Test 1 Flashcards
1
Q
Why will you always see the Dalmatian and the Cow?
A
Figure Ground Segmentation, fill in details around what you can now see
2
Q
Process of seeing illusions
A
Top Down
3
Q
Monkey Business Illusion
A
When you are looking for a gorilla you often miss other unexpected events
4
Q
Inverse problem of shape
A
Any 2D image projected onto the retina can come from an infinite number of 3D shapes, visual system has to infer the real shape from many possible shapes
5
Q
Inverse problem of brightness
A
Perceived brightness can come from infinite combos of illumination, surface reflectance and transmittance, visual system must infer reflectance of object
6
Q
Difference between Distal stimuli and Proximal stimuli
A
Distal: Objects or events in the real world
Proximal: Retinal images
7
Q
Introspection to the Problem of Perception
A
Manipulate perceptual experience to generate and test hypotheses by introspection
8
Q
Neuroscience Experiments to the Problem of Perception
A
Measuring brain activity at multiple levels. Lesioning, electrophysiology, imaging, stimulation etc.
9
Q
Behavioural Experiments to the Problem of Perception
A
Device perceptual tasks and measure threshold, accuracy, and/or response time (psychophysics, patient studies)
10
Q
Why does one leaning tower lean more?
A
Railroad example, if parallel but not converging then one should lean more. Visual interprets as 3D, does not happen for 2D images.
11
Q
Cognitively Impenetrable definition
A
Cannot be penetrated by knowledge. The leaning towers illusion is an example of this, face perception is another.
12
Q
Perception as an unconscious inference with faces
A
Both Margret Thatcher’s look similar upside down but when flipped only one is right. When shown whole face identical eyes can appear to show different emotions
13
Q
Definition of Psychophysics
A
Studies the relationship between physical stimuli and psychological/perceptual experience. Involves precise control and allows for insights of underlying mechanisms
14
Q
Define Absolute (detection) Threshold
A
The minimum stimulus intensity that can be perceived “just detectable” e.g. darkest grey or slowest speed
15
Q
Define Relative (difference, discrimination) Threshold
A
The minimum difference between stimulus intensities that can be perceived “just detectable” e.g. differences in grey brightness or slope of lines
16
Q
The three ways that psychophysics measures thresholds
A
Method of adjustment
Method of limits
Method of constant stimuli
17
Q
Method of adjustment - Psychophysics
A
Observers adjust stimulus level until the response changes from seen to not seen or vice versa
18
Q
Method of limits - Psychophysics
A
Gradually decrease/increase stimulus level until observers report change from seen to not seen or vice versa. Approach from both ends to find the crossover point.
19
Q
Method of constant stimuli - Psychophysics
A
show different stimulus levels in random order repeatedly, for each level tally number of yes responses and plot against stimulus level, steepest part of the curve is threshold
20
Q
Psychometric curve
A
Curve of best fit. S curve is characteristic of human responses for most psychophysics experiments
21
Q
Method of forced choice - psychophysics
A
Sidesteps criterion by forcing observers to choose between two or more stimuli (2AFC, 3AFC etc.)
22
Q
2AFC
A
A forced choice. 2AFC means two stimuli on each trial, NOT one stimulus with two choices
23
Q
Weber’s Law Equation
k = dI/I
A
k = Weber’s constant dI = relative threshold I = baseline/reference level
24
Q
Weber’s Law
A
Relative threshold is proportional to background level. This can be generalized across senses. brightness k= 0.08 line length k= 0.03 weight k= 0.02 loudness k= 0.05
25
Fechner's Law
Two signals that are just noticeably different are separated by one unit of perceptual/internal response. Relationship is logarithmic, think of financial status
26
Cornea and Lens
Direct light rays and photons onto retina, non-neural processing
27
What is the optic disc?
Where ganglion cells exit the eye through a hole and form the optic nerve that projects to the rest of the brain. No photoreceptors here make it a blind spot.
28
Major theme of perception
It is an active process, the visual system makes unconscious inferences about the world based on sensory information
29
What is the fovea?
Where we have sharpest acuity. We move our eyes to get fine grained visual details from objects like faces or words.
30
What is macular degeneration?
common form of blindness, 15% of over 65 have it, causes blindness at the fovea.
31
Stabilised images
blood vessels are fixed relative to retina, visual system is not sensitive to images that are stable
32
What are the ganglion cells?
fire when they see things, are the start of neural processing, the retina is an outpost of the brain
33
What are visual angles/degrees?
A standard measure that takes into account size and distance
34
Distribution of cones and rods across the eye
All cones at fovea, rods dominate periphery
35
Rods, cones and light
Cones work best in bright light, as you move away from the fovea you get better acuity in low level illumination
36
connection of rods and cones to retinal ganglion cells: light
Many rods attach to one ganglion while cones almost connect one to one. Ganglion cells need 10 unit of rod/cone responses to fire. Rod pathway is more sensitive to dim light because it sums individual responses of many rods where the cone pathway does not.
37
connection of rods and cones to retinal ganglion cells: sharpness
cones allow for sharper vision. rod pathway cannot tell where exactly photons come from, cone pathway is precise because it preserves information about which individual cone is stimulated
38
Major themes of visual systems
```
Overall Parallel Pathways
Functionally distinct
Anatomically distinct
Complete coverage
Recombine
```
39
Visual themes for rods and cones: functionally, anatomically, complete coverage, recombine
Functionally distinct: Rods for low light levels; cones for high light levels.
Anatomically distinct: Rods and cones have different shapes, although they’re located in same layer of retina.
Complete coverage: rods cover entire visual field except fovea.
Recombine: Rod/cone streams recombine at ganglion cells.
40
Retina ganglion cells: ON-center/OFF-surround
A bright spot in the center increases response; a bright stimulus in the surround inhibits response. Little to no response to a spot of light that covers both center and surround because of cancelation (lateral inhibition)
41
Retina ganglion cells: OFF-center/ON-surround
It gets inhibited by a small spot of light in the center, and it gets excited by a bright annulus in the surround.
42
Midget vs parasol cells
70/80% of ganglion are midget, project to different parts of the LGN. Midget (parvocellular). Parasol (magnocellular).
43
visual themes for midget and parasol cells: functionally, anatomically, recombine
Functionally distinct. Parasol cells get inputs from many photoreceptors, poorer spatial resolution, respond faster to moving stimuli. Midget cells get inputs from fewer photoreceptors, higher spatial resolution but slower processing.
Anatomically distinct. Parasol dendritic trees are big. Midget trees are small.
Recombine. Midget and parasol streams stay segregated in lateral geniculate nucleus (LGN) as parvocellular and magnocellular pathways, recombine later in cortex
44
Primary/retinogeniculate visual pathway
Optic nerve leads from eye to optic chiasm; optic tract leads from optic chiasm to lateral geniculate nucleus (LGN); optic radiation leads from LGN to primary visual cortex (V1)
45
Lateralisation
Right brain gets sensory input and sends motor commands to left body and vice versa. In vision, right brain gets input from left visual field (not left eye!) due to crossover at optic chiasm
46
Retinotopic map in LGN
LGN has 6 layers. All cells are monocular. Layers alternate inputs from two eyes.
Top four are parvocellular layers. Parvo (small) LGN cells get inputs from (small) midget ganglion cells.
Bottom two are magnocellular layers. Magno (large) LGN cells get inputs from (large) parasol ganglion cells.
Ganglion cell axons from the retina preserve their order in LGN.
47
Cortical magnification
Central 5° of visual field takes up 40% of V1. This distortion of retinotopic map in V1 is called cortical magnification.
48
Simple cells in V1
Respond selectively to lines or bars in particular orientations with separate ON and OFF regions. They can be monocular or binocular. They are sensitive to length (bigger response with increasing line length up to a point)
49
Complex cells in V1
Like simple cells, complex cells are orientation selective and are sensitive to length. But they are mostly binocular and have no separate ON/OFF regions. They are sensitive to length (bigger response with increasing line length up to a point)
50
Complex-direction selective cells in V1
Some complex cells show selectivity to both orientation and direction. These cells are important for motion processing
51
Hypercomplex cells in V1
Like complex cells except they obey end-stopping rule: increase response as line gets longer up to optimal length then decrease response if line gets longer still.
52
Direction selective cells
Some V1 cells respond to particular directions of motion.
53
Making simple cells from LGN cells
Each simple cell pools outputs from several LGN cells that have aligned receptive fields. This effectively builds an elongated receptive field from multiple center-surround receptive fields.
54
Ocular dominance
cells are organised systematically based on eye of origin
55
Cortical blindness
V1 damage leads to cortical blindness in the affected visual field (damage to left V1 affects right visual field).
Can show response to stimuli above chance in the blind spot. Blindsight suggests the existence of a subcortical and subconscious visual pathway that bypasses V1
56
How to define a “visual area”: PhACT
Physiology. Cells in different areas handle different visual attributes (motion in MT, colour in V4, objects in IT, etc.).
Architecture. Cytoarchitecture (cell size, cell density, number of layers, density of axons, etc.) can differ between visual areas, helping to identify distinct brain regions.
Connection. Revealed using chemical tracers that are picked up by cell bodies or axon terminals.
Topography. Each of the early visual areas contains a retinotopic map of visual field.
57
seeing the WWF panda
V1 follows what is physically present. V2 activates with what we perceive is there.
58
Ventral pathway goes to the temporal lobe, down from V1
Allows us to recognise shape, size, objects, faces, and words.
Terminates in medial temporal lobe, hippocampus, amygdala.
Contributes mostly to visual recognition, memory, emotion.
59
Dorsal pathway goes to the parietal lobe, up from V1
Handles aspects of the spatial layout such as location, distance, relative position, position in egocentric space, and motion.
Route feeds into motor cortex in frontal lobe.
Contributes mostly to visually-guided action and attention
60
Dark adaptation
Rods and cones change their sensitivity when we move from a well-lit area to a dark place. Cones are more sensitive for 10 mins then rods take over. After 30mins that's the best you will get.
61
Retinal ganglion and V1 cells light contrast
Retinal ganglion and V1 cells responses depend on contrast between light intensity at ON and OFF regions. ON-OFF regions within receptive field.
62
Koffka ring illusion
The grey ring at top left looks uniform, but when the halves are split or misaligned they look different
Continuous object cue wins over the other cues
63
Mechanisms of light 1 and 2
Light normalisation
| 3D Interpretation
64
Mechanisms of light: Light normalisation
Factors out average luminance, compute relative contrast of surfaces
65
Mechanisms of light: 3D interpretation
Brightness perception considers real world 3D factors such as geometry and shading
66
Colour in the periphery
We are colour blind in the periphery, it is cones that perceive colour
67
What is light?
Stream of photons, wave energy. Wavelength is the distance between the peaks of the waves
68
Light and selective transmission
A transparent object will have its colour determined by what light is able to go through the object
69
Cones, wavelengths and colours
Mostly red cones which see longer wavelengths, a lot of green cones for middle wavelengths, a few blue cones for short wavelengths
70
Subtractive mix of light
happens with paint etc, each pigment has its own selective reflection. Fewer wavelengths are reflected as each paint subtracts wavelengths from the mix.
71
Additive mix of light
Each additional light adds wavelengths so more wavelengths are reflected
72
Define metameters
Light. the two fields are physically different but are perceptually identical
73
Trichromatic Theory
percentage of light absorbed by each cone, 3 cones cover the whole spectrum. Colour perception is comparison of response of each cone. These ratios stay the same even if light intensity changes therefore colour is constant
74
Opponent process theory
Light. 3 opponent mechanisms, responding in opposite direction to different wavelengths.
75
Colour constancy in the brain
V1 responds to the wavelength of the stimulus. V4 responds to the colour in context.
76
Mechanisms of light: Lateral Inhibition
Retinal ganglion cells don’t care much about uniform fields, what they care about (and enhance) are borders and edges
77
Tilt aftereffects: v 1 mapping
population response is given by the shape of the bars. neurons will adapt and become tired. A similar tilt will cause less activation in the tired neurons and others will be able to respond more, hence seeing an incorrect angle.
78
Orientation selectivity
V1 cells respond maximally to lines/bars in select orientations but weakly or not at all to other orientations. The gradual decrease of response as stimulus tilts away from preferred orientation forms the cell’s tuning curve.
79
Difference between high and low spatial frequency
Low -> things that change gradually
| High -> part that gives fine details, detects variation in luminance that is abrupt and sudden
80
Characteristics of sine wave gratings
Frequency: How many cycles within a space
Contrast: How different the light and dark bars are
Orientation: Which way the grating is tilted
Phase: Does it start with a light or dark bar
81
Fourier transform
We can create any image by simply summing multiple sine wave gratings. Less info to store this way. There is a pathway more sensitive to coarse things, another more sensitive to higher resolution
82
Contrast sensitivity function
Humans are most sensitive at 6-8 cycles/degree (about 12 black/white stripes in a thumb at arm’s length); gradually less sensitive to lower/higher frequencies. The highest frequency you can see defines your visual spatial acuity.
83
Spatial frequency adaptation and channels
adaptation -> Reduced sensitivity in CSF but only for gratings near the adapting frequency
Each channel is sensitive to a relatively narrow range of frequencies but together they add to create the CSF
84
Perceptual organization
| How do we start getting things out of stuff?
Segmentation i.e. figure ground separation
| Grouping: Proximity, similarity, continuity, closure
85
Perceptual organization figure cues
Classical: Convexity, smaller area, symmetry, enclosure
| Non-classical: Familiarity, wider base, lower region, protrusion
86
Monocular cues of depth
```
Occlusion
Texture
Shadow casting
Aerial view
Shading and contour
Linear perspective
Motion Parallax
```
87
define binocular/retinal disparity
Difference in image location of an object on left and right eyes because of the eyes’ horizontal separation
88
Horopter
Imaginary surface that includes fixation and other locations in space that produce corresponding retinal points. At the Homoptera disparity is zero.
89
Crossed and uncrossed disparity
All non-corresponding retinal points (crossed (negative)for near objects, uncrossed(positive) for far objects) produced by non-horopter locations
90
Correspondence problem
What features or objects in the retinal images come from the same real world objects
91
What is a Stereoscope/stereogram
Present two offset 2D images to the left and right eye of viewer; brain combines them to create depth perception.
92
Random dot stereogram
Shows that depth can be computed without other cues (e.g., perspective, motion parallax) and that binocular fusion can happen before form/object is perceived.
93
Binocular rivalry
What happens if we present very different images to the two eyes? We will only see one at a time, not a merging
94
Binocular or disparity-selective cells
This cell responds to a line shown to both eyes with the right orientation, direction of motion, and binocular disparity
95
green red binocular rivalry
When it is green there is an increase in signal. For every red band it goes down. Can dissociate physical from perceptual but not as strong as physical.
96
Binocular rivalry in V1
V1 neurons ramp up response when higher contrast grating is seen and ramp down when lower contrast grating is seen.
97
Binocular rivalry in higher visual areas
Neural response in face/place areas tracks perceived object type not physical stimuli on retina. Does not care if it is on top of each other, cares about what is in your head
98
Ames room
Two people are assumed to be at same distance (when they’re actually not), so you infer a giant and a dwarf. You can only view the room as a rectangle.
99
motion after effects
motion will transfer between eyes but not colour
100
specific cues for distance and depth
Motion parallax, stereokinetic effect, cast shadow, figure ground segmentation
101
Inflow vs outflow models of motion analysis
Inflow model compares retinal data to feedback from eye muscles; outflow model compares them to brain commands.
102
Motion blindness (akinetopsia)
like seeing the world in snapshots, like watching a movie missing multiple frames
103
Area MT (middle temporal) or V5
A secondary or extrastriate visual area that plays a major role in motion perception, especially in processing global motion information from local motion signals
104
Single-cell studies in MT
MT cells are selective for motion directions. Strikingly, their firing patterns can account for motion perception threshold in psychophysics studies.
105
Aperture problem
Motion direction signal in early neurons with smaller receptive fields (e.g. V1) is ambiguous
106
Motion-induced position shift
Bottom patches appear shifted towards direction of motion.
107
Tennis ball outs, physics
more likely to make predicted error in direction of motion
