Cog Psych 303

Question 1

Rosche introduced the concept of 3 levels of object identification, what are they

Answer 1

Superordinate (general category ie animal)
Basic (common name - dog)
subordinate (specific breed ie collie)

Question 2

Whats jolicoeurs dual route model

Answer 2

Dual-Route Model for Object Identification: Jolicoeur’s dual-route model posits that object identification occurs via two processes: a holistic route that normalises an object to an upright orientation, and a feature-based route that identifies objects based on unique visual features. Repeated exposure to an object reduces the reliance on orientation, as the feature-based route becomes more efficient.

Question 2

________level identification is the fastest, and suggested that more semantic processing is required for _____ordinate identification, while ________ identification requires more visual processing.

Answer 2

Basic - dog is fast because its a high freq word (more commonly used)
semantic processing is needed for superordinate - need to search semantic memory for group name
subordinate requires more processing to distinguish specific breed from general categorie

Question 2

Jolicoeur (1985): Conducted experiments to examine object identification when images were rotated. what did he find?

Answer 2

He found that naming latencies increased with the degree of rotation, suggesting that our visual system normalises objects to match their upright long-term representation.

Question 2

whose results supported Roschs notion that subordinate identification requires more visual processing, while superordinate identification involves semantic processing.

Answer 2

Jolicoeur, gluck and kosslyn

Question 2

Hamm and McMullen (1998): Investigated orientation effects in object identification, finding what level of identification was affected by orientation and what was not affected?

Answer 2

Hamm and McMullen (1998): Investigated orientation effects in object identification, finding that subordinate-level identifications (e.g., differentiating a Collie from a Poodle) are affected by orientation, (more visual info is needed) while basic and superordinate-level identifications are not.

Question 2

What was Murrays explanation to explain faster identification times for inverted objects

Answer 2

he suggested we Flip rather than spin in the picture plane

Question 3

Murray suggested that is we did not attend to an object the orientation effect would….?

Answer 3

not go away

Question 3

discuss Object Identification as a Multistage Process:

Answer 3

The process of identifying objects is not instantaneous but involves multiple stages, with different types of information being extracted at each stage.
Initial identification can occur without detailed processing, meaning the brain can categorize objects at a basic level (e.g., dog) before more specific details (e.g., Collie) are processed.

Question 4

What is meant by Variability in Visual Information?

Answer 4

The information projected onto the retina changes based on the object’s characteristics (e.g., shape, size, texture) and environmental factors (e.g., lighting, distance, angle).

Yet, the brain can maintain stable object recognition, which highlights the sophistication of the visual system in filtering out irrelevant variations.

Question 5

What is Influence of Word Frequency and Learning:

Answer 5

The speed of naming objects is influenced not only by visual factors but also by linguistic properties such as word frequency, word length, and the age at which certain terms are learned.

Basic-level names tend to be more common and learned earlier, which contributes to their faster recall.

Question 6

Visual similarity within and between categories is a major factor in identification speed, give an example

Answer 6

Objects within a basic-level category (e.g., different dog breeds) are visually similar, which can complicate subordinate identification. On the other hand, objects across superordinate categories (e.g., birds, fish, and dogs) are less visually similar, aiding in faster identification at the basic level.

Question 7

Identification can be referred to as the point….

Answer 7

where processing meets our long term representation

Question 8

_________’s research showed that this delay can be modeled as a linear increase in naming time with rotation, implying the brain has to “___________” the orientation before identification can proceed.

Answer 8

Jolicoeur, normalise

Question 9

What level or identification are atypical words such as penguin?

Answer 9

Basic

Question 10

according to Murray, what is the role of Attention and Object Recognition?

Answer 10

Murray (1995) demonstrated that attention plays a crucial role in object identification. If subjects are not attending to an object, even prior exposure does not reduce the effects of orientation on identification, meaning the visual system relies heavily on active processing for efficient recognition.

Question 11

Whats the issue with strict template matching

Answer 11

o Strict template matching, which suggests that a visual stimulus is directly matched to a stored template in the brain, is impractical because it would require a separate template for every possible variation in an object’s appearance (e.g., different sizes, rotations, positions). This approach is computationally overwhelming and doesn’t account for how the brain efficiently identifies objects despite these variations.

Question 12

shrinking something gradually or adjusting the dial until the luminance matches is a for of what?

Answer 12

normalisation

Question 12

Whats luminance coding?

Answer 12

Luminance coding is the realative brightness compared to the representation- caterpillar’s head is always 3 shades lighter than body even if its in the dark or the shadows

Question 12

What is object centered coding

Answer 12

Items are coded depending on where the pixels are relative to the center

Question 13

comparison by ratio matching is know as what coding

Answer 13

invarient

Question 14

comparing size to an average representation is known as ____

Answer 14

invarience coding - extracting something that doesnt change

Question 14

How would Ullman suggest we identify objects

Answer 14

he suggested a 3 feature alignment - we grab 3 ‘parts’ like curves or intersections and all memory representations with those matching are activated and all item templates in LTM get rotated/transformed to match what we are seeing

Question 15

Matching something to a specific memory trace through adjustment is

Answer 15

normalisation

Question 15

What did biderman and cooper show when comparing pianos and dogs etc,

Answer 15

when asked to name the object, size didn’t matter, like you accessed a size invariant representation, but when asked if they’d seen it before they were slower because they had to normalise it to see if it matched - you can do both depending on the task

Question 16

What did Navons figures show us?

Answer 16

Navon introduced figures (Navon figures) to study hierarchical processing in visual perception. His work showed that global (larger) shapes interfere with the identification of local (smaller) shapes, but not vice versa, suggesting that the visual system processes general shapes faster than specific details.

we start with big global and work our way done, identification starts at a LOW frequency (would make sense to identify at basic level)

Question 17

Whats the pigs in blanket thing

Answer 17

more pigs - more osccilations - higher frequency - more definition like a tv with more pixels

Question 18

Selfridge developed the “Pandemonium” model, what is it?

Answer 18

a feature-based model of letter recognition. In this model, various “demons” (features like lines or edges) activate to identify a letter by combining the detected features in the correct configuration.

has zero templates

Question 19

whats the issue with selfridges Pandemonium model

Answer 19

it over identifies things and has no spatial relationships

Question 20

whats the difference between template-based and feature-based models of object recognition

Answer 20

. Template-based models suggest that the brain matches incoming stimuli to stored mental templates, while feature-based models focus on detecting basic visual features (like edges or lines) to identify objects. These two approaches are not mutually exclusive, and modern theories often integrate aspects of both.

Question 20

what was Corballis critique of Jolicoeur’s notion of normalisation,

Answer 20

pointing out that if a stimulus isn’t yet identified, it cannot be normalised to an upright position, as the “upright” is defined by the object’s identity. Jolicoeur said you needed to flip then identify

Question 21

How is it thought that we reduce the number of templates

Answer 21

through normalisation or invariant coding

Question 22

______ is when the visual system extracts information that remains constant despite changes in object size, position, or brightness.

Answer 22

Invariant coding

Question 23

, the brain centering an object or standardise its size before attempting to match it to a template is know as ______

Answer 23

Normalisation

Question 24

what do feature-based models suggest?

Answer 24

suggest that instead of matching the whole object to a template, the brain first detects basic visual features (like edges, lines, or contours) and then uses these features to identify the object

Question 25

Object identification is often viewpoint-dependent. Jolicoeur’s research showed that objects take longer to identify when they are rotated away from their canonical (upright) orientation, suggesting that the brain performs a mental rotation to normalise the object’s view before identifying it. This contrasts with models like Biederman’s “geons” or Marr’s theory, which suggest ——

Answer 25

that identification should be viewpoint invariant.

Question 26

what are the three stages of Marrs computational model

Answer 26

 Primal Sketch:
Raw primal - The initial stage of visual processing where changes in light intensity across the retina are mapped but without depth information. CODES PARTS AS PLACE TOKENS (edges, bars, terminators, blobs). Once these are grouped, it, is passed on to.

Full primal sketch - groups larger place tokens to create low freq information

 2 ½-D Sketch: Adds depth information to the visual scene by calculating distances from the viewer, combining depth with surface features.
 3-D Model: This final stage involves creating a three-dimensional, view-invariant model of the object that can be matched with long-term memory representations.

Question 27

2. Biederman’s Recognition by Components (RBC):
   Biederman’s model refines Marr’s by introducing …

Answer 27

geons (geometric primitives), which are analogous to phonemes in language.

Question 28

What do Marrs cylinders/wireframes do

Answer 28

Marr uses wireframe models to conceptualise how objects can be represented in memory without dependence on the viewer’s position, enabling object constancy.

lines are the representations, it doesn’t care whenere you put the lion, just that the colours match

Question 29

What reason does biderman give for a dip in recognition speed?

Answer 29

Geons are considered view-invariant, meaning their identification is not affected by changes in object orientation. However, the relationships between geons (e.g., above-below, side-to-side) can change with rotation, which affects recognition speed.
o Biederman’s explanation of the “rotation dip” (objects rotated by 180° being easier to identify than those at 90°) relies on the fact that certain spatial relationships between geons are less disrupted at 180°.

Question 30

Whats the issue with Marrs 3D model

Answer 30

Marr’s theory does not account for why naming rotated objects takes longer if the 3-D model is truly view-invariant.

Both Marr’s and Biederman’s models attempt to handle object rotation, but neither fully explains why rotated objects take longer to identify.

Question 31

10. Ongoing Debate Between Models:

o Although both models contribute significantly to our understanding of object recognition, they differ in fundamental ways. _________ focuses on parsing objects based on concavities and axes of elongation, while ________ highlights the importance of vertices and intersections. The models use different spatial frequencies to process visual information, leading to different predictions about how objects should be identified.

Answer 31

Marr, Biederman

Question 32

1. Shepard and Metzler (1971):
   o Conducted foundational research on mental rotation, showing that people can mentally rotate 3D objects to determine if two images are the same or mirror images. They found that response times increase linearly with the degree of rotation, suggesting that …..

Answer 32

the closer to upright the faster the RT

people on average use mental rotation in proportion to the orientation - ie if its halfway to upright you rotate half the time

Question 33

cooper and shepard suggested we dont need to rotate if…

Answer 33

its closer to upright

Question 34

how was it demonstrated that spatial info is preserved in mental imagery during a map reading task?

Answer 34

9. Kosslyn, Ball, and Reiser (1978):
   o Studied mental map reading and found that the time taken to mentally navigate a map was proportional to the real-world distances between landmarks. This demonstrates how spatial information is preserved in mental imagery.

Question 35

Prior to rotation we must have
1
2
3

Answer 35

identified the stimulus
know its orientation
determine which way to rotate it

Question 36

Mental rotation is employed to ……

Answer 36

Resolve spatial conflict

Question 37

What would trigger the ‘flip’ of mirrored letters

Answer 37

spatial conflict AND conflict with the LTRepresentation

Question 38

who backed the View-Dependent Theories (2x)

Answer 38

View-Dependent Theories

Jolicoeur - Normalisation Theory

Jolicoeur suggested that object recognition involves mentally rotating objects to match a stored upright view, a view-dependent process. This theory implies that recognition times are influenced by the degree of rotation needed to align with a familiar viewpoint.

Shimon Ullman - Alignment Theory

Ullman proposed an alignment process whereby objects are recognised by transforming the image to match a stored template. His model suggests that objects are not view-invariant; rather, they must be adjusted or aligned to a standard view for recognition to occur.

Question 39

Who (2x) took a view-invariant standpoint?

Answer 39

David Marr - Computational Model
Marr proposed that object recognition progresses through three stages, ultimately achieving a 3D model that is viewpoint-independent. His theory suggested that the brain constructs a structural representation of objects that remains stable across different perspectives.
Irving Biederman - Recognition by Components (RBC)
Biederman’s RBC model uses geons (geometric primitives) to describe objects. He argued that these basic components and their spatial relationships allow for viewpoint-invariant recognition. According to this theory, as long as the geons are identifiable, the object can be recognised regardless of orientation.

Question 40

what is the difference between the view-invariant and view-dependant theory?

Answer 40

View-Invariant Theories
Core Idea: View-invariant theories propose that the brain encodes objects in a stable way regardless of the viewpoint. This means that recognition occurs because the brain has a single, abstract representation of the object that doesn’t rely on specific orientations or angles.

View-Dependent Theories
Core Idea: View-dependent theories suggest that the brain stores multiple specific views or perspectives of an object, meaning that recognition is based on comparing what we see to familiar views stored in memory.

Question 41

1. Phil Light & Hamm (2008):
   o This study on top-bottom discrimination involved participants identifying whether a dot was placed at the top or bottom of a rotated object. Surprisingly, they found a strong orientation effect— which is seen by…..

Answer 41

participants were slower to respond as the object became more rotated, though the effect was much smaller compared to classic mental rotation tasks.

the spatial location of the dot creates spatial conflict if the dot is at the bottom of the cat but you have to say top
when the dot was at the cats head - congruent - it was faster

Question 42

what does Beiderman say about why there is an orientation effect??

Answer 42

o Biederman’s geon theory explains orientation effects as disruptions in the spatial relationships between an object’s components (geons). For example, when an object is rotated, spatial relationships such as “A on top of B” might change to “A below B,” which could cause a delay in recognition. This theory provides an alternative to the mental rotation explanation.

Question 43

Whats the dot location effect?

Answer 43

spatial location of the dot creates spatial conflict - you have to say top when the dot is at the bottom but the cats head is at the top

Question 44

whats the difference between conceptual compatibility and the simon effect

Answer 44

Simon Effect
Core Idea: The Simon Effect occurs when response times are influenced by the physical location of a stimulus, even if the location is irrelevant to the task. If a stimulus appears on the right side of a screen and requires a left-hand response, reaction times are slower than if the stimulus and response location match (e.g., right side stimulus with right-hand response).
Type of Compatibility: Spatial compatibility—based on the physical alignment between the stimulus location and the response location.
Example: In a simple task where participants press a left or right button based on colour (e.g., green means left, red means right), responses are faster if the colour appears on the same side as the button location, even if colour location is irrelevant.
Why It Occurs: The Simon Effect is thought to arise because spatial location activates an automatic response tendency, creating interference when the response doesn’t match the stimulus location.

Conceptual Compatibility Effect
Core Idea: This effect occurs when there is a match between the meaning of a stimulus and the nature of the response it requires. For example, if the task involves categorising a word like “high” or “low” by pressing a button located higher or lower on a screen, response times are faster when the word’s meaning aligns with the button’s position (e.g., “high” with an upper button).
Type of Compatibility: Semantic or conceptual—based on the match between stimulus meaning and response characteristics.
Example: In a task where participants respond to words like “left” or “right” by pressing a left or right key, responses are faster when the word aligns conceptually with the response location, even if the words are presented centrally.
Why It Occurs: Conceptual compatibility enhances processing by reducing cognitive effort when the stimulus meaning (concept) aligns with the response’s conceptual characteristics.

Question 45

Which of the following researchers found that mental rotation is a linear process where response time increases with the degree of rotation?
o a) Biederman
o b) Shepard & Metzler
o c) Corballis & McLaren
o d) Hamm & McMullen

Answer 45

B

Question 46

The Flanker Compatibility Effect (FCE) suggests:
o a) Only attended stimuli are identified.
o b) Both attended and unattended stimuli are identified.
o c) Only compatible stimuli facilitate identification.
o d) Response times are faster for neutral stimuli

Answer 46

B

Question 47

What is the flanker compatability effect

Answer 47

The Flanker Compatibility Effect (FCE) occurs because unattended stimuli (flankers) are processed for their identity, even though attention is directed elsewhere (to the central target). When these flankers are compatible with the target (i.e., they would map to the same response), response times are faster. When they are incompatible (mapping to a different response), response times slow down due to interference. This suggests that both attended and unattended stimuli are identified and can influence response selection.

Question 48

What is Amusia?

Answer 48

disorder in the perception or production of music

Question 49

the Montreal battery for evaluation of Amusia tests 6 scales what are they?

Answer 49

Scale interval meter rythm memory contour
“some ideas create real memorable memories”

Question 50

What is the deficit in congenital
amusia?

Answer 50

The most obvious is the impairment on fine grained
pitch discrimination (being able to tell that two tones
of similar pitch are actually different).

Question 51

Tillman et al tested spatial processing in Amusia… how?

Answer 51

used a line bisection task, to test spatial processing and a mental rotation but found no differences - he also lacked females in the task

Question 52

Pigeons and identification by Delius and hollard - what was the test and finding

Answer 52

pigeons were trained to match from a sample or pick the odd one out

• choices were mirror/normal or identical vs slight physical variation

pigeos showed no orientation effect for mirror normal, unlike humans who were also slower, showing that pigeons process visual stimuli differently than in humans

Question 53

Pigeons and go-no go (hamm, matheson, honig) task

Answer 53

shown 6 items - 3 + and 3-, so they learnt to discriminate between which ones they got rewarded for

Probe trial, one was rotated, birded showed an orientation effect (showed a M shape with a dip in the middle like humans)
can compare BETWEEN TWO images just fine but when its a single one they take longer to access their LTM
they responded to it more as if they didn’t know what it was so they just started speaking

showed that they are doing some kind of mental rotation