed that these patients have problems with early processesing of information in the visual system.
(p. 38 in physical book)

Question 3

Dorsal pathway

Answer 3

…also known as the “where” pathway, is the path from the primary visual cortex to parietal regions and is responsible for processing spatial information and for coordinating vision with action.

Question 4

Associative agnosia

Answer 4

A type of visual agnosia where patients are unable to recognise complex objects.

However, these patients are able to recognise simple shapes and can succesfully copy drawings of complex objects.

It is believed that these patients have intact early processing but problems with late information processesing in the visual system, i.e. pattern recognition. (p. 38)

Question 5

What is the principle of proximity?

Answer 5

We tend to group elements that are located closely together.

Question 6

Now, the retina

Answer 6

The light-sensitive tissue layer at the back of the eye that detects light focused by the lens. The retina contains photoreceptors (i.e. rods and cones) that convert light into neural signals (p. 39 in physical book)

Question 7

Ventral pathway

Answer 7

…also known as the “what” pathway, is the path from the primary visual cortex to the temporal lobe that is responsible for identifying objects

Question 8

What is the principle of similarity?

Answer 8

Elements that look similar tend to be grouped together

Question 9

What is the principle of closure and good form?

Answer 9

Our visual system tends to fill in missing parts to perceive complete/whole objects (closure) and organizes elements into the simplest whole (good form).

Question 10

Cones

Answer 10

Photoreceptor cells in the retina of the eye that are involved in color vision and produce high resolution, meaning greater visual acuity.

These cells are mostly concentrated in the fovea.
(p. 39 in physical book)

Question 11

Place of articulation

Answer 11

The spot in the vocal tract where airflow is blocked or narrowed to form a consonant (p.185, pdf version)

Question 12

Categorical perception

Answer 12

The perception of stimuli as clear categories and the failure to notice small differences within each category (p.190, pdf version)

Question 13

Top-down processing

Answer 13

Perceptual processing of a stimulus, in which we use context, expectations and prior knowledge to recognize a stimulus

Question 14

Word superiority effect

Answer 14

The superior recognition of letters when they appear in a word than when shown alone

Question 15

FLMP (fuzzy logical model of perception)

Answer 15

…refers to Massaro’s theory that perception comes from combining stimulus information with context information

Question 16

What is feature analysis?

Answer 16

Theory on how we recognize objects by combining all the features they they are made of

Question 17

What is deep convolutional networks?

Answer 17

It’s a network designed to recognize patterns. It’s called convolutional because it looks at small parts of an image at a time, and it’s called deep because it has many layers

Question 18

Prosopagnosia

Answer 18

A condition caused by damage to the temporal lobe, where people have difficulty recognizing faces

Question 19

Fovea

Answer 19

A small pit at the back of retina responsible for the sharpest, most detailed vision of the eye. It features a large density of cone cells and allows for high-resolution visual input to the brain. (p. 41 in physical book)

Question 20

A feature map

Answer 20

The spatial representation of visual features. The visual system analyzes a stimulus into many independent features in specific locations. Since different pathways have cells differently sensitive to color, orientation, movement, there is a map for each of these features.

Question 21

receptive field

Answer 21

The patch of visual space that a neuron is responsive to if an appropriate stimulus is presented there.

Question 22

Phoneme-restoration effect

Answer 22

Hearing missing phonemes in speech when the context makes them expected, even if they weren’t actually spoken

Question 23

Change blindness

Answer 23

The inability to notice a change in a scene when it fits the surrounding context

Question 24

What is the principles of good continuation?

Answer 24

We tend to follow smooth lines and paths instead of seeing breaks. Just like in figure 2.12 in book, we see a line from A to B and C to D

Question 25

What is template matching?

Answer 25

It's a theory about how our brain stores a template and we recognize things by matching what we see to that template

Question 26

What is fusiform gyrus?

Answer 26

It is an area in the temporal lobe that becomes active when we see faces

Question 27

What is phonemes?

Answer 27

They are the smallest units of sound in a language. E.g. “bat” consist of 3 /b/, /a/, and /t/

Question 28

What is consonantal feature?

Answer 28

It’s the property of a sound that makes it behave like a consonant

Question 29

Bar detectors

Answer 29

Cells in the visual cortex which respond positively to light in the center and negatively to light in the periphery or vice versa.

Question 30

Edge detectors

Answer 30

Cells in the visual cortex which respond positively to light on one side of a line and negatively to light on the other side or vice versa.

Question 31

2½-D sketch

Answer 31

This sketch allows us to identify where objects are in space, relative to the viewer. It employs cues such as texture gradient, stereopsis and motion parallax.

Question 32

3-D model

Answer 32

The late stage representation of objects in the visual field. It identifies how parts go together to form an image of the objects we see. Gestalt principles take us to a 3-D model.

Question 33

Primal sketch

Answer 33

In the early stage of perception, features are extracted to make initial sense of the information and thereby make a primal sketch.

Question 34

What is visual agnosia?

Answer 34

Impairment in recognition of objects presented visually. Two types: apperceptive agnosia and associative agnosia

Question 35

Describe the lens

Answer 35

Located behind the pupil. Actively focuses or bends light as it enters the eye so it falls on the fovea (also known as accommodation).

Question 36

What are rod cells?

Answer 36

Photoreceptor cells in the retina of the eye that aid in night vision and motion detection. They provide poor colour vision and low resolution, resulting in lower visual acuity.

Question 37

Describe the pupil

Answer 37

The opening in the middle of the iris. Changes size to allow different amount of light to enter the eye.

Question 38

Bottom-up processing

Answer 38

The processing of a stimulus in which information from the physical stimulus, rather than from general expectations/knowledge, is used to help recognise the stimulus. That is, information travels 'up' from the stimuli, via the senses, to the brain.

Question 38

What is the Ebbinghaus illusion?

Answer 38

An illusion where a central circle appears larger or smaller depending on the size of the surrounding circles, despite the central circle’s actual size remaining constant.

Question 39

What happens during the early and later phase of visual perception?

Answer 39

In the early phase, shapes and objects are extracted from the visual scene. In the later phase, shapes and objects are recognised.

Question 40

What is texture gradient?

Answer 40

Elements tend to appear more closely packed together as the distance from the view increases.

Question 41

What is stereopsis?

Answer 41

The ability to perceive 3-D because two eyes receive slightly different view of the world.

Question 42

What is motion parallax?

Answer 42

Provides 3D information when an object is in motion. As more distant points move they will move more slowly across the retina than closer points.

Question 43

What does Warren (1970) show in regards to the phoneme-restoration effect?

Answer 43

- Illustrates the phoneme-restoration effect (good evidence exists for the role of context in the perception of speech) - Participants listened to “The state governors met with their respective legislatures convening in the capital city,” with a 120-ms tone replacing the middle s in legislatures - Only 1 in 20 participants reported hearing the pure tone, but couldn’t locate it correctly

Question 44

Yin (1969)

Answer 44

Study that showed that people are much better at recognizing faces presented in upright orientation than other categories of objects, such as houses in same orientation. But when faces and object presented upside down, there is a dramatic decrease in its recognition, but not in the same way for other objects

Question 45

Beck, Rees, Frith, and Lavie (2001)

Answer 45

- Study of the ability to detect changes in people’s faces - They found greater activation in the fusiform gyrus, when changes were detected, than when they were not

Question 46

Rumelhart and Siple (1974) and Thompson and Massaro (1973)

Answer 46

- They explain why people are more accurate when identifying the letter in the word context (see figure 2.26) - In figure 2.26.a, if only shown the very last letter or the first three letters, the participant wouldn’t be able to intuitively guess correct. Only when combining the two it becomes clear that the whole word must be WORK. It’s an unconscious inference. Same as in figure 2.26.b where multiple letters are unrecognizable, but when combining the fragments only one word is possible

Question 47

How do Reicher (1969) and Wheeler (1970) demonstrate the word superiority effect?

Answer 47

- Participants were presented with either a letter (such as D) or a word (such as WORD). Then given a pair of alternatives and had to tell which alternative they had seen. E.g. If shown letter D, then D and K could be alternatives. If shown the word “WORD”, then WORD and WORK could be alternatives. Both scenarios differed only by the letter D or K - Results: 10% more accuracy in identifying the word than letter alone. Better discrimination between D and K better in the context of a word than as letters alone

Question 48

Ratcliff & Newcombe, 1982 - study on associative agnosia

Answer 48

- a patient with associative agnosia ( visual agnosia with regards to object recognition) was presented with a drawing of an anchor and was supposed to recreate it - patient was able to recreate the drawing relatively, but couldn't recognize it (he said it was an umbrella) - suggests that people with associative agnosia don't struggle with early processing, but with pattern recognition, that occurs later

Question 48

Benson & Greenberg, 1969 study of patient with apperceptive agnosia

Answer 48

- a soldier suffered brain damage, which resulted in a visual agnosia - he was able to recognize objects by their feel, smell, and sound, but couldn't distinguish a circle and a square from each other, recognize letters or faces, he could, however, distinguish colors from each other or tell in what direction an object is moving - suggests that there is a part of a brain responsible for transforming visual information into perceptual experience, another part of the brain is responsible for sensory information; visual perception is more than just "seeing"

Question 49

Goodale, Milner, Jakobson, & Carey, 1991 - study of patient with associative agnosia

Answer 49

- a patient with visual agnosia, due to damage in temporal lobe (and no damage in parietal lobe), could correctly reach out and grasp the door handle, even though she couldn't recognize the object - suggests that the "where" pathway (neural pathway carrying visual information from primary visual cortex to parietal lobe) is specialized in action

Question 50

Hubel and Wiesel (1962) - study of ganglion cells in a cat

Answer 50

- study on a cat's primary visual cortex - they found differently configured receptive fields in primary visual cortex than in ganglion cells and cells in the lateral geniculate nucleus - they found cells that are elongated, in contrast to circular receptive fields of the on-off and off-on cells; there are edge detectors and bar detectors

Question 51

(Gross, 2008)

Answer 51

- study on the primary visual cortex - it was found that besides cells that respond to particular patterns in particular locations, there are also cells that respond to particular patterns in many locations

Question 52

Desimone, Albright, Gross, & Bruce, 1984

Answer 52

- study on a macaque monkey - they found a neuron in the inferior temporal lobe of the monkey's brain that would only fire up when the monkey was shown a picture of a hand or something highly resembling a hand, while seeming somewhat insensitive to the hand location - this study suggests that there are cells that respond to particular complex patterns (like hands or faces) in many locations

Question 53

Gibson, 1950

Question 54

David Marr - proposed levels of visual perception from early to late visual processing

Answer 54

Light energy - primal sketch - 2 1/2 - D sketch - 3 - D model - recognized objects

Question 55

Palmer (1977)

Question 56

Hoffman and Richards (1985)

Question 57

Visual pattern recognition — Template-matching

Question 58

Visual pattern recognition — Feature analysis Gibson (1969)

Answer 58

Gibson (1969) proposed how the features underlying the recognition of letters is divided. E.g. the capital letter A can be seen as consiting of a horizontal, two diagonals in opposite orientations, line intersection, symmetry and "vertical discontinuity". In comparison to the template model, some of the advantages are, simpler features making it easier to correct for the difficulties of the template model in recognizing full patterns Also highlights the most important parts of the relationships to the pattern. For example with A being three lines that intersect, two diagonals and one horizontal. The feature template also eliminates the need for a large amount of templates.

Question 59

Kinney, Marsetta, & Showman, (1966)

Answer 59

In relation to evidence for the existence of features as componets in pattern recognition. Kinney et al. looked into the "open H", which is when letters share features, people are more prone to confuse them. Participants were briefly exposed to a letter, and then tested to see what letter they saw. What they found out was, that their participants made 29 errors when presented with the letter G, 21 of the misclassifications were the letter C, 6 of them were the letter O, 1 misclassification was B, and 1 misclassification was 9. No other errors occured. Conclusion is, that participants were more often misclassfying the letter G with the letters sharing similar features with it.

Question 60

Pritchard (1961)

Answer 60

Looked into feature-analysis Looked into psychological nystagmus which is the small drift that happens with a rate of 30 to 70 cycles per seconds in the eye, and how it ties in together with perception. What they specifically tested was disintegration of an image that is stabilized on the eye. What they found out, was that that the partial outlines show various patterns reported as the stabilized image began to disappear.

Question 61

Krizhevsky, Sutskever, & Hinton, (2012)

Answer 61

Made a deep convolutional network for object recognition In the 8-layerd model, image processing starts with the stimulus, then followed by five layers of pattern recognizers. Elements in small regions of pixels converge on elements in layer 1. Layer 1 elements converge on layer 2, then elements that recognize patterns of layer 1 elements that recognize patterns continue onto layer 3 and so on.

Question 62

He, Zhang, Ren, & Sun, (2016)

Answer 62

Made a deeper deep convolutional network than Krizhevsky et al. with 150 layers instead of the 8-layerd model for object recognition. Works by elements in small regions of pixels converge on elements in layer 1. Layer 1 elements converge on layer 2, then elements that recognize patterns of layer 1 elements that recognize patterns continue onto layer 3 and so on.

Question 63

The McGurk effect (McGurk & MacDonald, 1976)

Answer 63

Involves watching the lips of someone. Hearing "ba" while seeing the lips move to say "ga" is often perceived as "da" by the listeners. Even when they know that the sound is ba, they often hear something else. So, they merge the auditory stimulus with the context provided by the lips.

Question 64

Lisker and Abramson (1970) - Auditory categorical perception

Answer 64

Experimented with computer-generated syllables in which the delay between the release of air and the onset of voicing was varied from −150 ms to +150 ms. The participants had to identify which syllables began with 'b' and which with 'p'. At about 25 m/s there was a switch from 'b' to 'p' At 10 ms participants agreed that the sound was a 'b' and at 60 ms, they agreed that the sound was a 'p'. Because of this, perception of this feature was referred to as categorical.

Question 65

Gauthier, Skudlarski, Gore, and Anderson (2000) - greebles

Answer 65

Researched whether the fusiform gyrus is only specialized for face recognition. Gauthier et al. found that bird experts and car experts showed high activation in the fusiform gyrus when judging birds or cars, and another study showed that people who practiced recognizing unfamiliar objects called greebles also activated the fusiform gyrus. These studies showed that since we are very familiar with faces, we are good at making judgments about them, but similar detailed processing can happen with any type of object we have a lot of experience with.

Question 66

Miller and Nicely (1955)

Answer 66

Looked into which sounds were most often confused. Participants had to identify phonemes 'b', 'd', 'p', 't' by listening to the sounds 'ba', 'da', 'pa', 'ta' presented in noise. The result was that the participants often confused one sound in the noise for another, and they mostly confused consonants that differed by only one feature. E.g. when hearing 'p', the participants more often thought they heard 't', which only differs in place of articulation, rather than 'd', which differs in both place of articulation and voicing. Similarly, when hearing 'b', participants often thought they heard 'p', which differs only in voicing rather than 't', which differs in both features.

Question 67

Goldstone (1994), Goldstone & Hendrickson (2010) Visual categorical perception

Answer 67

Goldstone (1994): Goldstone trained participants to categorize novel visual stimuli. where categories were determined by either either size or brightness of stimuli. People were better at discriminating to which category the stimuli belonged to when the aspect was relevant for discrimination. Goldstone & Hendrickson (2010) People are better at discriminating across categories (acquired distinctiveness) and decreased discriminabilty within categories (acquired equivalence).

Question 68

Pisoni (1977) Auditory categorical perception

Answer 68

There are clear boundaries for speech signals. Pisoni created nonlinguistic tones that had a distiguishing accoustic feature (comparable to voice-onset time in voicing, see p. 61 figure 2.24). It was either a low frequency tone, simultaneously presented with a high-fr. tone or it was lagged. Participant showed abrupt boundaries for speech signals. Thus for auditory categorical perception the signal doesn't need to be speech.

Question 69

Massaro 1979 (FLMP Model for Combination of Context and FEature Information)

Answer 69

In his experiment to study how participants combine stimulus information from a letter with context information from the surrounding letters, Massaro showed 4 variations with different amounts of contextual evidence (see p.65, fig 2.27): 1: only "e" can make a word; 2: only "c" can make a word; 3: either "c" or "e" can make a word; 4: neither "e" or "c" can make a word. Massaro found when e looked less ambiguous and more like "e", prob. for identifying e increased, similiarly when context increased (see fig. 2.28). This model is called FLMP (fuzzy logical model of perception). Thus it can be concluded that contextual information combines independently with stimulus information to determine what pattern is perceived.

Question 69

Kuhl (1987) Auditory catgeorical perception

Answer 69

Kuhl trained chinchillas to discrimnate between syllabels "da" and "ta". Allthough these animals don't have a human vocal tract, they showed the same perceptual boundary humans do. Thus, categorical perception the signal doesn't depend on whether the perceiver has a human vocal or auditory system.