colour 2 Flashcards
what is the physiology and neuroanatomy underlying colour processing of cone signals
chromatically opponent retinal ganglion cells
chromatic tuning in LGN, V1 and V2
V4 (and cerebral acrhomatopsia)
what psychophysical evidence is there for opponency
coloured after effects arise from adaption at multiple stages
habituation experiments
how can colour processing explain colour appearance
explaining the unique hues
flexible relationship between lights (the physical stimuli) and appearance (our perception)
how is colour information extracted from cone readout
comparison between cone classes
the relative variation in light’s spectral energy as a function of e(𝜆) read from ratio of cone excitations
preserved even when light intensity changes
when the spectrum is altered the ratios change
as long as these are not metamers, the colour will change
how is chromatic opponency implemented in ganglion cells
gaglion cells make the comparison of ratios explicit
midget ganglion cells are L-M opponent and have on-off receptive fields that are spatially and chromatically opponent (although unclear whether the surround is M cone signal specific or randomly wired with L and M cones)
small bistratified ganglion cells are S-opponent
compared with LM signals combined
only chromatically opponent
how are ganglion cell RFs organised
ganglion cells feed three separate pathways of the LGN, they differ in size RF and cone contact
what are the properties of parasol ganglion
10% of all retinal ganglion cells, large spatially opponent, chromatically non-opponent RF, projects to magnocellular layers in LGN, L and M cones input via diffuse bipolar cells (probably not S cones), poor chromatic selectivity
what are the properties of midget ganglion cells
60-80% of all retinal ganglion cells
RFs are 2-3x smaller and spatially and chromatically opponent
project to the parvocellular layers of the LGN
L or M input to centre, antagonistic surround from complementary type or L&M mixture (probably not S cones)
good chromatic selectivity: cherry-teal
what are the properties of small bistratified ganglion cells
4% of all retinal ganglion cells
larg spatially non-opponent, chromatically opponent RF
project to the koniocellular layers of the LGN
S input, opposed by spatially overlapping L plus M inputs
good chromatic selectivity lilac-lime
what are the different retinogeniculate pathways
magnocellular
parvocellular
koniocellular
what are the properties of the magnocellular pathway
not colour selective
poor spatial resolution
fast
(movement)
what are the properties of the parvocellular pathway
colour selective
good spatial resolution
slow
(objects)
what are the properties of the koniocellular pathway
colour selective
very poor spatial resolution
very slow
how is colour represented in cone-opponent DKL space
L+M+S signal: white-black (intensity)
S-(L+M) signal: lilac-chartreuse (small bistratified ganglion)
L-M signal: cherry-teal (midget ganglion)
Hue corresponds to Azimuth (angle around the plane)
Saturation corresponds to elevation in the plane
how are cells in the visual pathway chromatically tuned
plotting populations of chromatically opponent cells on the axis elevation and azimuth reveals that they cluster around 0-90* (Derrington et al., 1984)
90 azimuth K-cells, S-opponent small bistratified
90 elevation M-cells, parasol
0 azimuth P-cells L/M opponent midget
repeating these measurements in different areas of the visual system shows that in V1 chromatic tuning is more evenly distributed around the hue circle
how is V4 chromatically tuned
neurons in macaque V4 respond strongly to colour, but also orientation and form - functional role in colour vision not clear (Desimone & Schein, 1987)
cerebral achromatopsia is the result of damage to V4 - strong link to colour (Zeki, 1990)
but may be that form selectivity is predominant function (Heywood et al., 1992)
how are colour, form and texture separated
either one brain region for analysing all surface properties
or multiple foci independently extracting different surface properties
what did cavina-pratesi et al.’s 2010 fMRI study show about how shape, colour and texture are processed in the brain
participants monitored a sequence of stimuli for repeats with either a shape, colour or texture change to spot
recordings showed that regions were selectively activated highlighting a separation between areas responsible for the processing of each feature
subsequent study with neuropsychological patients indicates a double dissociation
patient MS with achromatopsia has disruption to the surface-feature area loss of texture and colour, no loss of shape
patient DF shape agnosia disrupted LOC, spared surface0-feature areas, loss of shape discrimination no loss of texture or colour
can we separate gloss from colour and texture
patient MS with acrhomatopsia can complete an oddity task for glossiness for real and rendered surfaced but they are impaired
glossiness does not depend exclusively on processing in the same regions that are necessary for the perception of surface colour or texture
what is an after effect
behavioural evidence of physiology
The visual system adapts to the colours of an image after prolonged exposure and this adjusts your adaption to the subsequent image.
E.g looking at a sand colour will fatigue M cones so the afterimage will be blue and dominated by S cone signals as we adapt S opponent ganglion cells.
Aftereffects evidence extensive complementary colours.
why has it been said that reducing perception to the Young-Helmholtz theory is an oversimplification
Difficult to find evidence for Hering’s opponent processing theory beyond phenomenological evidence
ot logically the case that adaption has to occur to produce after effects.
not all cones adapt together, suggesting it is unlikely there are only red-green, blue-yellow and white- black channel
what evidence does Krauskopf et al., 1982 provide for the selective adaption of opponent colour mechanisms
Presented participants with a field that gradually changes in colour following the shape of a sine curve.
Thresholds for detecting changes in colour (e.g reddish and greenish changes from white) are raised following viewing a field sinusoidally modulated in colour over time (in a reddish-greenish direction but not after viewing one varying in a yellowish-bluish direction and vice versa).
Thresholds for changes in luminance are raised following viewing a field varying in luminance but not altered by exposure to purely chromatic variation and vice versa.
As this selectivity is only found for these directions and not intermediate directions in colour space, it follows that they are cardinal.
Signals varying along these directions are carried along separate, fatigable, second stage pathways.
The two chromatic cardinal directions correspond to the two chromatic axes of the DKL opponent colour space, and thus the two populations of cells in the LGN.
This is behavioural evidence for opponent processes.
what are the unique hues
unique blue (neither red nor green), unique red (neither yellow nor blue), unique yellow (neither green nor red), unique green (neither blue nor yellow).
are the cardinal axes compatible with Hering’s theory
there are opponent mechanisms that can adapt but the cardinal axes don’t seem to correspond to the unique hues
more likely that the properties of the external world have been internalised in our physiology
how does the physical light eliciting a given percept change
changes with the chromatic filter in the eye
changes with the chromatic statistics of the environment
what evidence is there for recalibration over time
All perceived colour is filtered through the lens.
The chromatic filter of the eye changes with age such that the lens yellows and darkens.
An increasing amount of blue is missing, beginning to absorb short wavelength light and only transmit longer wavelength light
Observers were asked to make achromatic settings where they had to adjust the colour of a light until it was perfectly achromatic (white).
Conditions: before/after cataract surgery
Found that before surgery their perception of white was the same as standard white but after surgery their perception jumps to yellow.
This is because there is long scale adjustment to the light that’s hitting the retina and the colour that the light is perceived
what calibration happens across the visual field
The macular pigment in the fovea appears yellow as it absorbs S wavelengths and transmits L wavelengths.
The colours that are most affected by this reflect a lot of s wavelength light (purple)
Percept changes depending on their position in the periphery
Participants asked to view an object in the periphery and adjust a colour patch in the fovea until they match.
There appears to be calibration across the visual field to compensate for the macula pigment as experience has taught us to predict what the relationship is, but this is not perfect
how does colour appearance depend on context
Take two identical patches of colour
Place them in different surrounds and they appear slightly different
Simultaneous colour contrast effect
Make the two surrounds more similar and the test rings now appear very different
The strength of colour induction depends heavily on spatial structure
Strong interactions between the colour context and the spatial layout of that variation
summary
To disentangle changes in wavelength from changes in intensity, the visual system must compare the outputs of the (univariant) cone mechanisms. The post-receptoral S-opponent and L/M Opponent pathways make these comparisons explicit.
Colour-selective neurons in retina and LGN respond maximally to colour variations along either the S-(L+M) or (L-M) directions in colour space. Colour-selective neurons In V1 and V2 show colour preferences either to these or to intermediate colour directions. Adaptation can occur at multiple stages of the visual pathway.
Adaptation of opponent mechanisms requires special stimuli. These mechanisms don’t correspond perfectly to the “red vs green” and “yellow vs blue” mechanisms of “opponent colours theory”.
Colour appearance can depend strongly on previous adaptation, and on concurrent spatial context. These effects may be a consequence of colour constancy mechanisms…
