NEU 325 - Exam #1 Flashcards
What is perception?
There is stuff in the world, and our viewing of it forms percepts.
Essentially the process for:
- extracting information via the senses
- forming internal representations of the world
Epistemology
The theory of knowledge
Where does knowledge come from? (2 answers)
- Psychological Nativism: the mind produces ideas that are not derived from external sources
- Empiricism: all knowledge comes from the senses (A newborn is a ‘blank slate’)
Proponents: Hobbes, Locke, Hume
Resembles the nature vs nurture debate – extreme positions at both ends are a bit absurd
Metaphysics
Theory of reality
What kind of stuff is there in the word (2 answers)
- Dualism
- there are 2 kinds of stuff, usually ‘mind’ and ‘matter’ - Monism
- there is only one kind of stuff, “materialism” (physical) and “idealism” (mental)
What did René Descartes believe?
- undertook a program of ‘radical skepticism’: decided to discard any idea that could be doubted’
- a senses can be fooled
- all sense data could be caused by an ‘evil demon’
- concluded that the only thing he could be certain of was that he existed (I think, therefore I am)
Did Descartes believe in Dualism or Monism?
Dualism: there are two kinds of stuff
- mental stuff (non-spatial, non-physical)
- physical stuff (possesses no mental properties)
There was a problem though, because how can the physical and mental stuff interact?
What are the modern versions of dualism?
“homunculus” - little man
- there is a person who sits inside our head and is responsible for ‘perceiving’ what we see
- prevalent conception in neuro today
What is the relationship between ‘things in the world’ and ‘representations in our heads’? (2 answers)
- Naive Realism: we perceive the world as is, our minds have direct access to reality
- (Epistemological) Idealism: the only reality is that of mind/ideas; there is no evidence for/reason to believe in an external world
- supported by Bishop Berkeley, idealist, empiricist
What is the modern variant of Epistemological Idealism?
- brain in a vat
What is representative realism?
We perceive the external world indirectly and imperfectly, via intermediate ‘sense data’
What is the philosophical position of the course?
- empiricism: knowledge from senses (obviously)
- materialism: only one kind of stuff (matter/energy)
- representative realism: indirect knowledge of world, via the senses
- functionalism: understanding the ‘function’ of the sensory systems is all we need to know to ‘understand’ them
What does ‘understanding perception’ mean in this course?
- we can write down an algorithm for how a perceptual task is performed
- knowing where and how the algorithm is implemented in the nervous system
Why is naive realism wrong? (Include reason, and type of illusions)
- lightness illusion, and comparison patch illusion
- the fact that we are sometimes mistaken in our sensory perceptions indicates that we do not directly perceive the world
- thus naive realism is false!
What is the importance of eye movements?
- we aren’t aware of them, but they are essential for vision
- if you stabilize the eye, you become blind within several seconds
____ is also critical for perception
Top-down information (e.g. memory)
What is happening when there illusions arise from conflicting information?
- brain sometimes comes up with entirely new percepts in response to conflicting sensory cues
How do cochlear implants work?
In the ear there is a microphone that takes it to a transmitter and receiver that sends the electrode array in the cochlea. The cochlea then gets that information and sends it on it’s way to the brain
What are the conclusions from the arguments about illusions?
- perceptual representations are not always accurate (naive realism is wrong)
- we can understand why the world looks the way it does by studying the algorithms/computations used to generate percepts
- if we understand the computation, we can replace neural circuits with computer chips
What are the methods of study for perception?
Ecological (phenomenological, naturalistic)
Psychophysical
Neurophysiological
Modeling/Reverse Engineerings
What is the ecological approach?
- observe and draw some conclusions
- use of rich, naturalistic stimuli
- emphasizes the environment in which the system evolved, developed, and lives
What is the main strength and main weakness of the ecological approach?
main strength: takes in richness of sensory behaviors and the evolutionary constraints that shaped them
main weakness: lack of scientific rigor in stimulus control and self-report data (which can be unreliable)
What is the psychophysics approach?
- use of carefully controlled laboratory stimuli
- carefully measured quantitative data
- scientific theory of the relationship between mind and matter
- founded by Gustav Fechner (co-founder of experimental psychology)
What are the main strengths and main weaknesses of the psychophysics approach?
Main strength: scientific rigor, non-invasive
Main weakness: use of impoverished stimuli that are rarely if ever encountered in nature
What is the neurophysiology approach?
- use of carefully controlled laboratory stimuli (like psychophysics)
- measures the response of neurons or groups of neurons at various locations in the perceptual system
What is a weakness of neurophysiology?
Sometimes it is not meaningfully connected to behavior/perception
What is computational modeling?
- developing quantitative theories or computer simulations to predict behavior or neural function OR
- construct artificial systems that perform like human perceptual systems
What is Fechner’s Law?
S = klogR
S is sensation intensity, R is physical stimulus intensity
What is Weber’s Law?
- law about how stimulus intensity relates to detectability of stimulus changes
- as stimulus intensity increases, magnitude of change must increase proportionately to remain noticeable
What is Weber’s Fraction?
- ratio of change magnitude to stimulus magnitude that is required for detecting the change
(dR)/R
dR = change in stimulus
R = stimulus intensity
What is the Just-Noticeable Difference?
- smallest magnitude change that can be detected
What happens when you differentiate both sides of Fechner’s law?
You get Weber’s Law!
So detectability (“how much the percept changes”) is determined by the ratio of stimulus change (dR) to stimulus (R)
It’s really the same law –> Weber-Fechner Law
The membrane of neurons are ____, which means________….
Polarized; voltage difference between inside and outside (neuron is like a battery)
What happens where there are “spikes” in the nervous system?
- spikes are like the currency of the nervous system
- channels open, the current flows in, and the mebrane becomes depolarized. This raises the membrane potential
What happens at the synapse of a neuron?
- action potential triggers release of vesicles
- transmitter molecules bind to receptor
- post-synaptic electrical signal
How can you measure neural activity?
Invasive methods:
- electrophysiology (electrodes)
- imaging (voltage sensitive dyes)
Non-invasive methods
- fMRI (functional magnetic resonance imaging)
- EEG (electroencephalography)
- MEG (magnetoencephalography)
What is Stevens’ Power Law?
S = kR^b
What are the problems with Stevens’ Power Law?
- subjective
- based on rating data
- no ‘right’ answer: just a mapping between one unknown scale (‘pain’) and another unknown scale (‘numbers’)
How can you measure perception?
A percept is internal
So psychophysics tells us that you can use:
- detection (yes/no)
- discrimination (e.g. bigger than)
- estimation (report the stimulus exactly)
And all provide indirect measure of internal mental state
What is the psychometric function of detection?
Relates to the physical quantity of sound/percept to the probability of detecting it
Describes probability of saying “I heard it” as a function of stimulus intensity
What is signal detection theory?
A psychophysical theory that quantifies the response of an observer to the presentation of a signal in the presence of noise
What are the four responses to percept under the signal detection theory?
Hit: stimulus is presented and observer responds “yes”
Miss: stimulus is presented and observer responds “no”
False alarm: stimulus is not presented and observer responds “yes”
Correct rejection: stimulus is not presented and observer responds “no”
What is ‘noise’ distribution?
values arising when stimulus not present
What is “signal” distribution?
values arising when signal + noise is present
What is Type I error?
Rate of ‘false alarms’, or false positives
What is Type II error?
Rate of ‘misses’, or false negatives
What is light, waves, and particles?
Light: electromagnetic radiation within a narrow energy range
Wave: can be bent by lenses
Particle: “photons” - can travel through a vacuum, have minimum that can be emitted/absorbed (quanta)
What happens to light?
It is:
- absorbed: energy (e.g. light) that is taken up, and is not transmitted at all
- scattered: energy that is dispersed in an irregular fashion (most light does this!)
- refracted: energy that is altered as it passes into another medium (e.g. light entering water from the air)
- reflected: energy that is redirected whe nit strikes a surface
Why do we need optics?
- without optics, light from everything hits the whole retina/screen/film
- with optics, we form an image
- i.e. light from a single point in space hits a single spot on the retina
Pinhole Camera Problem
- the larger the pinhole, the blurrier the image
- smaller pinhole = sharper image
Smaller pinhole
- fewer rays
- sharper image
What is diffraction
The bending of waves around small obstacles or through small apertures
What is the cornea?
The transparent ‘window’ into the eyeball (carriers 2/3) of eye’s total refractive power
What is the aqueous humor?
Watery fluid behind cornea
What is the lens?
Allows changing of focus in the eyeball
What is the pupil?
the dark circular opening at the center of the iris in the eye, where light enters the eye
What is the vitreous humor?
transparent fluid that fills main cavity of the eye (gel-like; may contain ‘floaters’)
What is the retina?
light-sensitive membrane in the back of the eye that contains rods and cones
What is the photic sneeze reflex?
tendency to sneeze when walking from a dark room into bright light
- affects 18-35% of the population
- usually 1-10 sneezes followed by refractory period
- aka ‘pepper on the sun’
Explanations:
- Aristotle thought the ‘sun heats the nose’
- Bacon thought when you closed your eyes, you didn’t sneeze
- current theory: ‘crossed wiring’ in the trigeminal nerve
How is an image formed with the lens in the eye?
The goal is to focus the light rays emanating from a single point to a single point on the imaging surface. The eye converges it.
What is refraction?
Necessary to focus light rays, carried out by the lens; includes:
- accommodation: process in which the lens changes its shape, altering its refractive power
- emmetropia: no refractive error
What are the refractive errors in vision?
Near-sightedness - myopia
- eye is too fat and powerful because the eye is too long
- can’t get far objects into focus
Far-sightedness - hyperopia
- eye is too thin/not enough accommodation because the eye is too short
- can’t get near objects in focus
What is astigmatism?
Visual defect caused by the unequal curving of one or more of the refractive surfaces of the eye, usually the cornea; so if you have an astigmatism, some lines will be blurrier
What is the camera analogy for the eye?
The aperture (f-stop) = iris/pupil
- regulates the amount of light coming into the eye
Focus = lens
- changes shape to change focus
Film = retina
- records the image
What does the retina do in the eye?
- Transduction
- Processing
What is the process of transduction, as it happens in the eye?
Conversion of energy from one from to another (i.e., ‘light’ into ‘electrical energy’)
What is the process of processing (lol) as it happens in the retina?
- amplification of very weak signals (1-2 photons can be detected)
- compression of an image into more compact form so that information can be efficiently sent to the brain; optic nerve = ‘bottleneck’
What is the process of sending information to the brain from the retina like?
Light comes in from the eye, and hits the retina. From the cone, it goes to the bipolar cell, which then takes it to the retinal ganglion cell. The retinal ganglion cell then takes it to the optic disc (blind spot), which then takes it to the optic nerve
What is crazy about the process of light transferring to the retina?
It has to pass through all the other ‘junk’ in our eye before getting to the photoreceptors
Cephalopods (squid/octopus) have their photoreceptors in the innermost layer
Could be:
- accident of evolution
- may be better to have photoreceptors near the blood supply
What is phototransduction?
converting light to electrical signals
What are rods and cones?
Rods:
- respond in low light (“scotopic”)
- only one kind: don’t process color
- 90M in humans
Cones:
- respond in daylight (“photopic”)
- 3 different kinds: responsible for color processing
- 4-5M in humans
What are the outer segments of rods and cones in phototransduction?
- packed with discs that have OPSIN in them (proteins that change shape when they absorb a photon)
- different opsins are then sensitive to different wavelengths of light
- rhodopsin are opsin in rods
What is photopigment?
A general term for molecules that are photosensitive (like opsins)
What is the dark current
In the dark, membrane channels in rods and cones are open by default (unusual!)
- current then flows in continuously
- the membrane is depolarized (less negative)
- neurotransmitters are released at a high rate
How is transduction & signal amplification different from the dark current process?
In the light, photon is absorbed by an opsin
- the channels close (dark current turns off)
- membrane becomes more polarized (more negative)
- neurotransmitters are released at lower rates
- this is a graded potential, not spikes
Photoreceptors are not ____ distributed across the retina - why, and what are the implications?
Evenly
Why: fovea is mostly cones, periphery is mostly rods
Implications: not much color vision in the periphery and highest sensitivity to dim lights
What is the visual angle?
The size an object takes up in your retina (in degrees)
- vision scientists measure the size of visual stimuli by how large an image appears on the retina rather than how large the object is
What are ‘on’ and ‘off’ ganglion cells?
On: responds when the light hits the center
Off: responds when the light hits the surrounding space
What is a receptive field?
What makes a neuron fire – weighting function that the neuron uses to add up its inputs
What are the different types of retinal ganglion cells?
On and off retinal ganglion cells’ dendrites arborize (‘extend’) in different layers
PARVOCELLULAR: ‘small’, feed pathway processing shape and color
MAGNOCELLULAR: ‘big’, feed pathway processing motion
What are the 4 channels in visual processing?
- ON, M-cells (light stuff, big, moving)
- OFF, M-cells (dark stuff, big, moving)
- ON, P-cells (light, fine shape/color)
- ON, P-cells (dark, fine shape/color)
RETINA = on or off, p-cells or m-cells
OPTIC NERVE = light/dark or big/fine or moving/color
Luminance adaptation
- there is an incredible range of luminance levels to which we can adapt
- two mechanisms for luminance adaptation:
1. pupil dilation
2. photoreceptors and their photopigment levels (the more light, the more photopigment gets ‘used up’ –> less available photopigment, retina becomes less sensitive
What is luminance adaptation from light to dark
We are pretty bad at estimating the overall light level
- all we really need (from evolutionary standpoint) is to be able to recognize objects REGARDLESS of the light level
- this can be done using light differences, also known as: contrast
- contrast: difference in light level, divided by overall light level
C = (ΔI)/I
- from an ‘image compression’ standpoint, it’s better to just send information about local differences in light
How does the brain begin to process information to extract a visual interpretation?
Early visual pathway, and specifically the lateral geniculate nucleus (LGN) in the thalamus AND primary visual cortex (V1) in the cortex
What is topography?
Mapping of visual space onto visual cortex
Contralateral representation
Each visual field (L/R) represented in opposite hemispheres, as opposed to ipsilateral representation
Cortical magnification
Unequal representation of fovea vs. periphery in cortex
Acuity
Measure of finest detail that can be resolved
How do you measure Visual Acuity?
Smellen E test
- the strokes on the E form a small grating pattern
What kinds of circuits support acuity?
Sensitivity in rods and cones
What is acuity in the retina?
- it is a striped pattern in a ‘sine wave grating’
- visual system “samples” the grating at cone location
Visual acuity declines in an orderly fashion with _____, which is the ______
Eccentricity; distance from the fovea
What is the lateral geniculate nucleus (LGN)
One on each side of the brain; this is where axons of retinal ganglion cells synapse
Organization:
- represents contralateral visual field
- segregated into eye-specific layers
- segregated into Magnocellular and Parvocelluar layers
What is the primary visual cortex?
- also known as the striate cortex, or V1
- it is the first place in cortex where visual information is processed
- the previous two stages (retina & LGN) are pre-cortical
What do LGN cells respond to?
They respond to either one eye or the other, never to BOTH
What do V1 cells respond to?
Can respond to input from both eyes, but still tend to have a preferred eye — they spike more to input from one eye
What is the major change in sensory representation in V1?
- retina & LGN: circular RFs, and 1M fibers from RGCs
- V1: elongated, orientated
Orientation tuning
Neurons in V1 respond more to bars of certain orientations
- response rate then falls off with difference from preferred orientation
What do many cortical cells in receptive fields in V1 respond well to:
- moving lines
- bars
- edges
- gratings
- direction of motion
What is ocular dominance?
Cells in V1 tend to have a ‘preferred eye’ (respond better to inputs from one eye over the other
Simple vs Complex Cells
These are cells in V1 that respond best to bars of light rather than to spots of light
- ‘simple’ cells: prefer bars of light, or prefer bars of dark
- ‘complex’ cells: respond to both bars of light and dark
What is a column, orientation column, and ocular dominance column?
Column: a vertical arrangement of neurons
Orientation column: for a particular location in the cortex, neurons have same preferred orientation
Ocular dominance column: for a particular location in the cortex, neurons have the same preferred eye
What is a hypercolumn?
It contains all possible columns
- 1mm block of V1 containing ‘all the machinery necessary to look after everything the visual cortex is responsible for, in certain small part of the visual world’
- each one contains a full set of columns — has cells responding to every possible orientation, and inputs from left right eyes
What is spatial frequency?
The number of cycles of a grating per unit of visual angle (usually specified in degrees)
- think of it as # of bars per unit length
Fourier decomposition
The visual system breaks down images into sine waves of different frequencies
- mathematical decomposition of an image (or sound) into sine waves
What is the “Fourier Decomposition” Theory of V1
The role of V1 is to do the Fourier Decomposition, i.e. break images down into a sum of sine waves
- summation of two spatial sine waves
- any pattern can be broken down into a sum of sine waves
Retinal Ganglion Cells are tuned to _____
Spatial frequency
- low frequency yields weak response
- medium frequency yields strong response
- high frequency yields weak response
What is the tilt after-effect?
- perceptual illusion of tilt, provided by adapting to a pattern of a given orientation
- supports the idea that the human visual system contains individual neurons selectively for different orientations
Adaptation
The diminishing response of a sense organ to a sustained stimulus
- an important method for deactivating groups of neurons without surgery
- allows selective temporary ‘knock out’ of group of neurons by activating them strongly
Selective adaptation alters _____ and _____
Neural responses; perception
What is the perceptual effect of adaption?
Repulsion away from the adapter
What is adaptation for spatial frequency?
It is evidence that the visual system contains neurons selectively for spatial frequency
What is the orthodox viewpoint of selective adaptation?
- if you can observe a particular type of adaptive after-effect, there is a certain neuron in the brain that is selective (or tuned) for that property
- THUS (for example): there are no neurons tuned for spatial frequency across all orientations, because adaptation is orientation specific c
What are ways to study the vision of infants who can’t speak yet?
- Preferential-looking paradigm: infants prefer to look at more complex stimuli
- Visually evoked potentials (VEP): measure brain’s electrical activity directly
What is difficult about studying the development of spatial vision?
Young children are not very sensitive to high spatial frequencies
visual system is still developing
- cones and rods are still developing
- retinal ganglion cells still migrating and growing connections with the fovea
- fovea: not fully developed until 4 years of age
What are the relevant brain areas for object vision?
The extrastriate cortex: aka the general term for regions outside V1
- V2, V3, V4, Inferotemporal Cortex, etc
What is the view-dependent model?
A model that will only recognize particular views of an object
So also a template-based model
What is the problem with the template-based model?
You need a neuron (or ‘template’) for every possible view of the object — quickly run out of neurons
What is middle vision?
After basic features have been extracted and BEFORE object recognition and scene understanding
- involves perception of edges and surfaces
- determines which regions of an image should be grouped together into objects
How do you find the edges of objects?
Cells in primary visual cortex have small receptive fields
What is the issue for computer-based edge detectors?
Not as good as humans
- they find TOO many edges
Illusory contour
A contour that is perceived even though no luminance edge is present
What is Gestalt Psychology?
“Whole is greater than the sum of its parts”
- opposed to other schools of thought (e.g. structuralism) that emphasize the basic elements of perception
What do structuralists say about perception?
- perception is built up from ‘atoms’ of sensation (color, orientation)
- challenged by cases where perception seems to go beyond the information available (e.g. illusory contours)
Gestalt grouping rules
A set of rules that describe when elements in an image will appear to group together
What is good continuation?
A gestalt grouping rule that states that two elements will tend to group together if they lie on the same contour
What are the gestalt grouping principles?
Similarity AND proximity
What are the gestalt figure ground assignment principles?
- surroundedness: surrounding region is likely to be group
- size: the smaller region is likely to be figure
- symmetry: a smaller symmetrical region tends to be seen as figure
- parallelism: regions with parallel contours tend to be seen as figure
- extremal edges: if edges of an object are shaded such that they seem to recede in the distance, they tend to be seen as figure
What is accidental viewpoint?
It produces an regularity in the visual image that is not present in the world
The visual system will not:
Adopt interpretations that assume an accidental viewpoint
What is a non-accidental viewpoint?
A ‘typical’ viewpoint interpretation that won’t change if you move the camera a little bit
What are t junctions, y junctions and arrow junctions?
T junctions: indicate occlusion
Y junctions: indicates corners facing the observer
Arrow junctions: corners facing away from the observer
All of these features are still present if object is shifted, scaled, or rotated by a small amount
Viewpoint Invariance
The idea that we should be able to recognize an object from any viewpoint
Geons
“recognition-by-components” model of object recognition
- visual system identifies objects by recognizing component shapes (‘geons’) that compose
What is the problem with view-invariant theories?
Object recognition is not completely viewpoint invariant
- viewpoint does affect object recognition
- the farther an object is rotated away from a learned view, the longer it takes to recognize it
What are the main takeaways from viewpoint invariance?
Object recognition is somewhat but not entirely viewpoint invariant
- observers do seem to store certain preferred views of objects
- makes sense from an evolutionary standpoint: we generate representations that are as invariant as we need them to be for practical applications
What are the two facts that constrain any models of object recognition in the visual system?
- Visual processing divided two cortical streams:
- separate pathways for ‘what’ and ‘where’ information
Dorsal = where pathway
Ventral = what pathway
AND
- Object recognition is fast
What is Area V4?
- cells tuned to stimuli such as spirals, pinwheels, concave and convex shapes
- difficult to know what V4 neurons do/what stimuli drive them best (but not simple spots or bars!)
What is the inferotemporal cortex (IT)?
Receptive field properties:
- very large – some cover half the visual fields
- don’t respond well to spots or lines
- do respond well to stimuli such as hands, faces, or objects
- when IT cortex is lesioned, it leads to agnosias (e.g. object agnosia, prosopagnosia)
What is agnosia?
Failure to recognize objects in spite of the ability to seem them
How can you identify brain regions associated with object recognition?
Functional imaging (fMRI) decoding method:
- train a computer to identify images using functional images of brain activity
- then examine which brain areas allow for objects to be decoded most accurately
- reveals good decoding of objects from IT and other temporal lobe area
What is a grandmother cell?
It is the idea that a single neuron responsible for representing some complex object (e.g. your granny)
- long considered ‘idea that could never work’
- how could you have a different neuron for every possible object you know how to recognize?
- what if that neuron died? Could you still recognize your grandmother?
What is the IT cortex and grandmother cell crossover?
- high selectivity to people/things, independent of viewpoint
What does it suggest about object recognition being fast?
Suggests operation of a feed-forward process
What is a feed-forward process?
computation carried out one neural step after another, without need for feedback from a later stage
- this is still debated, but it’s agreed there’s not much time for feedback
What is the pandemonium model of object recognition?
- Oliver Selfridge’s (1959) simple model of letter recognition
- perceptual committee made up of ‘demons’ – demons loosely represent neurons. Each level is then a different brain area
What is the hierarchical ‘constructive’ models of object recognition?
- explicit description of how parts are combined to form representation of a whole
Metaphor example: ‘committees’ forming consensus from a group of specialized members
- perception results from the consensus that emerges
What is the modern version of object recognition?
There are deep neural networks
- the last 10-20 years have seen rapid progress in ‘deep learning’ methods for object recognition & scene understanding
What is a deep-learning based approach?
“task based” or “goal based” approaches
1. train a network (CNN/DNN/RNN) to perform the task
2. regress units in trained network against neural data
- recent work emphasizes pre-trained networks
- use of RNNs/LSTMs/GRUs to capture time-course of responses
- current debate about whether we can ever ‘understand’ V1 (or whether that is even a worthwhile goal)
What is the physics background to know about light and color?
- most of the light we see is reflected
- typical light sources: sun, light bulb, LED screen
- We see only part of the electromagnetic spectrum (between 400 and 700 nm)
How many numbers would you need to write down to specify the spectral properties of a light course?
It depends on how you ‘bin’ up the spectrum
- one number for each spectral ‘bin’
What is a hyper-spectral camera?
- measures amount of energy in each range of wavelengths
- can use thousands of bins (or ‘frequency bands’), instead of just the 13 shown here
What is a spectral, illuminant, and illuminant power spectrum?
Spectral - referring to the wavelength of light
The illuminant - light source
Illuminant Power Spectrum - this curve – the amount of energy (or power) at each frequency
What is the illuminant with the most power at long wavelengths?
A reddish light source
An illuminant with most power at medium wavelengths?
A greenish light source
An illuminant with power at all visible wavelengths
a neutral light source, or ‘white light’
How many measurements of this same spectrum does the human eye take (in bright conditions?)
Only 3! One from each cone class
What are the cone types?
S = short (blue)
M = medium (green)
L = long (red)
What does color vision rely on?
Color vision relies entirely on comparison of responses from three cone types (S, M, L)
What is the absorption spectrum?
Describes response (or ‘light absorption’) of a photoreceptor as a function of wavelength
What is the problem with the absorption spectrum?
Response from a single cone is a ambiguous, because all the photoreceptor gives you is a ‘response’
- can’t tell which light frequency gave rise to this response (blue or orange)
- plus, can’t tell a weak signal at the peak sensitivity from a strong signal at an off-peak intensity?
- all 3 of these lights give the same response from this cone
Cone response = ______ x ______
Cone response = absorption spectrum x light intensity
What is the problem of univariance?
Infinite set of wavelength + intensity combinations can elicit exactly the same response
- so a single cone can’t tell you anything about the color of light
Metamers
Illuminants that are physically distinct but perceptually indistinguishable
- cone sensitivities define a 3D subspace of color perception
- metamers differ only in the null space!
What is the implication of metamers and the project of univariance?
- many things in the natural world have different spectral properties, but look the same to us
- that wouldn’t be the case if we had more cone classes
How did they figure out that there were 3 ‘primary’ lights and that any color can be made by combining 3 suitable lights?
James Maxwell’s color-matching experiment
- given any ‘test’ light, you can match it by adjusting the intensities of any three other lights (2 is not enough; 4 is more than enough)
What is trichromatic color vision?
Three lights needed to make a specific color percept, due to use of 3 distinct cones with different sensitivities
- colors uniquely defined by combinations of cone activations
What did Isaac Newton show?
That white light can be separated into colors w/ a prism - and then he reassembled to make white light w/ a second prism
What was Newton’s Theory
- there are seven kinds of light –> seven kinds of photoreceptor
BUT NEWTON was totally WRONG
What is Color Space?
A 3D space that describes all possible color percepts
What are the ways to describe that space?
- RBG color space: defined by the outputs of Long, Medium, Short wavelength (or R, G, B) lights
- cone responses entirely determine our color percepts
- ‘non-spectral hues’: a percept couldn’t be produced by any single-wavelength light - HSB color space: defined by hue, saturation, and brightness
What is hue, saturation, and brightness?
- Hue: the chromatic (color) aspect of light
- Saturation: the chromatic strength of a hue
- Brightness: the distance from black in color space
What is 2D slice of HSB color?
- hue around the edge
- saturation increasing from center to edge
- brightness not shown
What is the opponent color theory?
Perception of color is based on the output of three channels, each based on an opponency between two colors
What are the opponent channels?
- L-M (red-green)
- S- (L+M) (blue-yellow)
- L+M - (L+M) (black-white)
Some retinal ganglion cells have _____ with ‘color-opponency’?
center-surround receptive fields
for example: red-green (L-M) color-opponent cell
- carries info about red vs. green
What is a (Negative) Afterimage?
- Visual image seen after a stimulus has been removed
- polarity is the opposite of the original stimulus
- colors are complementary:
— red produces green afterimages
— blue produces yellow afterimages
— light stimuli produce dark negative afterimages
What properties of an object determine the properties of the reflected light that hits our eyes?
Surface reflectance function:
- describes how much light an object reflects, as a function of wavelength
- think of this as the fraction of the incoming light that is reflected back
What is the process for how color vision works?
- Illuminant: defined by power (or ‘intensity’) spectrum, which is the amount of light energy at each wavelength
- Object: defined by its reflectance function, certain percentage of light at each wavelength is reflected
- Cones: defined by absorption spectrum, each cone class adds up light energy according to its absorption spectrum
- Cone responses: three spectral measurements, convey all color information to brain via opponent channels
What is color constancy?
The visual system uses a variety of tricks to make sure things look the same color, regardless of the illuminant (light source)
It is the tendency of a surface to appear the same color under a wide range of illuminants
To achieve color constancy, we must discount the illuminant and determine the surface color, regardless of how it appears
How does the cylinder and grey patches illusion work?
The effects of lighting/shadow can make the colors look different that are actually the same
- exact same light coming to your eye from these 2 patches
- but the brain infers that less light is hitting this patch, due to the shadow
- so CONCLUSION: the lower patch must be reflecting a higher fraction of the incoming light (i.e. it’s brighter)
Rationale for Color Constancy?
Visual system tries to discount the effects of the illuminant: it cares about the properties of the surface, not the illuminant
- still unknown how the brain does this: believed to be in cortex (V1 and beyond)
- BUT: color-constancy is not perfect
- it is possible to fool the visual system
— using a light source with unusual spectrum (most light sources are broad-band; narrow-band lights will make things look very unusual)
— showing an image with little spectral variation (e.g. a blank red wall)
What does the white/gold & blue/black dress show?
Percepts depend on inferences about the light source
- but we have no idea (so far) why people are making such radically different inferences about light
Mixing of lights vs. mixing of paints
Mixing of lights = additive
- if light A and light B both arrive at the eye, the effects of those two lights add together
- Georges Seurat’s painting: illustrates the effect of additive color mixture, and reflected light from nearby dots add together when blurred by the eye
Mixing of paints = subtractive
- if pigment A and B mix, some of the light shining on the surface will be subtracted by A and some by B. Only the remainder contributes to the perception of color
What is the typical population that has color blindness?
- about 8% of male population, 0.5% of female population has some form of color vision deficiency
What is the cause of color blindness?
- mostly due to missing M or L cones (sex-linked; both cones coded on the X chromosome)
What are the types of color-blindness?
- Dichromat - only 2 channels of color available (contrast with ‘trichromat’ = 3 color channels)
- Monochromat: true ‘color-blindness’, world is black-and-white
- Cone monochromat: only have one cone type (vision is truly b/w)
- rod monochromat: visual in b/w AND severely visually impaired in bright light
What are the three types of dichromats?
- Protanopia: absence of L-cones
- Deuteranopia: absence of M-cones
- Tritanopia: absence of S-cones
