Psych 120B exam 1 material
Terms
undefined, object
copy deck
- Level of representation in Vision
-
Photons
Retina -> lines, edges
V1 -> Orientation, Spatial Frequency
V2 -> Illusory Contours
V3 -> (not sure, Possibly combining information to construct 3D percepts)
V4 -> Complex features, shapes, COLOR
MT (Medial Temporal)->Motion & Stereoscopic depth
IT (Infero Temporal)-> Object form/identity (Shapes, form, faces); specific parts of recognizable objects. - Why Sensation & Perception?
-
All organisms, in order to survive and reproduce, must adapt to their environment
Plants grow to face the leaves to the light (rudimentary systems)
Animals move about (complex-guides locomotion)
In order for an animal to move about successfully (not to fall or be eaten), it must be able to detect the (spatial) structure and (temporal) events in its surroundings
This ability is called PERCEPTION - Morpheus’ answer to Neo in The Matrix, 1999
- What is real? How do you define real? If you’re talking about what you can feel, what you can smell, what you can taste and see, then real is simply electrical signals interpreted by your brain. This is the world that you know.â€
- Sensation
-
In order to detect, an animal has to be sensitive to at least one form of energy that provides information about the environment
Sources of information
Chemicals in the air or water (smell, taste)
Mechanical pressure on body surface (touch)
Sound pressure in the air (hearing)
Perhaps most important of all ⬦light - Light Sensation (vision)
-
Information about distant and silent
environment
inanimate world and silent animals
Most animals use vision for long distance silent environment
except bats, dolphins who use ultrasound - (Visual) Sensation
-
A process
Detecting physical energy, e.g., light from the world
Encoding it into neural signals - Visual Perception
-
A process
Selecting ⬦
Organizing ⬦
Interpreting visual sensory information
Processing RELATIONS in incoming signals
Allowing us to recognize meaningful objects and events - Sensation vs. Perception
-
Traditionally, S is considered as sending RAW DATA to P, the so called bottom-up process
SENSATION^PERCEPTION
But there is debate
Boundary hard to define
SENSATION^PERCEPTION^COGNITION
or Top-down in addition?
SENSATIONvPERCEPTIONvCOGNITION
later processes affect earlier ones -
Sensation and Perception -- Going a bit deeper
Is ignoring this distinction a good idea? -
YES: It does not have all that much practical importance.
NO: We lose some important understanding by doing this. -
Sensation and Perception -- Going a bit deeper
Definitions:
Thomas Reid (1785) -
SENSATIONS: Experiences within the person; sensory
qualities characteristic of our brains
“I feel painâ€
PERCEPTION: Experiences that refer outward to external
objects
“I see a cat†- Intentionality (Brentano, 1838 - 1917)
-
.... is the characteristic of some mental states that
they are about something.
Perceptions are intentional (refer to something out in the world).
Sensations are conscious experiences that do not
refer outward.
Other aspects of cognition can be intentional.
“I believe we are in Los Angeles.†-
Three Reasons to Understand Distinction
between Sensation and Perception -
1) Energy vs. Matter
2) The Mystery of Sensory Qualities
3) Talking to Physicists / Sorting out REALITY - Energy vs. Matter
-
PERCEPTION is mostly concerned with
material properties:
Objects, Spatial Arrangement, Events
We get information about these by means of
energy, but energy is not the focus
Although we do encode energy qualities
in sensations, they are not the
outputs of perception.
Example: We have relatively poor sense of
how much light is coming to our eyes
from a surface - The Mystery of Sensory Qualities
-
Press your eyeball gently=get sensations of light.
"Specific energy of nerves" Johannes Muller (thought nerves determined sense--Wrong, brain area of connection)
Nowadays: cortical projection areas
Color and brightness are in the visual area not in the world. - REALITY
-
Important to Distinguish Qualities in Us
from things in the World
Sensations: Belong to one sense
Perceptions: Usually available in more than
one sense - The Mind-Body Problem
-
Contemporary: “Closet mentalists / dualistsâ€
Mental and physical have different properties (Descartes (1596–1650))
• We have no idea how physical /chemical processes give rise to conscious experiences.
• All of the data we use to form ideas about the physical world come in through perception -- which is “Mind.â€
Contemporary: “Common-sense physicalists" Materialists
Science has made great progress understanding the physical world.
• Laws of physics, chemistry and biology will eventually explain
mind and consciousness. - A Framework for Perception
-
Physical world
Information
Biological Mechanisms
Processes->Algorithms/programs of brain
Representations - Levels of Analysis
-
All three levels are needed to understand
perception (Marr, 1982).
Ecological Level: What is the task? What is the information?
Information Processing Level: What are the algorithms
and computations involved?
Biological Level: How are these processes
implemented in neural systems? - CONNECTIONS across LEVELS
-
PSYCHOPHYSICAL
IND. VBL. IS PHYSICAL STIMULUS
DEP. VBL. IS PERCEPTUAL RESPONSE
PSYCHOBIOLOGICAL
E.G., SINGLE UNIT RECORDING WHILE STIMULATING EYE OF ANIMAL SUBJECT
fMRI
IND. VBL. IS PHYSICAL STIMULUS
DEP. VBL. IS PHYSIOLOGICAL MEASUREMENT - Berkeley
-
Esse est percipi
To be is to be perceived
Despite limitations in perception, all our knowledge comes from it.
Descartes:
What can we reasonably doubt?
Can't trust perceptions.
Dubito ergo cogito; cogito ergo sum - Psychophysics
-
Study of sensations evoked by physical stimuli
e.g., light intensity, mechanical pressure
Accuracy and reaction time are usually the dependent variables - Measurement (Psychophysics)
-
SENSATION: SENSORY LIMITS
WHAT IS DIMMEST LIGHT YOU CAN
DETECT?
WHAT IS SMALLEST DIFFERENCE IN
PITCH YOU CAN DISTINGUISH?
PERCEPTION: WHAT IS OUT IN WORLD?
WHAT IS SMALLEST DEPTH DIFFERENCE YOU CAN SEE?
IS ONE OBJECT BIGGER THAN ANOTHER? - GUSTAV FECHNER (1801 - 1887)
-
German physicist, philosopher
PHYSICIST / NERVOUS BREAKDOWN
⬢ "EXACT QUANTITATIVE RELATION BETWEEN MATTER AND MIND"
⬢ INVENTED MANY OF THE BASIC CONCEPTS WE USE TODAY - THRESHOLD CONCEPTS
-
Absolute threshold
Difference threshold - Absolute threshold (Method of Limits)
-
50% rule, just detectable
Method of limits (e.g., detecting light spot)
Increasing intensity until just visible
Decreasing intensity until just invisible
Back and forth
Take the average, this is the threshold - Absolute threshold (Method of Adjustment)
-
Method of adjustment
Turning a knob up until just visible
Turning the knob down until just invisible
Back and forth
Take the average - Absolute threshold (Method of Constant Stimuli)
-
Use strong lights -- visible
Use very dim lights -- invisible
Use lights in between
Present all above in random order
Fit the curve, and find the intensity that corresponds to 50% (or chance) accuracy
E.g., 2 pencil tips on the skin
Question: “One point or two?â€
Uneven sensitivity, finger tips vs. elbow - What Absolute Threshold isn’t
-
Is there one absolute characteristic of response?
No, fluctuations
Is there an absolute level that is just barely
detectable?
No, a probability of detection that increases with signal intensity - Difference threshold
-
The smallest difference to discriminate 2 stimuli apart
Also called just-noticeable-difference (JND)
The absolute threshold methods apply here as well. - Weber’s law
-
Difference threshold/standard=constant
Change in intensity/intensity=k
Example: JND for weight
Difference threshold = 5g
Standard = 100g
Difference threshold = 10g
Standard = 200 g - Fechner - Weber’s law
-
How to make a scale of magnitude out of JNDs
P = k log S
P = perceptual or sensory magnitude (psychological steps)
S = physical intensity (stimulus - physical steps)
k = a constant - Direct Scaling:Magnitude Estimation
-
When a stimulus is above threshold
Example question:
If we have 2 light spots of different intensities
What should a 3rd light’s intensity be so that it is PERCEIVED subjectively as half way in between?
Physically, it may not be the mid point of the 2 intensities (it is only when it is linear) - Stevens’ Law
-
P = kS^n
P = perceptual or sensory magnitude (psychological steps)
S = physical intensity (stimulus - physical steps)
k = a constant
n = exponent, specific to each sensory dimension - Magnitude Estimation EXAMPLES
-
If light intensity is physically double, does it look twice as bright? – less than doubled (skewed right)
If a pencil is physically doubled in length, does it look twice as long? -- yes (straight 45deg line graphed)
If an electronic shock is physically doubled in intensity, does it hurt twice as much? – more than doubled (skewed left/up) - Magnitude estimation in log units
-
Brightness=almost horizontal (1.1ish to 1.5)
Length=diagonal from bottom left to top right
Shock=almost vertical (bottom a little past 1.0, top around 1.5) - In log-log plot
-
All relations are linear (straight lines)
Light intensity example
Slope of the line is smaller than 1: compression
Pencil length example
Slope of the line is equal to 1: veridical
Electric shock example
Slope of the line is greater than 1: exaggeration - RESPONSE PROBLEMS with threshold estimation
-
VARYING RESPONSE TENDENCIES INFLUENCE THRESHOLD
EXAMPLE: DESCENDING LIMITS
"YES" ON EVERY TRIAL
COOPERATIVE SUBJECT OR TRYING TO LOOK GOOD
“CATCH†TRIAL(old remedy)
Subject BECOMES VERY CONSERVATIVE
NEW RESPONSE TENDENCY:
HAVE TO BE VERY SURE TO SAY YES - SIGNAL DETECTION THEORY
-
SEPARATE SENSITIVITY FROM RESPONSE TENDENCIES
CRITERION (bias) VS. SENSITIVITY
BOTH ARE VARIABLE IN CERTAIN WAYS - SENSITIVITY
-
INTERNAL PERCEPTUAL EFFECT
But, there is always a:
NOISE DISTRIBUTION
Criterion line is under conscious control
Overlap between distributions creates uncertainty
Hit, Miss, False Alarm, Correct Rejection
Improves the more your hit rate exceeds your false alarm rate - "d prime"
- Distance between the peaks of the S+N and N distributions in Signal Detection theory
- ROC
-
Receiving operating characteristic
further ROC curve bends from "d-prime=0" line (top left half) means more sensitive - Electromagnetic spectrum
-
Visible light
400 - 700 nm(1 nm = 10 -9)
Differentially absorbed and reflected by objects in the natural world - Visual Pathway
-
receptors
neural pathways (optic nerve, chiasm)
relay station (LGN)
separate alerting system (superior collicus)
V1 - First step for vision
-
Light enters the eye and is absorbed by photopigment
in receptors (“bleaching†of photoreceptors)
Chemical change brought about by light absorption
triggers electrical signal in photoreceptor.
Subsequent signals for vision are all electrical (neural). -
Preserving Information:
Refraction and focusing of light -
Ray tracing: Light moves in straight lines.
But, refraction by the cornea and lens "bends" the path of light.
How do we figure out where rays go on the retina?
Preservation of directional information (each point on retina gets light from one direction)
Nodal point
Definition: The nodal point of an optical system is the
point through which all undeflected rays pass. - Refraction
- Light bends toward perpendicular "normal" line when entering a new medium and then bends away from the normal, parallel to initial direction, when reentering the old medium.
- Visual Angle
-
Angular measure of extent of image on retina
Visual Angle: Theta=S / D (approximation; good for S <<D) - Retinal image
-
Retinal image is upside down
It is also left-right mirror reversed
So the left world goes to the right retina, and vice versa
This is true for both eyes - Visual Receptors
-
Transduction: Light energy --> Chemical reaction --> Electrical signals
A receptor has many pigment molecules
1 photon (smallest packet of light)
1 photon comes in and is absorbed by retinal, which then changes its shape
This is called isomerization
One isomerized pigment triggers many others, and many others more
The receptor then generates an electronic signal. - Hecht, Shlaer, Pirenne (1942)
-
When a spotlight is just detectable, it has on average 100 photons
50 are reflected off the cornea or absorbed in the eye before reaching the retina
43 reach the retina but pass through
Only 7 hit receptors, and almost certainly hit 7 receptors
So 1 photon suffices to excite a receptor - Retinal cell organization
-
Pigment Epithelium->
Rod and Cone receptors->
Horizontal Cells (lateral inhibition--center-surround receptive fields)->
Bipolar Cells->
Amacrine Cells->
Ganglion Cells
(incoming light) - Dark adaptation
-
Cones get more sensitive faster (3-4 min)
It takes rods 20-30 min
After complete adaptation, rods are more sensitive than cones
So in the dark, rod vision is at work, that is why we see no colors (cones are responsible for color vision) - Spatial Resolution
-
Acuity much better in fovea
Decreases as you move into periphery
Magnification factor, packing density - Rods vs Cones: pooling
-
Rods pool their outputs more than cones
More rods than cones send their outputs to a common ganglion cell
So even when light is dim, a ganglion cell can still receive strong enough a signal from the rods to make a response
The price rod vision has to pay is that it loses spatial details - Organization of V1
-
Retinotopic map
Adjacent things are adjacent
Cortical magnification factor
The retinotopic map is still retinotopic, but distorted
Some locations are ‘expanded’ relative to others
This is due to different cell densities at the retina than at V1 - Cortical Magnification Factor
-
Cells are distributed evenly on the cortex...
...no matter where they came from (fovea or periphery)
The cones on the fovea are packed more densely than in the periphery
Ganglion cells follow this dense packing
There are more of them per unit of space: 50,000 per mm (fovea) vs. 1,000 per mm (periphery)
This means that the fovea takes up more space on the cortex than the periphery since there are more cells to account for. - Cells in V1
-
Simple cells
-orientation
-spatial frequency (scale)
Complex cells
-orientation / spatial frequency
-motion direction
End-stop
-corners
-motion direction - Single Cell Recording
-
The specific visual stimulus that a cell prefers can be found by displaying various stimuli to the eye and recording the electric signal from a cell
When a cell ‘sees’ something it likes, it will begin to fire
In such cases we often say that the cell has become more ‘active’
These spikes in activation can be signaled with sound - Simple cells
-
Respond to bars of specific orientations
This is a different receptive field structure than the center/surround structure of cells in LGN - Complex Cells
-
Orientation specific (like simple cells)
But, they are most active when the oriented stimuli is moving
Some cells respond to motion in any direction, some respond to motion only in very specific directions - End-Stopped Cells
-
Like complex cells that respond to lines of only a certain length
Orientation specific
Motion specific
Length specific - V1 organization: Columns
-
The cortex is organized into ‘columns’ that represent specific values of a feature
Features so far: location, orientation, which eye the signal came from
Types of columns:
-Location columns
-Orientation columns
-Ocular dominance columns - Location Columns
-
Columns of neurons that all have their receptive field on roughly the same location on the retina
Run perpendicular to the cortex: - Orientation Columns
-
Column of cells that all prefer the same orientation
Aligned perpendicular to cortex (like location columns) - Hypercolumns
-
All these types of columns are organized in a systematic fashion
Hypercolumn: complete set of orientation and ocular-dominance columns (right and left eye) for a given retinal location
Like a ‘super-column’ that contains all the column types - Spatial Frequency
-
V1 cells are also tuned to spatial frequency
Similar to orientation tuning
What is spatial frequency?
Essentially, the amount of ‘stuff’ present per unit length
-e.g., More luminance changes per unit length
Something is happening more ‘frequently’ - Frequency and Sine waves
-
Luminance changes can be represented by graphs, like the sine function
Fourier analysis: All images can be mathematically considered as a summation of sine waves of different spatial frequencies
As we add sine waves of higher frequencies, the original (top) wave gets more square - Spatial Frequency: Definition
-
Within one degree of visual angle, the number of cycles a sine wave repeats itself
Remember visual angle...
High frequencies
Sharp edges, details
Low frequencies
Overall shapes, blobs - Adjusting Spatial Frequency
-
Fourier decomposition
By subtracting certain sine waves, we can alter the appearance of natural images
(Groucho Marx slides) - Spatial Frequency and Contrast
-
Contrast
((Highest intensity-lowest intensity)/2)/Average intensity - Contrast Sensitivity Function
-
Contrast Sensitivity Function (CSF)
-The psychophysical contrast threshold is a function of spatial frequency - Beyond V1
-
After V1, information is sent along two separate streams for further processing
Depending on which type of ganglion cell initiated the signal
Ventral Stream
Dorsal Stream
Following these streams will display the modularity of the visual system - “What†and “Where†pathways
-
What (Ventral)
Shape, form
Where (dorsal)
Location (e.g. left vs. right)
More recently described as the How path
We know “how†to interact with something--takes into account size, shape, and location
How path is used to plan action - Visual Pathway:M-ganglion
-
M-ganglion cell
Magno LGN
V1
V2
V3 (with a link to V4)
MT [motion]
Parietal - Visual Pathway:P-ganglion
-
P-ganglion cell
Parvo LGN
V1
V2
V4 [color] (with a link coming from V3)
IT [form]
Temporal - Methods: Double Dissociation
-
How we know one area is responsible for one particular function...
... and not others as well?
Double Dissociation: demonstrate that one action is present and one absent in case A
and that the opposite can be true in case B
TV: busted speaker = picture, no sound
TV: busted glass = sound, no picture
Means these 2 function involve different mechansims - MT: a motion module
-
Module: a particular structure for a particular function
MT seems to primarily process motion
MT does not exclusively process motion: we saw that V1 cells respond to motion as well
90% of MT cells respond to motion, in contrast to less than 10% of V4 cells - IT: a form module
-
Cells in IT respond to patterns
Organized in a columnar structure, like V1
Primary cells respond to simple shapes
Elaborate cells respond to more complex stimuli - IT and Faces
-
Faces are an example of an extremely complex form
Certain cells in IT respond only to any general face or face-like picture
Faces are special to humans...
... possibly special enough to dedicate a special brain area to.
Fusiform Face Area (FFA): a structure that responds specifically to faces
There is some debate that this area is for processing any complex form that demands fine distinction (like breeds of dog), not just faces
Specific IT neurons respond to scenes with faces, respond more to just faces, and respond little to scenes with no faces - Prosopagnosia
-
Failure to recognize specific faces, like those of friends and relatives
A patient knows that they are looking at a face, they just don’t know who’s face it is...
...even if it is their own reflection!
Perception of non-face shapes is ok (single-dissociation) - Ecology of differences:
- What is the relation of differences in brightness and color at the retina to objects and surfaces in the world?
- COLOR
-
Less important than brightness
differences
Experiments: Equiluminant boundaries
Casual observation:
Black and white TV
-Special aesthetic qualities - How do shading Differences Matter?
-
Gradual changes: Could be shadows or subtle variation within a surface
Abrupt changes: More likely to be object edges - Lateral Inhibition
-
Enhancement Mechanism
Discovered by Ernst Mach
-Good example of flow of discovery from psychophysical to physiological model
From his observations, Mach hypothesized that there must be lateral inhibitory connections among receptors.
Hermann grid, Benary cross, Simultaneous Contrast - Edge Detection
-
FIRST STEP IN OBJECT PERCEPTION
⬢ LATERAL INHIBITION IS RETINAL -- enhances raw data that is sent along
WHAT IS FURTHER ALONG? Cortex
⬢ COMPUTATIONAL ANALYSIS
suspected function of early cortical areas
edge anaylzer: sum values of brightness map to get edge map; zero crossings = edges - Some problems of edge detection
-
edge analyzer has trouble with specks and gradual changes
Solution:
Detectors at Multiple Levels of Scale
-A sharp edge is the combination of responses of detectors in the same position at multiple scales. - Brightness
-
PHYSICAL Dimensions
ILLUMINANCE
LUMINANCE
REFLECTANCE
Sensory quality relating to amount of light
“It’s dark in here.â€
Sensory magnitude of brightness related to physical intensity of light is power function (with exponent less than 1) - ILLUMINANCE
-
Amount of light hitting surface per unit time per unit area
candelas / meter squared - LUMINANCE
-
Amount of light reflecting off surface per unit time per unit area
lumen / meter squared - REFLECTANCE
-
Proportion of light hitting a surface that it reflects
Range? 0 ≤ R ≤ 1 - Lightness
-
A persisting property of a surface
Gray / Black / White
Lightness tends to stay the same despite changes in illumination - Brightness vs. Lightness
-
With changes in illumination, the amount of light coming from a surface changes; the sensory quantity that changes is BRIGHTNESS
With changes in illumination, the apparent shade of gray of a surface does not change; this persisting property of a surface is LIGHTNESS - Lightness CONSTANCY
-
The way perceived lightness stays the same despite changes in illumination
a) Distal stimulus is the reflectance of the object
b) Proximal stimulus is the luminance of the object
c) Percept matches the distal, not the proximal stimulus
How??
Perceived lightness depends on ratios of luminance in a scene.
--Wallach's experiment - SIZE CONSTANCY
-
As a person walks away, their image size on our retinas
shrinks. Yet we do not see the person as getting smaller.
a) Distal stimulus is the actual size of the object
b) Proximal stimulus is the size of the retinal projection
c) Percept matches the distal, not the proximal stimulus
->requires taking distance into account - SHAPE CONSTANCY
-
We see object as retaining the same physical shape despite differences in the shapes they project to the retinas (caused by changes in vantage point).
EXAMPLE: A rectangle tilted away from you projects a trapezoidal shape to your eye.
Yet, we see the rectangle as remaining rectangular.
a) Distal stimulus is the actual shape of the object
b) Proximal stimulus is the shape of the retinal projection
c) Percept matches the distal, not the proximal stimulus
->requires taking distance/viewing angle into account - Wallach's Ratio Rule
-
Wallach (1963) present subjects with a grey patch surrounded by a brighter annulus (standard) and a similar configuration in which the annular patch was ten times as bright and asked subjects to adjust the central patch in the second (test) configuration so that it match the appearance of the central patch in the standard display. He found that subjects did not make veridical matches rather they set the grey levels so that the patches had the same grey level ratio as the standard.
Wallach's experiment highlights the role of ratios of luminance in lightness perception. Subjects appear to be matching on the basis of lightness rather than brightness. The ratio rule makes sense if the visual system is more concerned with recovering the reflectance properties of the scene than the luminance. Two pieces of paper, one that reflects all of the incident light and one which reflects 50% of the incident light have the same luminance ratio whatever the level of incident light that falls on them, since both reflectances are multiplied by the same illuminance value, whatever its level. - General Theories of Perception
-
Perception is Indirect
• involves inference
• involves learning in the life of the individual
• involves “taking into accountâ€
Perception is Direct
• involves direct response to relations in the stimulus
• involves mechanisms developed by evolution - Helmholtz’ principle
-
“Those objects are imagined to be present in the field of view that have most often given rise to similar sensations in the past.â€
Perception is an act of imagination!
Inference / Expectation / Motivation - Perception is Direct arguments
-
Most important claim:
Indirect theories underestimated the information available for perception
Instead of inference, guessing, expectation, you can pick up invariants that specify what is going on in the world - Perception is Indirect arguments
-
Stimulus (luminance) changes; therefore perception of constant lightness cannot come from the stimulus.
Must be an inference.
Takes illumination into account. - Hochberg & Beck (1954) experiment
-
Rectangle laying down or trapezoid standing up?
Monocular view: Rectangle lays down --> looks darker
Wave pencil behind: Rectangle stands up --> looks lighter
Does this prove the “taking into account†theory?
Later work: Co-planar ratio principle
Visual system computes lightness based on ratios within planes
CAN INTERPRET THIS AS DIRECT OR INDIRECT - Physics of light
-
White natural light
has all wavelengths with equal contributions
Color of a surface
Is the “color†of light reflected away by the surface (the rest of light is absorbed)
Why is the sky blue, the sun yellow, and setting sun red?
The atmosphere scatters more bluish light - Psychological Dimensions of color
-
Hue
Saturation
Brightness
Hue is related to wavelength (relation is not simple)
Brightness is total amount of light coming to the eye
Saturation relates to the purity or selectivity of hue (how much white is mixed in) - Function of Color Vision
-
Enhances differences
Aesthetics
Not photometry
Any color experience can be achieved by a variety of different wavelengths - Color Mixing
-
Additive: Adding lights
-more light reflected
Subtractive: Adding pigments, paints, etc.
Each pigment reflects a certain band of wavelengths
-more light absorbed - Color Phenomena
-
1) Contrast
-Simultaneous
-Successive
Afterimages
2) Intuitions
-Simple and Compound colors
-Exclusivity
3) Combinations
-Cancellation
4) Color Blindness
-Pairings
--Red-Green; Blue-Yellow
Can't imagine redish-green, blueish yellow, etc.
-Rod monochromats: no cones=no color - Trichromatic theory
-
Young + Helmholtz
At least 3 independent lights are needed to mix in order to generate any color (3 cone types - correct)
Independent
-Mixing any two cannot create the COLOR of the 3rd
Matching experiment - Color vision Physiology
-
3 cone types (S, M, L)
Short-, Medium-, Long-wavelength cones
Different because pigment molecules are different
Differ in spectral sensitivity
One cone cannot give you the wavelength of any wavelength of light (A single receptor’s output can be duplicated by any wavelength light (or wavelength combination) merely by scaling (changing intensity)) - Principle of univariance
-
Only one dimension of output: firing rate…
Related to only one input quantity: Number of photons absorbed
Univariance means there’s only one quantity that can vary.
Consequence is that any stimuli that have the same output on that one dimension of output will be indistinguishable
(by that mechanism alone) - How is a color neurally coded?
-
By the firing patterns of the 3 cone types
Anything that duplicates the outputs of the three mechanisms will match the color experience - Yellow
-
Problems for Young-Helmholtz theory
Contrast: Blue/Yellow
Intuitions: Yellow does not seem to be compound, nor does it seem to be a greenish red.
Combinations: Adding Red + Green light gives gray or white, not yellow.
Color Blindness:How do color-blind individuals with no red or green sensing ability see yellow perfectly well? - Opponent Process Theory
-
Idea: Mapping of mechanisms into color sensations is more complex not one mechanism --> one color
4 Primary Color Sensations: Red, Green, Blue, and Yellow…
… organized into two opponent pairs
Hering’s arguments
People who are blind to red also are to green
People who are blind to yellow also are to blue
Color-normal people can’t see red-green together, nor blue-yellow together
Hard to imagine greenish-red, or bluish-yellow - Hurvich & Jameson (1957)
-
Green can cancel red (and vice versa).
Blue can cancel yellow (and vice versa).
Need to use unique hues
Example: Unique blue: neither reddish nor greenish
Unique hues: 485 nm (blue)
505 nm (green)
580 nm (yellow) - Physiological evidence for Opponent Process Theory
- Opponent neurons (DeValois, 1960)