Final Exam Review Flashcards
3 stages of high level vision
Image
Surfaces
Objects
Approaches for object recognition
Template matching
Structural descriptions
View- dependent recognition
Template matching
Receptive fields as templates
- match image to stored representation in brain
Structural descriptions
Recognition by components
- interpretation depends on arrangement
- geons
What is a geon?
A simple 2D or 3D shape/ structure
- cylinder etc
Object recognition in the brain
Inferotemporal (IT) cortex
- neurons respond to complex stimuli
Viewpoint - dependent responses
Not all recognition is viewpoint invariant
- can read words upside down
View - dependent recognition
Stored characteristic views of objects
- can tell what an object is from different angles
Different levels of description
Objects can be classified at different levels
Superordinate classification of objects
Something super broad like an animal
Entry level classification of objects
More specific than “animal” but still just ‘dog’
Subordinate classification of objects
Specific to breed of animal such as “Labrador”
Neuropsychology
Inferences about function from impairments following brain damage
Visual agnosia
Loss of object recognition despite normal low-level vision
What are the 2 types of visual agnosia?
Apperceptive agnosia
Associative agnosia
Apperceptive agnosia
Disruption of basic form processing
- cannot complete images in brain
Associative agnosia
Impaired access to stored visual representations in memory
- may think that an octopus is a spider etc
Example of associative agnosia
Can copy drawings but cannot recognise own work
Neuropsychological inference
Patterns of dissociation
Separate brain systems for different mental functions
- “what” vs “where” pathways
Agnostic patient DF - perceptual orientation judgment
- subject had a circle with a rectangle in the middle and had to match the orientation of the rectangle shown to them
- The control showed an accurate straight line depiction but DF was scattered and nothing like the control
Agnostic patient DF - visuomotor task
Dissociation of recognition vs action
- had the subject post a thin object through a rectangle in a circle
- the results were much closer to that of the control this time.
Dissociating perception and action
Ebbinghaus illusion
Ebbinghaus illusion
Size can be manipulated by surrounding shapes.
- 2 orange circles are the same size - each represent the centre of a flower
- Orange circle 1 looks much smaller because the petals are larger
- Orange circle 2 looks much larger because the petals are smaller
Single vs double dissociations
Unequal effects on performance could arise from damage to one system.
- get a stronger inference from double dissociation
Looking for a double dissociation
Has object recognition but no vision for action
- Optic ataxia
Optic ataxia
Damage to parietal cortex
Two pathways for visual perception
Dorsal and ventral pathways
Dorsal pathway
“Where” pathway
- parietal cortex
- magnocellular input
Ventral pathway
“What” pathway
- lateral occipital and inferotemporal cortex
- parvocellular input
Prosopagnosia
Inability to recognise familiar faces
- family faces
- famous faces
- themselves in a mirror
Identifying people when you have prosopagnosia
People with prosopagnosia can identify people through other means:
- basic vision is intact
- recognises voices
- the way people walk
Origin of prosopagnosia
. “Acquired” due to brain damage
. Congenital “face blindness”
Ways prosopagnosia is acquired
Brain damage through strokes, car accidents etc…
Congenital “face blindness”
. A lifelong impairment
- 2% of people are born with it, it develops for the rest
- genetic
Prosopagnosia is single dissociation
Most people with prosopagnosia have intact object recognition but impaired fascia recognition
- possible damage to only one system?
Faces require within category discrimination; they all look so similar and are more difficult to tell apart.
Double dissociation between faces and objects example
Patient CK
Case of patient CK
Visual agnostic - head injury at age 27
Impaired object recognition:
- can’t recognise food in a cafeteria
- can’t distinguish toy soldiers he collected as a child
Normal face recognition:
- can see, describe and recognise faces
Face recognition in CK case
Vegetable face
- can see the face that is made out of vegetables and can describe what feature each vegetable is representing.
Face inversion effect
“Special” face processing only occurs for upright faces
Upside down faces are processed slower and less efficiently
Thatcher illusion
Face is turned upside down and looks normal, even though some features such as mouth and eyes stay the correct way.
Upside down it looks normal but the correct way up it looks extremely bad.
Holistic processing of faces
Whole vs part advantage
- more accurate recognition of individual parts
Composite face effect
When the top and bottom half of different faces are put together, they’re easier to recognise when misaligned than when directly lined up on top of one another.
Face distortion aftereffect
After looking at a distorted face for a while, the original face looks overly distorted in the opposite direction
- if looking at an extremely thin face, the original will look wider than it actually is
Face aftereffects
Adaptation moves perception in opposite direction
- shifts the neutral point
Neural basis of face perception
“Face cells” in IT cortex
The fusiform face area (FFA)
Specialised part of brain activated i facial recognition
Damaged in acquired prosopagnosia
Representation of person identity
Intracranial recordings of hippocampus
- recognise someone so have a “mum” cell or a “Jennifer Aniston” cell that only recognises that face.
FFA feeds into regions associated with memory
Anterior IT
Amygdala
Delusional misidentification syndromes
Capgras syndrome
Fregoli syndrome
Capgras delusion syndrome
Belief that family members have been replaced with identical looking imposters
FFA and Amygdala are disconnected
Fregoli syndrome
Belief that different people are all the same person in disguise
Neuroimaging of high-level vision
Using an MRI machine to see how blood flow changes are associated with neural firing
Parahippocampal Place Area (PPA)
Recognises places but not faces
- geographic layout of environment (houses, rooms, outdoor scenes)
Topographagnosia
Damage to the PPA
Loose the ability to navigate
Otherwise have normal vision
- know where they are but everything looks unfamiliar
Body perception lobe
Extrastriate body area (EBA)
- static images of human bodies or body parts
- lower response to faces, objects, animal bodies
EBA damage - body part agnosia
Distorted self image (“too fat”) even when severely underweight
- changes in EBA structure and function
- excessive food restriction
Body shape aftereffect
Whichever body type you spend most of your time looking at becomes the norm
- social media filters distort bodies
Static vs dynamic body perception
EBA processes the structure of body parts
Superior temporal sulcus (STS) looks at biological movement
Superior temporal sulcus (STS)
Biological motion
Sensitivity to biological motion
“Uncanny valley” in film production
- when something looks and moves like a human but isn’t quite a human (dolls and zombies) its more scary
Brain activity for uncanny motion
Larger STS response to android human vs robot human
. Biggest response to something that looks human but doesn’t move like one
Biological motion
What contributes to the perceived attractiveness of biological motion?
Gender, weight, emotional state etc…
Sexual dimorphism
An exaggerated idea of attractiveness
Attractiveness of point light walkers
Hyper feminine movements appear to attract more mates
Measure heterosexual mens appeal to the way point light silhouettes walk
How do we use information about bodies?
Gestures
Proxemics
Mirroring
Social cognition
Conscious gestures
The peace sign to some places is considered swearing to other places and people.
Unconscious gestures
Generally universal (shrugging shoulders)
Some have specific cultural meanings
Gestures during communication
Deaf children have spontaneous gestures
Proxemics
Cultural perceptions of personal space:
- where do you sit on an empty bench?
Unconscious rules of how we arrange ourselves in an elevator
Personal space
Damage to the amygdala causes someone to have no concept of personal space.
Different levels of personal space based on how well you know someone.
Standing face to face versus not making eye contact based on comfort level with that person.
Mirroring
Unconsciously reproduce posture and gestures of other people we like and respect.
Body posture and emotion
Facial cues carry information about emotion.
- Intense emotions can be difficult to differentiate between.
- same face on 2 different bodies - rely on body cues to tell emotion
Expression in animals
Body posture and gestures of animals tell us their emotions
Evolution of facial expressions
Primate facial musculature
- mainly forehead, nose and lip wrinkles
Development of facial expressions
Early sensitivity to faces in babies
- more reactive to something with a face on than something without a face
Babies produce facial expressions in first few hours of being born in reaction to sweet, sour and bitter stimuli
Origins of expression production
Are facial expressions learnt by being observed or are they hardwired?
Blind athletes still show facial expressions
Universal expressions
Anger
Fear
Disgust
Shock
Joy
Sadness
Universal emotions across cultures
When tribes were studied they had the same emotional reactions to events as we do
Asymmetry in emotion processing
Being presented with an asymmetrical face makes it harder to distinguish emotions
Head and gaze direction
The rotation of your head and gaze direction can impact the emotion or expression you display
“Reading the mind in the eyes”
Some people interpret eyes as playful and joyful whilst others interpret them as bored, upset or tired etc…
What makes a face attractive?
Symmetry
Averageness
Sexual dimorphism
Symmetry
Symmetry is a physical indication of reproductive fitness
Averageness
An average of two faces is more attractive than either of the original faces
- more attractive faces = more attractive average
Sexual dimorphism in attractiveness
Sex hormones influence face shape
Sexual dimorphism in females
Neonate features (“baby face”)
High cheekbones and narrow cheeks
Sexual dimorphism in males
Stronger jaws and heavier brows
- attractiveness may depend on female menstrual phase
Define attention
Selection: attention selects information for awareness
Slow, serial, limited capacity
Preattentive processing
Image measurements in V1 - automatic (bottom-up)
Properties must be bound together by attention
Two types of preattentive processing
Exogenous
Endogenous
Exogenous attention processing
Reflexive or involuntary attention
Generated by external stimuli - brightness, colour, shape etc…
Endogenous attention processing
Goal - directed attention
Voluntary attention selected by location or feature
Effortful
Selection in space
Posner cuing task for covert attention
Posner cuing task
Measures manual eye-movement reaction times in relation to cues
Selecting for features
When we look at a scene we often look for certain features - some are easier to find than others.
Visual search for features
Finding a friend in green is much easier when everyone is wearing red than when everyone is wearing green
- like where’s wally
Feature selection
When looking at something, some features ‘pop out’
Other features must be attended one at a time - none stand out
Visual search experiments
Subject reports whether a horizontal line is present in an array of vertical lines of different amounts
Parallel vs serial searching
Serial processing = only one object can be processed at a time
Parallel processing = various objects are processed simultaneously
Serial processing
Only one object can be processed at a time
Parallel processing
Various objects can be processed simultaneously
Conjunction search
Features need to be bound into objects
Feature interrogation theory
Attention as “glue” binding features together
High level features may be preattentive
Certain features guide our attention
Disorders of attention
Dorsal processing stream
- “where”/“how”
Visual neglect (hemispatial)
People only have responses in one side of their visual field
When copying a drawing, they only copy one half
Measuring visual neglect in the lab
Line cancellation task
- have to cross out all the lines shown and only one half of the lines are crossed through
Line bisection task
- have to section a line in half and they cut it at 3/4 or 1/4 because they only see one half of the line
Visual extinction
When 2 stimuli compete for the subject’s attention and the subject cannot identify the object
Simultanagnosia
Subject can only focus on one object at a time
Term for when subject can only focus on one thing at a time
Simultanagnosia
Balint’s syndrome
Bilateral parietal damage
3 symptoms:
. Simultanagnosia
. Difficulty fixating
. Optic ataxia
Anosognosia
Patient denies or is unaware of medical condition
Term for when patient denies or is unaware of medical condition
Anosognosia
Anton - Babinski syndrome
Denial of blindness due to occipital damage
Blindsight
Damage to V1 leaves patient totally blind.
Basic visual function is preserved without awareness.
Limits of physiology
Too much information to process everything
Attention and high level vision
Increasing effects at higher levels of processing
Capacity limits of attention
Space and time limits on attentional capacity
Attention has limited capacity
Space: crowding
Tracking: multiple object tracking
Time: attentional blink
Visual crowding
Objects which are easily recognised when alone are harder to spot in a crowded visual field
Visual crowding vs acuity
Not just poor acuity in periphery
Multiple object tracking
Ability to track multiple objects as they move simultaneously
Temporal limits of attention
Attentional blink
Rapid serial visual presentation
Attentional blink
Attention selection takes time
Inability to detect a second stimulus straight after the first
Attentional bottleneck
We have limited attention resources that we can use at the same time
Dividing attention
Multitasking
Phones cause 1 in 4 car crashes - hands free devices are no improvement
. Task switching
Attending to the world
Attention guides our perception of the world but we can also use what we know about the world to guide our attention
Object finding in scenes
Easy to find a tap in a picture of a kitchen because it’s a common object in that scene.
Difficult to find a TV in a kitchen scene because it’s not as common.
Fast and efficient scene perception
When shown an array of images, it’s easy to identify if there is an animal present even though we only see each image very quickly.
Scene ensemble statistics
Central tendency of visual properties in scenes
Memory for scenes
When shown a bunch of pictures together for 5 seconds and asked to remember as many as possible, we can remember the general facts such as there was a picture of pigs, but we cannot identify specifics such as how many there were.
Change blindness
When looking at two very similar images, its often difficult to distinguish between them, even when something changes
Change blindness in videos
When watching a video, a woman’s jumper changes colour before our eyes and it is easily missed even when we are looking straight at it.
Real world change blindness
Continuity errors in movies are often missed
When people switch roles during interactions, people often don’t realise that they are talking o someone completely different.
What does change blindness mean for the world?
Our view of the world is relatively sparse but sudden disappearance and change is relatively rare.
Inattentional blindness
Whilst watching the white team and counting how many times they pass the ball, people miss the occurrence of a gorilla in the middle of the screen
Science of magic
Magic is all about misdirecting attention so that actions can be performed unnoticed
