Wk 6: Colour perception, colour mixing and types of colour vision

6a: Colour perception

  • how colour info is encoded

    • colour mixing

    • metamers

  • models of colour processing

  • retinal based colour ‘blindness’

    • tests of colour abnormalities

  • colour blind visual processing

  • ways to define diff colours

colour - perceived colour

  • defined by wavelength \lambda lambda

  • EMR: electromagnetic radiation 

    • energy that travels in waves and includes a broad spectrum

    • visible 400-700 nm (10^-9m)

wavelength order: ROY G BIV

  • red, orange, yellow, green, blue, indigo, violet

definitions

hue:

  • the actual colour

  • e.g. red, green

saturation: 

  • amount of colour relative to white

  • great saturation = great vividness

  • less sat = washed out colour

brightness

  • light intensity

narrow or broadband

cells that mediate (bring about) colour perception can either be broadband or narrow band

  • narrowband: 

    • sensitive to a narrow stimulus dimension

    • cell responds strongly to a narrow range of wavelengths

    • typically tuned to a specific colour

  • broadband:

    • sensitive to a broad range of stimulus dimension

    • cells respond to a wide range of wavelengths

    • often spanning multiple colour channels

drawback of a narrowband system

  • need lots of cells to encode the entire stimulus dimension

  • low sampling density

    • with more cell types spread out across the retina, each type becomes less densely packed

    • hence low spatial resolution

    • fine details not detected well

  • poor spatial acuity

    • because each cone type is spaced further apart, the system struggles to resolve small or closely spaced visual features

    • this is especially problemati in tasks requiring sharp vision

    • e.g. reading or detecting edges

tuning of individual cones

  • individual cones are tuned/selective to colour

    • specific cones need to feed into specific colour-sensitive retinal-ganglion cells to give them their specific colour sensitivity

    • these specific RG cells feed into specific colour sensitive v1 cells

  • individual cones are not tuned to orientation or motion

    • v1 simple cells are tuned to orientation so that means a given cone can feed into many simple cells that are tuned to different orienations

    • same for complex cells that are tuned to motion

    • reasons why the brain can have narrow-band tuning for orientation and motion and not impair spatial acuity for those stimulus dimensions

### 🟥 1. What cones are tuned for: Colour

- Cones = colour detectors in your retina (red, green, blue types).

- Each cone connects to a specific retinal ganglion cell that’s sensitive to colour.

- These ganglion cells then pass colour information to V1 cells (in the brain’s visual cortex) that are also tuned to colour.

- 🔑 Key idea: Colour sensitivity is preserved all the way from cone → ganglion cell → V1 cell.

---

### 🟦 2. What cones are not tuned for: Orientation or Motion

- Cones themselves don’t care about which direction a line is tilted (orientation) or whether something is moving.

- But V1 cells do care about those things:

- Simple cells in V1 are tuned to orientation.

- Complex cells in V1 are tuned to motion.

- A single cone can send signals to many different V1 cells, each tuned to different orientations or motion directions.

- 🔑 Key idea: This setup allows the brain to be very precise about orientation and motion without losing detail (spatial acuity).

---

### 🧠 Summary

- Colour tuning starts early (at the cone level) and stays specific.

- Orientation and motion tuning happen later (in the brain), using flexible input from cones.

- This division helps the brain process different visual features efficiently without overlap or loss of detail.

---

Let me know if you’d like a diagram or analogy to go with this—like a “cone-to-brain” signal pathway or a metaphor for how tuning works!

broadband and narrowband: effect on mixing

  • mix stimuli that are processed by a broadband system

    • e.g. add red and green together, you’ll perceive a new colour - yellow

    • broadband systems blend input into one unified perception

  • mix two stimuli that are processed by a narrowband system

    • e.g. combine vertical and horizontal linesm you’ll perceive both line separately 

    • these systems don’t blend

    • they keep distinct features apart

pooling of broadband and narrow cells

  • how the outputs of those cells are pooled

  • applies to cells that are looking at the same region of visual space

  • broadband tuning

    • e.g. colour

    • total population average

    • relative activation of the three cone types

  • narrowband tuning

    • e.g. orientation and motion

    • local populaiton average of cells tuned to similar orientations

    • but not cells tuned to markedly different orientations/direction

    • only pool output from cells that are processing that stimulus and so depends upon the tuning bandwidth of the cells

population response

definition: how a group of neurons responds to a stimulus. The shape of the response (unimodal vs bimodal) depends on how neurons are tuned (broadly or narrowly)

  • broadband tuning

    • neurons respond to a wide range of stimulus types

    • e.g. colours

    • when you mix colours, the brain looks at the relative activation of these cones

    • create a unimodal response

    • one clear peak of activity

    • you perceive a single colour as a result 

  • narrowband tuning

    • responds to very specific stimuli

    • e.g. line andle or motion direction

    • bimodal response

    • two separate peaks of activity

    • result: perceive both stimuli distinctly

when is a stystem narrow or broadband?

spectrum: energy at each wavelength for a given light

monochromatic light: light of a single wavelength, spectrum is a spike at that wavelength

composition of white light

  • can create white light with 2 complimentary colours

  • 3 correctly chosen colours can generate all colours

  • major colour in the spectrum

    • red, orange, yellow, green, blue, indigo, violet

    • ordered in terms of decreasing wavelength. 

6b: Colour Mixing

what do we perceive when we add diff colour together 

how does the brain process colour iinfo

metamers: physically diff stimuli but perceived the same

  • colour mixing studies tell us: people determined how the colour system worked

  • behavioural studies tell us: a function map of the system

metamers

e.g. yellow

  • spectral yellow

    • a single wavelength of light (a)

  • mixed yellow

    • combine multiple long-wavelength light (red+green) (b)

why this happens

  • our visual system uses three types of cones (red, green, blue)

  • diff light combinations activate these cones in the same way

  • hence we perceive the same colour

types of colour mixing

mixing type

additive

subtractive

medium

light sources

pigments/filters

primary colours

red green blue

cyan magenta yellow

result of full mix

white

black/dark

perception mechanism

combined cone activation

reflected wavelength after absorption

additive mixing

primary colours: RGB

  • combining light sources of different colours

  • resultant spectrum is the sum of the component spectrums

  • more light in the mixture than in the components

  • brain interprets the combined cone activation as a new colour

methods

  • projection

    • project different colour onto the same region

  • spatial summation

    • project to points in close proximity

    • e.g. red, green, blue points of light that are so close together the brain can’t spatially resolve them, so the brain averages them 

    • like a TV/phone scren

  • temporal summation

    • temporally interleave different colours faster than the brain can resolve them

    • faster than our temporal acuity limit

    • the brain adds up signals that arrive close together in time

    • if brief flashes of light occur rapidly enough, they’re perceived as continuous or brighter

    • if diff colours are flashed quickly in sequence, the brain can average them over time

subtractive mixing

primary colours: CMY cyan magenta yellow

  • light is removed, so there is less light in the mixture than in the components

  • only the light that is common to both components remain

  • combining materials that absorb certain wavelengths

  • perceives wavelengths that are left over after absorption

methods

  • add colours to remove light

  • less light in the mixture than in either component

  • e.g. adding paints together, stacking lenses together

mixing the spectral colours

what if we mix all ROYGBIV colours

colour matching

split-field display and additive mixing

process of recreating a target colour by combining other colours

  • minimum no. of primary colours

    • 3 to maky any colour

    • 2 to make white

implications for colour mechanisms

  • for people with standard colour vision, any colour can be matched by three appropriately chosen primaries

  • related to having three cones types

    • long, medium, and short LMS lambda sensitive

    • also called RGB

6c: Types of Colour Vision

consider the type of vision a person would have if they had one, two or three diff cone types

monochromatic vision

  • only one type of cone

  • this system cannot discriminate colours, only light intensity

  • the response of the receptor to diff wavelengths can be matched by varying the relative intensities of the lights

    • e.g. it might abosorb more light at 500nm than 550nm

  • a dim light at one wavelength can produce the same response as a brighter light at another wavelength

    • e.g. 500nm has same response as 550nm with twice the intensity

  • can only encode intensity: black and white vision

    • shades of grey

    • similar to how the rod system works

  • monochromatic light at neutral point will drive them both equally

    • ratio of response 1:1

  • monochromatic light at neutral point is perceived as white

    • white light also drives both types of detectors equally and so is signalled b relative activity of 1:1

dichromatic vision

neutral point: red arrow
  • two types of cones

  • can encode colour

  • limited range of distinct colours that can be perceived

  • only two distinct colours

  • need only two primaries to match any colour

  • presence of a neutral point

    • wavelength at which the two curves cross over

    • excites both receptors equally

trichromatic

  • colour perception based the relative activity of three receptor types

  • need three primaries to match all colours

  • no neutral point

  • no wavelength excites all receptors equally

tetrachromatic

  • four types of cones

  • affects on human colour perception

    • subdivide the spectrum into more bands (10 vs 7)

trichromatic theory

  • three types of ‘fibres’ in the eye

  • perceived colour is based on the ratio of LMS activity

  • physical ratio of LMS cones are about 32:16:1

results that can’t be accounted for by trichromatic theory

  • colour afterimages

    • e.g. view red, green afterimage

  • colour naming

    • asked to assign a name to monochromatic lights

    • can’t describe all colours using RGB

    • need to add Y

    • can’t get certain combinations

      • e.g.can’t get RG but can get RB

opponency model

colour vision is mediated by three opponent mechanisms

opponent channels: 

  • red, green

  • blue, yellow

  • white, black

trichromacy vs opponency

pairing

opponent

cone

red, green

LM

BY

S, ML

white, black

LM, LM

our colour perception is based up

  • having three cone types and their relative activation

  • opponent mechanisms is how the brain encodes their relative activity

  • note RGB the same as LMS

  • LMS often used to make it clear that they are not only sensitive to RGB

colour bliind processing

  • Magnocellular and parvocellular distinction

    • M: only luminance sensitive (colour blind), no colour info

    • P: colour and luminance sensitive