Unformatted text preview: Color Vision Based on 3 classes of cones, each with 2010 conesFri., Oct.29, a different visual pigment, which absorbs a different Color Vision visible spectrum band of Colour Vision short wavelength (S) cones: absorbance peak at about 420 nm; uncommon - Based on 3 classes of cones Long wavelength - Each w/ a diff visual pigment (opsin), which absorbs a diff band of middle wavelength (M) cones: 530 nm visible spectrum long wavelength (L) cones: 552-557 nm
PSL302Y: Lecture 20, by Prof. MacKay! October 29, - Cones classified based on absorbance peaks (opsins vary btwn individuals, so peaks also vary) - Short wavelength (S) cones: ~420nm (uncommon) - Middle wavelength (M) cones: 530nm - Long wavelength (L) cones: 552-557 nm - Graph: Lots of overlap btwn M & L absorbed wavelengths - Absorbance peak of rhodopsin btwn S & M - Melanopsin to be discussed later - Each cones absorb broad band of visible light spectrum - L cones absorb red, yellow, and even greens; peak at orange - M cones absorb some reds, up to blue; peak at green - S cones absorb mostly violets but also greens - Distribution of photoreceptors on retina - L cones in red, M cones in green, S in blue; rods in grey - Few rods, lots of cones - S are rarest (~10%), relatively evenly distributed - L & M tend to be clumped Middle wavelength Short wavelength Visual pigment absorbance Color Vision October 29, 2010 Long wavelength retinal distribution of cones
Middle wavelength Short wavelength Rod Absorption - Rods fill the gap btwn S & M cones - BUT rods roughly 10x +sensitive to light (except red) than cones - At high light lvls required for cone fxn, rhodopsin bleached = rods non-fxnal - To derive more selective response to light: must do some processing - First stage of processing right in retina at colour-selective ganglion cells retinal distribution of cones 2 Colour selective ganglion cells: 2 types - Type 1: various sets are colour-selective b/c use center-surround organization of r.f.'s to subtract btwn responses of M & L cones - IOW: Centre + surround r.f.'s receive input from diff cone pop'ns, usually M + L types - So center part of r.f. only gets M cones, and surround gets L cones, or the rvs - Lateral inhibition subtracts response to stimulus, of one cone type from other - IOW: lateral inhibition btwn two parts of r.f.'s subtracts btwn responsiveness of M + L cones to particular stimulus - Ganglion cell responds to the differential effect btwn competing cone pop'n inputs - Restricted # sensory receptors: so look at ratio of activity btwn receptors 1 of 6 October 29, 2010 PSL302Y: Lecture 20, by Prof. MacKay! Fri., Oct.29, 2010 Diagram: L cones feeding center (ON ganglions), M cones feeding surround (OFF ganglions) -M absorbance peak: opposite b/c inhibitory effect on ganglion -L has excitatory effect on ganglion cell -Diffce traced out in middle fxn -> grey area = +ve diffce -Any wavelengths in this range -> more excitation of ganglion cell than inhibition -Subtraction is what colour-selective ganglions are doing! -If we go to range where L cones most effective (orange), we have = effect from M cones => zero effect on ganglion -If we go to shorter wavelengths (green) = net inhibitory effect on ganglion cell Single-opponent r.f.'s - Used c-s structure to look at luminance contrast (B&W) - Same structure in retina for narrowing down range in spectrum that ganglion cells respond to - Call this single-opponent r.f.: 1 cone pop'n in one part of field + other cone pop'n in antagonistic part of field Single-opponent r.f.'s - Combos of M + L inputs, ON/OFF, in centre or surround make up most colour cell r.f.'s - NOTE: optimal coloured stimuli give best response when they cover entire r.f., b/c Combinations of M and L inputs, ON or inputs based on diffial responses by diff cone pop'ns OFF, in center-or surround make up most Acuity in colour vision is less than B&W vision color cell r.f.'s - Ganglion cells do NOT respondColor Vision contrast to colour - colored stimuli a more NOTE: optimal They respond togive bestrestricted band of spectrum than do cones The narrow range response when- they cover entire r.f.of spectrum that cones are responding to, b/c cones have hopelessly wide response spectrum ganglionDiagram:NOT respond to color inhibitory OFF cells do excitatory ON center, contrast;surround they respond to a more restricted band of spectrum than L cones, surround by M cones - Center fed by do cones - "GREEN" stimulus: inhibit ganglion cell - "RED" stimulus: excite ganglion cell - "WHITE" stimulus: no activity - Must have L wavelength stimulus picked up more strongly than picked by M cone 4 = more excitation - Other possibilities for combining L & M cone pop'ns - "ON" centers as either L & M wavelength, or "OFF" centers as either too, and opposite in center-surround structure October 29, 2010 2 of 6 PSL302Y: Lecture 20, by Prof. MacKay!
Color Vision Fri., Oct.29, 2010 Diagram: L+ -> L wavelength is excitatory (left) vs. M+ -> M is excitatory (right) -Left: L+ in center vs. L + in surround = both r.f.s respond to "RED" light -Right: M+ in center vs. M+ in surround = both r.f.'s respond to "GREEN" light Coextensive opponent r.f.'s -BUT diff proportions of cones being activated -L+ in surround territory >>> L+ in center territory -> pushes Since red spectrum (closer to number activity of ganglion cell further along S cones are few in number, they are not relegated to a center or surround field orange) -Clumping of cones on retina S cone inputs are collected from entire field -All the cones feeding into center/surround need not be just one S cones provide ON input, combined M + L cone type -> mostly just 1 cone type due to clumping, but cones provide O input p OFF p others are possible S cones also mixed in other to - Depending on the proportion of cones types on retina, shift actual activity of ganglion cellcombinations diff range of spectrum 5 - M & L wavelength cones: "bread & butter" of visual system - S wavelength cones: tiny fraction of visual system - Not enough of them to fill a center/surround set (dispersed, not clumped) - Must take part in diff type of colour-selective ganglion cell set... Coextensive opponent r.f.'s - Since S cones are few in # (and dispersed, not clumped), they are not relegated to a centre or surround field - S cone inputs are collected from entire field - S cones provide ON input, combined M + L cones provide OFF input - Don't go thru center-surround ganglion cell sets - S cone inputs go thru 1 ON bipolar -> transmit excitatory input to colour-selective ganglion cell (blue-yellow combination cell, rather than red-green combination) - M & L cone inputs summed onto another set of OFF bipolars -> provide inhibitory input to b-y ganglion cell - Ganglion responds to S wavelengths + is inhibited by M & L wavelengths - Combination of M & L wavelengths -> peak in Color Vision yellow range = max inhibition at yellow wavelength - S cones also mixed in other combos - Most common of coextensive r.f.s: - S+ vs. M- & L- = "BLUE" - Rare but still existing: - M+ & L- vs. S+ = "CYAN" - L+ & M- vs. S+ = "MAGENTA" - M or L input to ON bipolar along with S input - These colours not part of light spectrum -> do not correspond to wavelengths of light = combos created in retina! October 29, 2 3 of 6 Color channels PSL302Y: Lecture 20, by Prof. MacKay! Fri., Oct.29, 2010 @ ganglion: Colour Channels - As a result of retinal processing, several colour-related channels are sent to brain - Huge # of colour-selective cones - Red-green: LARGE family w/ varying colours of peak response - Blue-yellow: SMALL group of ganglion cells responding to blue, cyan & magenta - Black-white: ALL cones combined = finest acuity b/c across entire r.f., not just in c-s - Recall lecture 19 @ thalamus: Colour cells in LGN - Processing in thalamus' LGN: same spectral responses as ganglion cells - Located in P layers (parvocellular) - Colour-selective cells kept separate from B&W cells - Fxn of LGN is alignment of r.f.s across all layers = sent to appropriate V1 hypercolumn - Project to layer 4C of V1 - Some colour cells in interlaminar spaces of LGN = K stream for koniocellular ("tiny cells") - Tend to receive synaptic input from b-y stream: afferents project btwn LGN layers - Project directly to C.O. blobs of V1 -> where most of colour analysis proceeds Color Vision @ visual cortex: V1 colour cells - Colour-sensitive stellate cells of layer 4C similar to LGN and ganglions - R-G info processed here, B-Y info processed at C.O. blobs - R.f.'s similar to r.f.'s at ganglions + LGN, but slightly larger due to convergence - Not oriented bars of light but points of appropriate colours - Project to colour-selective oriented cells (close to cytochrome oxidase blobs) and to blobs: respond to bars of particular colours - Found alongside B&W oriented cells in orientation columns, close to C.O. Blobs - Double-opponent colour cells are common in C.O. Blobs - Blob cells do not respond to oriented bars Color Vision - W/i C.O. blocks, never find oriented cells: respond to spots of light October 29, 2010 Double opponent cells Center surround r f Center-surround r.f. for optimal response, each part of r.f. must have different color stimulus responds to color contrast: e.g. green center and red surround October 29, 2010 Double-opponent cells - Two parts to r.f.: centre-surround r.f. - But center made up of selective responses already established in ganglion cell pop'n - Surround sets made up of OPPOSITE responses - Larger r.f. -> combining r.f.s of retina & LGN - For optimal response, each part of r.f. must have different colour stimulus - Sum double-opponent cells along linear axis - Responds to colour contrast: e.g. green centre and red surround, and vice versa - This is what the B&W ganglions do in retina V2 color cells - Processing at retina: narrows range of wavelengths (colour-selectiveness) of-> then continues in Processing color information the association areas V2 and V4 contrast processed up in brain V2 color cells are concentrated within the `thin stripe' zones of cytochrome oxidase - Bigger r.f. = lower acuity in colour vision both oriented (including directional, end( g , stopped, length-summing) cells and non- Double-opponent cells are characteristic of C.O. blobs oriented - Colour-selective oriented cells outside of C.O. blobs in orientation columns 9 10 4 of 6 PSL302Y: Lecture 20, by Prof. MacKay! Color Vision Fri., Oct.29, 2010
thin stripe thick stripe October 29, 2010 @ association areas: V2 colour cells - Processing of colour info continues in the association areas V2 and V4 - V2 colour cells are concentrated w/i the thin stripe zones of c.o. - Both oriented (incl. directional, end-stopped, lengthsumming) + non-oriented cells - Selectivity for colour more acute: narrow down response rates to particular shades - V1 is speckled w/ C.O. Blobs; in V2, C.O. sections in thick stripes btwn Color Vision thinner stripes - Thin stripes receive input from C.O. blobs in V1 - In thin stripes, find cells that are more selective for colour, i.e. pink, aqua, lime; Color Area (V8) also purple - Existence of responses to colours that culmination of color analysis: lesion results don't exist in visible light spectrum in achromatopsia (extraspectral, i.e. purple) neurons respond to perceived colors, - Entire colour system does not regardless of wavelength composition rely on detecting specific wavelengths processing checks for g p g global shifts in - To see yellow, don't pick up one particular wavelength -> in diff visual envmts,patch illumination; interpretation of one diff wavelengths reflected back but you can consistentlydepends on light from surrounding region recognize same colour - We recognize patterns: ratio of wavelengths from object VS everything else - Processing requires a lot of stages: looking at global effects across retina to determine what lighting means in visual scene - Global shift to adapt interpretation of wavelengths to actual lighting conditions Colour Area (V8) - At bottom of inferior temporal lobe, beside fusiform area: culmination of colour analysis - Lesion results in achromatopsia (everything looks grey, no colour) - Neurons respond to perceived colours, regardless of wavelength composition - Learns to recognize RED ratio out of all possible ratios of activity - Processing checks for global shifts in illumination - Interpretation of one patch depends on light from surrounding region October 29, 2010 11 1 12 5 of 6 Color Vision October 29, 2010 PSL302Y: Lecture 20, by Prof. MacKay! Fri., Oct.29, 2010 Responses of one neuron in `V8': -Expt: Mse response to RED -Wavelength composition from stimulus was exactly the same for each of these conditions, but it looked different b/c of diff visual envmts (what was around this stimulus, i.e. lighting conditions, surrounding lights) -When it LOOKS like red, very active -When it looks green or blue, virtually no response, etc. Color Vision Perceived color depends on more than wavelength; the light from the surrounding field is just as important. October 29, 2010 Responses of one neuron in `V8': Perceived color depends on more than wavelength; the light from the surrounding field is just as important. - So response to same wavelength depends on diff visual envmts - Colour analysis requires global analysis of illumination - Brain picks up SHADED side of cube: performs global shift to reinterpret wavelengths coming from this wall Color Vision October 29, 2010 - Assumes ORANGE shade is orange b/c must be lighter if properly illuminated BUT the actual wavelength leaving the picture is BROWN Orange? Orange? No----brown! 13
13 6 of 6
View Full Document
This note was uploaded on 03/27/2012 for the course PSL PSL300 taught by Professor Mackayfrench during the Fall '11 term at University of Toronto.
- Fall '11