Category Archives: vision

The evolutionary trajectory of color vision

In a recent comment, a reader of this blog had highlighted some concerns regarding an evolutionary trajectory of color vision evolution that I had proposed in conjunction with the evolution of color terms in human languages and as a possible explanation for the linguistic trend. At that time, I had proposed a possible evolutionary scenario, without doing due diligence investigation of existing evolutionary theories of color vision, as my post had more of a linguistic and developmental focus and the evolutionary conjecture was just that- a conjecture, which, if found true, would lend more credence to my linguistic trend. Thanks to Andreas, I reviewed the literature on color vision evolution and was surprised to find some support for my theorization.

Before I discuss the color vision evolution, I’ll strongly recommended reading two posts on the evolution of color vision and the evolution of retinal structures (more Avian focus here) , for getting some basic familiarities with the retinal structures involved in color vision and how they might have evolved.

To recap,

An animal has color vision if it has the capability of discriminating lights (scattered light as well as light sources) on the basis of the lights’ spectral content, even when those lights are of equal subjective brightness.

The front end requirement for such a system is that the animal must have at least two different spectral classes of receptor, where each class is defined by the sensitivity of the receptor to light as a function of wavelength.

The above succinctly defines what we usually mean by color vision. You can either have a dichromatic color vision, when you have two differently tuned receptors to detect different light wavelengths and the different signal combinations from these receptors yield different hues; or you can have trichroimatic / tetrachromatic vision where three/four independent color signals are combined to yield an entire Hue range. One familiar with the RGB color system used in computers, would note that it is based on the assumption of 3 pure colors, which can be mixed in different amounts to yield most of the color hues we see on the monitor.Pigeons, and birds in general, have a tetrachromatic color vision.

Now for some basic visual circuitry:

The retinal structures involved in vision, in mammals, are, pohotorecptors (classified as cones and rods), horizontal, bipolar, amaracine and ganglion cells.

However, for all vertebrates (mammals as well as reptiles and birds) and invertebrates as well, the receptor mechanism is conserved and is basically the same and we will discuss that first:

The first step in the transduction of light energy to a neural signal is the light-induced isomerization (change of shape) of a chromophore, specifically a vitamin A derivative. Each chromophore is bound to a membrane protein called an opsin. The main function of the opsin is to change shape after light absorption triggers the isomerization of the chromophore: the opsin is an enzyme that is activated by the chromophore’s isomerization. However, because of the linkage between the opsin and the chromophore, the opsin also serves to tune the wavelength dependence of the light induced isomerization reaction in the chromophore. That is, the chromophore’s sensitivity to light at a given wavelength is established in part by the opsin–different opsins (i.e. opsins with different amino acid sequences) bound to identical chromophores will have different absorption probabilities at each wavelength. The result is that photoreceptors which express the gene for only one type of opsin will form a different class than photoreceptors that express a gene coding for a different opsin. Although there are other mechanisms that animals could use to differentiate photoreceptor classes (most notably some animals use more than one chromophore, and many vertebrates have colored oil droplets that screen individual receptors) it seems that the expression of only one of their possible opsin coding genes in each receptor is the mechanism that all animals use.

The above clarifies, that in mammals, we associate color vision with cones or specialized photoreceptors that contain a single pigment and are responsive to a single wavelength range. In reptiles, we also have double cones, wherein, two photopigment/ receptors are part of the same cell and then there are other mechanism like oil droplets that are also involved in color vision (but thankfully not in mammals). Rods are also a type of receptors, tuned to a frequency, but we normally do not associate rods with color vision, because they are usually used for night vision and their signals are not combined to create the color hue; yet a limited form of monochromatic color vision is possible by having a combination of one rod and one cone receptor types.

Next we need to differentiate between the rhabodermic eyes of invertebrates (based on r-opsin and the ciliary eyes of vertebrates based on c-opsins. Pharyngula does an excellent job here.

Eyes can be further categorized as rhabdomeric or ciliary by the nature of the cellular elements that make up the photoreceptors, by the kind of opsin molecule used to transduce the light signal, and by the signaling pathway used to convert a conformation change of the opsin molecule into a change in the electrical potential across the cell membrane.

As many accounts of color vision evolution focus on the phylogentic tress of opsin genes evolution to make their case, it is important to distinguish between the levels of analysis. All the known Opsin genes can be classifies in seven sub-families: two of these the r-opsin families and the c-opsin families are pertinent to, and expressed in, the photorecptors found in invertebrates and vertebrates respectively.

Thus, if one wants to focus on mammal color vision evolution, one needs to focus on c-opsins mostly. Many studies have been conducted over these and the phylogentic data indicates that the vertebrate opsins too form a neat tree with five sub-families relevant for (color) vision and 3 other sub-families having non-visual functions.

Thus, in mammals we have a five types of opsins : one rhodopsin-type and expressed in rods, and four other chromatic types (detecting Red, Blue, Green and U/V colors) and expressed in cones.

One should pause here and note that the human S(short) or blue receptor actually belongs to the U/V (S) family; while the human L (red) and M (green) receptors both belong to the Red (L) family.

These 5 opsin families (Red, Gree, Blue, U/V and Rhodipsin)have been variously characterized as (L, ML, MS, S and Rh) or as(RH1, RH2, LWS, SWS1 and SWS2).

With this background, information, we can now go straight to the heart of the problem: the evolutionary trajectory of these different receptors / opsins and how the color vision evolved in humans. I’ll limit the discussion here to mammals first and then to primates , as my original thesis that color terms evolution follows the color vision evolution requires the analysis to happen only in that time frame in which linguistic abilities make sense. Assuming some proto-language in Apes and primates, it is reasonable to expect that whatever sequence of color terms we see in languages, would reflect the successive levels of color vision as experienced by Primates, and would be independent of how color was perceived in invertebrates. (I’m sure no one contends that the color terms of human languages should capture the early chromatic experiences of invertebrates).

Although I do not buy the bottleneck theory of Mammal evolution -stock, barrel and lock – I believe we can take that as a reasonable starting point. It posits that mammals were reduced to being a nocturnal burrowing species during the age pf the dinosaurs and thus were reduced to having just the rods, and lost the earlier, cones, double cones and oil pigments that reptiles still have. In any case, in mammals, rods seem to be older and more conserved than cones. (Pat on the back: one claim originally made defended to satisfaction!!)

Amongst vertebrates, the rod opsin seems to be the most conserved; cone opsins have arisen principally by duplication and subsequent mutation of the rod opsin gene.


Which of the two primary classes, rods or cones, is the ancestral photoreceptor? Given the tremendous variation seen photoreceptors across vertebrate and invertebrate species, this in not an easy question to answer based on simple phylogenetic assumptions. In addition, it is often difficult to clearly distinguish certain rod and cone types from each other, or classify them into one or the other category. Rods appear to be relatively more conserved in vertebrates in terms of pigments and structure than cones, and therefore could be considered the more ancestral form. However, rods in some respects are more morphologically complex than cones, having developed extreme sensitivity (capable of detecting as little as one photon of light).

Now,coming to the evolution(or re-evolution) of the cones or the chromatic system in mammals, it is instructive to pause here and note that having three cones does not necessarily mean that the two species will have the same qualia of color hues.

If we restrict ourselves to animals which have the same number of receptor classes, might we expect that their color vision systems are equivalent? The answer is a resounding no. Let’s compare the color vision systems of two animals that both have three photopic (e.g. active under bright illumination) photoreceptor classes. One is the human, the other is the honey bee (specifically the worker–I don’t know how the other castes are endowed). Does anybody here think that what a bee sees when it looks at a rainbow has the same appearance as what we see? We’ll ignore optical polarization (which the bee is sensitive to and we’re not) and focus on what we can infer about “color” based on, among other things, our knowledge of the bee’s receptor classes. To begin with, at the inside of the rainbow where the violet-appearing light fades off to invisibility for us, the bee will still see more rainbow. On the outside, where we see red, the bee would see nothing for although bees have an ability to see what for us is UV, we have the ability to see what bees might call infrared.

Also, it is instructive to note here how a higher level chromatic vision (dichromatic for instance) may arise form a lower level chromatic vision (monochromatic in this example). Although, along with the photoreceptors, we will need additional supporting neural wiring, in both the retina and the brain, for the opponent-processing mediated color perception to take place, we will restrict the discussion to the emergence of a new photoreceptor.

A new photoreceptor, may come into existence by a duplication and polymorphisms of an existing receptor (opsin) gene. The new receptor would have a slightly different frequency sensitivity than the original receptor and, by selectively expressing these two genes in different receptors, we can have two types of receptors. By processing and combining the two types of signals, one can now get dichromatic vision, from the original monochromatic vision.

Much confusion, in primate color vision evolution, depends on the fact that one takes as base the other mammals like dogs, and their blue-yellow world as a baseline from where to start. It should be emphasized that even though dogs may currently have two receptors, tuned to detect blue and yellow, we cannot conclude form that anything about humans or ancient ancestral mammals. In the human ancestry lineage, the dichromatic phase may have involved Red-Green perception. This is evident form the bee-human trichromatic example given above.

A very good paper summarizing the latest research on primate color evolution concludes that their are five types of primate color vision systems- beginning with a Monochromatic (L opsin only)_ system in nocturnal primates to a S + M+L (multiple copies) trichromatic system in humans.

It is interesting to note here that the human Green evolved, by replication and polymerization of the Red opsin present on the X chromosome. From the hierarchy of primate color systems, it is reasonable to conclude, that initially when we were nocturnal primates, we had a dysfunctional S-opsin gene and a functional L gene- conferring us the ability to perceive the red qualia to some extent.

In diurnal prosimians, the S become functional and they have two qualia- that of red and blue.

In the new world monkeys, the L gene is polymorphic (it is on X chromosome and as explained in the paper, if we have two alleles for that L gene, that encode for slightly different frequencies, then as females have two X chromosomes, they can have both the alleles; the males meanwhile have only one X chromosome; so at at a time they can have only one of the alleles present. By X chromosome inactivation process, all cells of a female new world monkey, will have only one of the alleles; but different cells may have different alleles expressed and thus, the females may have 3 types of receptors (one S type and two L types), thus endowing them with trichromatic vision. The Males meanwhile will have dichromatic vision, but as the gene is polymorphic, we will differences in their dichromatic perceptions. This is exactly what is observed.

The old world monkeys, have the full apparatus for trichromatic vision- with one S and two L genes. The second L (or rather M as it detects green) gene was formed by replication and polymorphisms of the L gene that detected red. thus, they had the qualia of Red, Blue and Green.

Lastly, the humans, are more or less the same as old world Monkeys; but their L gene shows polymorphisms. This has the effect of making some females tetrachromatic (as this polymorphisms will only affect females- only they have two copies of X chromosome) and it seems , that by fortuitous replication, we might get a fourth cone type in all humans. Till then, this polymorphisms will explain some of the color perception differences that we may exhibit.

Suffice it to say, that the evolution of color terms should follow the same trajectory- with Black and White (rod based) color terms preceding Red, Blue, Green and Yellow color terms.

A final note of caution: only receptor types do not guarantee that the qualia experienced would change. In an experiment with mice, in which the mice were endowed with human pigments, they could not still learn to distinguish Red, as presumably the latter opponent-processing wiring, required for that qualia generation was not present/ couldn’t develop.

Thats all for now. Hope you found this post Eye opening!! Do let me know via comments of any incompatible/recent evidences and arguments.

Incongruence perception and linguistic specificity: a case for a non-verbal stroop test

In a follow up to my last post on color memory and how it affects actual color perception, I would like to highlight a classical psychological study by Bruner and Postaman, that showed that even for non-natural artifacts like suits in a playing card deck, our expectation of the normal color or shape of a suit, affects our perception of a stimuli that is incongruent to our expectations.

In a nutshell, in this study incongruent stimuli like a red spade card or a black heart card was presented for brief durations and the subjects asked to identify the stimuli completely – the form or shape (heart/spade/club/diamond), the color (red/black) and the number( 1..10…face cards were not used) of the stimuli.

The trial used both congruent ( for eg a red heart, a black club) as well as incongruent stimuli (a black heart, a red spade).

To me this appears to be a form of stroop task , in which, if one assumes that form is a more salient stimulus than color, then a presentation of a spade figure would automatically activate the black color perception and the prepotent color naming response would be black, despite the fact that the spade was presented in red color. This prepotent ‘black’ verbal response would, as per standard stroop effect explanations, be inhibited for the successful ‘red’ verbal response to happen. I am making an analogy here that the form of a suit is equivalent to the linguistic color-term and that this triggers a prepotent response.

In these lights, the results of the experiment do seem to suggest a stroop effect in this playing-deck task, with subjects taking more trials to recognize incongruent stimuli as compared to congruent stimuli.

Perhaps the most central finding is that the recognition threshold for the incongruous playing cards (whose with suit and color reversed) is significantly higher than the threshold for normal cards. While normal cards on the average were recognized correctly — here defined as a correct response followed by a second correct response — at 28 milliseconds, the incongruous cards required 114 milliseconds. The difference, representing a fourfold increase in threshold, is highly significant statistically, t being 3.76 (confidence level < .01).

Further interesting is the fact that this incongruence threshold decreases if one or more incongruent trials precede the incongruent trial in question; or increases if the preceding trials are with normal cards. This is inline with current theories of stroop effect as involving both memory and attention, whereby the active maintenance of the goal (ignore form and focus on color while naming color) affects performance on all trials and also affects the errors , while the attentional mechanism to resolve incongruence affects only reaction times (and leads to RT interference).

As in the playing card study, no reaction time measures were taken, but only the threshold reached to correctly recognize the stimuli were used, so we don’t have any RT measures, but a big threshold is indicative of and roughly equal to an error on a trial. The higher thresholds on incongruent trial means that the errors on incongruent trial were more than on congruent trials. The increase in threshold , when normal card precede and a decrease when incongruent cards precede is analogous to the high-congruency and low-congruency trials described in Kane and Engel study and analyzed in my previous posts as well as in a Developing Intelligence post. It is intuitive to note that when incongruent trials precede, then the goal (ignore form and focus on color while naming color) becomes more salient; when normal cards precede one may have RT facilitation and the (implicit) goal to ignore color may become less salient.

Experience with an incongruity is effective in so far as it modifies the set of the subject to prepare him for incongruity. To take an example, the threshold recognition time for incongruous cards presented before the subject has had anything else in the tachistoscope — normal or incongruous — is 360 milliseconds. If he has had experience in the recognition of one or more normal cards before being presented an incongruous stimulus, the threshold rises slightly but insignificantly to 420 milliseconds. Prior experience with normal cards does not lead to better recognition performance with incongruous cards (see attached Table ). If, however, an observer has had to recognize one incongruous card, the threshold for the next trick card he is presented drops to 230 milliseconds. And if, finally, the incongruous card comes after experience with two or three previously exposed trick cards, threshold drops still further to 84 milliseconds.

Thus clearly the goal maintenance part of stroop effect is clearly in play in the playing-card task and affects the threshold for correct recognition.

The second part of explanation of stroop task is usually based on directed inhibition and an attentional process that inhibits the perpotent response. This effect comes into play only on incongruent trials. An alternate explanation is that their is increased competition of competing representations on incongruent trials and instead of any top-down directed inhibition, inline with the goal/expectation, their is only localized inhibition. The dissociation of a top-down goal maintenance mechanism ad another attentional selection mechanism seems to be more inline with the new model, wherein inhibition is local and not top-directed.

While RT measures are not available it is intersecting to take a look at some of the qualitative data that supports a local inhibition and attentional mechanism involved in reacting to incongruent stimuli. The authors present evidence that the normal course of responses that are generated by the subjects for (incongruent) stimuli is dominance, compromise, disruption and finally recognition.

Generally speaking, there appear to be four kinds of reaction to rapidly presented incongruities. The first of these we have called the dominance reaction. It consists, essentially, of a “perceptual denial” of the incongruous elements in the stimulus pattern. Faced with a red six of spades, for example, a subject may report with considerable assurance, “the six of spades” or the “six of hearts,” depending upon whether he is color or form bound (vide infra). In the one case the form dominates and the color is assimilated to it; in the other the stimulus color dominates and form is assimilated to it. In both instances the perceptual resultant conforms with past expectations about the “normal” nature of playing cards.

A second technique of dealing with incongruous stimuli we have called compromise. In the language of Egon Brunswik , it is the perception of a Zwischengegenstand or compromise object which composes the potential conflict between two or more perceptual intentions. Three examples of color compromise: (a) the red six of spades is reported as either the purple six of hearts or the purple six of spades; (b) the black four of hearts is reported as a “grayish” four of spades; (c) the red six of clubs is seen as “the six of clubs illuminated by red light.”

A third reaction may be called disruption. A subject fails to achieve a perceptual organization at the level of coherence normally attained by him at a given exposure level. Disruption usually follows upon a period in which the subject has failed to resolve the stimulus in terms of his available perceptual expectations. He has failed to confirm any of his repertory of expectancies. Its expression tends to be somewhat bizarre: “I don’t know what the hell it is now, not even for sure whether it’s a playing card,” said one frustrated subject after an exposure well above his normal threshold.

Finally, there is recognition of incongruity, the fourth, and viewed from the experimenter’s chair, most successful reaction. It too is marked by some interesting psychological by-products, of which more in the proper place.

This sequence points towards a local inhibition mechanism in which either one of the responses is selected and dominates the other; or both the responses mix and yield to give a mixed percept —this is why a gray banana may appear yellowish—or why a banana matched to gray background by subjects may actually be made bluish—as that of a blackish red perception of suit color; or in some cases there may be frustration when the incongruent stimuli cannot be adequately reconciled with expectations- leading to disruption- in the classical stroop task this may explain the skew in RT for some incongruent trials—-some take a lot of time as maybe one has just suffered from disruption—; and finally one may respond correctly but only after a reasonable delay. This sequence is difficult to explain in terms of top-down expectation model and directed inhibition.

Finally, although we have been discussing the playing card task in terms of stroop effect, one obvious difference is striking. In the playing cards and t e pink-banana experiments the colors and forms or objects are tightly coupled- we have normally only seen a yellow banana or a red heart suit. This is not so for the printed grapheme and linguistic color terms- we have viewed then in all colors , mostly in black/gray- but the string hue association that we still have with those colors is on a supposedly higher layer of abstraction.

Thus, when an incongruent stimuli like a red heart is presented , then any of the features of the object may take prominence and induce incongruence in the other feature. For eg, we may give more salience to form and identity it as a black spade; alternately we may identify the object using color and perceive incongruence in shape- thus we may identify it as a red spade. Interestingly, both kind of errors were observed in the Bruner study. Till date, one hast not really focussed on the reverse stroop test- whereby one asks people to name the color word and ignore the actual color- this seems to be an easy task as the linguistic grapheme are not tied to any color in particular- the only exception being black hue which might be reasonably said to be associated with all grapheme (it is the most popular ink). Consistent with this, in this reverse stroop test, sometimes subjects may respond ‘black’ when watching a ‘red’ linguistic term in black ink-color. This effect would be for ‘black’ word response and black ink-color only and for no other ink color. Also, the response time for ‘black’ response may be facilitated when the ink-color is black (and the linguistic term is also ‘black’) compared to other ink-colors and other color-terms. No one has conducted such an experiment, but one can experiment and see if there is a small stroop effect involved here in the reverse direction too.

Also, another important question of prime concern is whether the stroop interference in both cases, the normal stroop test, and the playing card test, is due to a similar underlying mechanism, whereby due to past sensory (in case of playing cards) or semantic associations (in case of linguistic color terms) the color terms or forms (bananas/ suits) get associated with a hue and seeing that stimulus feature automatically activates a sensory or semantic activation of the corresponding hue. This prepotent response then competes with the response that is triggered by the actual hue of the presented stimulus and this leads to local inhibition and selection leading to stroop interference effects.

If the results of the non-verbal stroop test, comprising of natural or man-made objects, with strong color associations associated with them, results in similar results as observed in the classical stroop test, then this may be a strong argument for domain-general associationist/ connectionist models of language semantics and imply that linguistic specificity may be over hyped and at least the semantics part of language acquisition, is mostly a domain general process. On the other hand, dissimilar results on non-verbal stroop tests form the normal stroop test, may indicate that the binding of features in objects during perception; and the binding of abstract meaning to linguistic words in a language have different underlying mechanisms and their is much room for linguistic specificity. Otherwise, it is apparent that the binding of abstract meaning to terms is different a problem from that of binding of different visual features to represent and perceive an object. One may use methods and results from one field and apply them in the other.

To me this seems extremely interesting and promising. The evidence that stroop test is due to two processes – one attentional and the other goal maintenance/ memory mediated – and its replication in a non-verbal stroop tests, would essentially help us a lot by focusing research on common cognitive mechanisms underlying working memory – one dependent on memory of past associations and their active maintenance- whether verbal/abstract or visual/sensory- and the other dependent on a real-time resolution of incongruity/ambiguity by focusing attention on one response to the exclusion of the other. This may well correspond to the Gc and Gf measures of intelligence. One reflecting how good we are at handling and using existing knowledge; the other how good we are able to take into account new information and respond to novel situations. One may even extend this to the two dissociated memory mechanisms that have been observed in parahippocampal regions- one used when encountering familiar situations/stimuli and the other when encountering novel stimuli. One essentially a process of assimilation as per existing schema/ conceptual metaphors; the other a process of accommodation, involving perhaps, an appreciation/formation of novel metaphors and constructs.

Enough theorizing and speculations for now. Maybe I should act on this and make an online non-verbal stroop test instead to test my theories!

Endgame: Another interesting twist to the playing cards experiment could be in terms of motivated perception. Mixing Memory discusses another classical study by Bruner in this regard. Suppose that we manipulate motivations of people so that they are either expecting to see a heart or a red color as the next stimuli- because only this desired stimuli would yield them a desired outcome, say, orange juice; then in this case when presented with an incongruent stimuli – a red spade- would we be able to differentially manipulate the resolution of incongruence; that is those motivated to see red would report seeing a ‘red spade’ and those motivated to see a heart would report a ‘black heart’ . Or is the effect modality specific with effects on color more salient than on form. Is it easier to see a different color than it is to see a different form? And is this related to the modality specific Sham’s visual illusion that has asymmetry in the sense that two beeps, one flash leads to perception of two flashes easily but not vice versa.

The Male and The Female Brain: from Back to Front and from Left to Right

I have been reading too many commentaries on The Female Brain (and also read an online chapter from the same today), so please excuse me if I too jump into the fray with my own discovery of a Dorsal visual stream bias in Males and a Ventral visual stream bias in Females. This is a novel departure from the usual left brain /right brain argument and deserves some attention!!

It has been often commented that the dorsal visual stream is specialized for location (and motion). Considering the combination of motion and location, one can easily see that if males process this stream more easily/predominantly then they are good at driving:-) and parking cars:-) vis-a-vis females who may not process data in this stream as preponderantly as in the Ventral stream.

Now, it has also been commented that the ventral stream is specialized for things like shape, color etc all of which enable us to identify the object. Thus, this stream is specialized for identifying objects. If women have more preponderant processing here, they would definitely be good at skills needing to treat objects like a whole- for ex relating to a person, recognizing faces etc.

I got thinking along these lines by reading a excellent commentary on Developing Intelligence regarding visual binding and you must read it before proceeding further.

As per the research mentioned there, it was experimentally found that object-location condition evoked longer looks from infants only when the objects were toys. It is evident that this ventral stream is a predominantly mean stream with focus on worldly objects and toys (I guess they had used cars as toys!)

It was also found that the object-identity condition evoked longer looks only when the objects were faces. Thus, the ventral stream it seems is tailor-made for females with their emphasis on interpersonal relationships and faces and persons as opposed to the more objective world of Men. Pardon me if reading too much Gilligan etc has gone to my head.

To me this seems as compelling evidence that not only do the female and male brains differ from left (hemisphere) to right, but also from front to back!

This post is written with a tongue-in-cheek but also takes forward some of the concepts like object and motion permanence that I mentioned earlier. It seems we need to distinguish now between object-identity permanence, object-location permanence, object-motion permanence and object-binding permanence!

Now I see it, now I don’t : object and motion permanence

Cognitive Daily has a good article summarizing the findings of recent study on 4 month old babies and how they perceive moving objects.

The study utilizes the fact that babies look longer at stimuli that are interesting or what they perceive as novel. The results of the study indicate that if a moving ball is occluded by a stationary object, then the motion prior to occlusion and posterior to occlusions would be perceived as the same motion if the time of occlusion or length of occlusion is small.

This is an interesting finding from two angles. First this study necessitates that one distinguish between object permanence and motion permanence. The former seems to be an easy to achieve property relying only on the static stimuli and should be judged only by the fact as to whether a child gropes for an object that has now been occluded and is out of sight. The latter, viz. motion permanence implicitly assumes that object permanence has been achieved. It doesn’t make sense to say that two motions that were temporally or spatially close are the same if the object undergoing that motion was not existent even when occluded.

Thus, these experiments provide further evidence that Piaget had misjudged the capacity of babies to achieve object permanence.

Endgame: does the existence of two visual pathways : one specialized for motion perception and other for location/shape/color/object mean that object permanence and motion permanence may be achieved at different ages and may have different underlying prognosis?

The power assumption : Do we really want to be the alpha males?

BPS research digest has just published some articles and one of the article that seemed to catch attention was related to pursuit of power. As per this, game theoretical experiments have demonstrated that, if given a choice, we would like to have more power over ourselves and our behaviors, than over others behaviors. This seems to be a groundbreaking study, that posits that in the dimension of power we are more motivated by being masters of ourselves than being an alpha male and dictating terms for others.

In other words, they believe we’re driven to increase our ‘personal power’ over
ourselves, but not necessarily our ‘social power’ over others.

In another post on BPS related to the brain centers engaged during eyes closed and eyes open situation in a dark room, it is posited that in the dark room condition, eyes open leads to an ‘exteroceptive’ state characterized by attention and oculomotor activity while eyes closed corresponds to an an ‘interoceptive’ state characterized by imagination and multi sensory activity. Here it is claimed that in eyes closed scenario, imagination and corresponding sensory activation are utilized. This seems to be counterintuitive, as in a dark room, with eyes open, we should be paying more attention to any threatening stimuli- why this does not lead to greater sensory acuity in visual/auditory senses needs explanation. In the eyes open, one could hopefully relax and indulge in night-dreaming.

What would be the brain activations in the closed eye and open eye conditions in a brightly lit room? Would that mean that different systems kick in, not only during eyes open and closed conditions, but also in night/day conditions?

Eve’s Lasting Legacy: The Serpent and The Apple

As per a recent scholarly article it seems that mammalian evolution may have been driven by the predatory presence of snakes. While some mammals adapted by becoming better snake sniffers, others developed immunities to serpent venom; while in the case of humans, the primates developed a good visual system to detect the snakes.

The other factor that drove human evolution (and hastened descent from the garden of eden after falling prey to serpent’s designs 🙂 ) was the fact that anthropoid ate fruits (substitute apples 🙂 ) and this frugivorus eating habit endowed them with enough-glucose-availability-in-the-brain to act as a pre-adaptation necessary to the evolution of brain matter required for visual acuity needed to detect snakes and take appropriate action.

Fox news has an excellent article on the same which is a needed reading before one can try to appreciate the excellent coverage of the same done by John Hawks.

I’ll try to summarize the arguments.

1. It is common knowledge that runaway arms-race between predators and preys lead to selective development of traits in a particular direction. For eg, the great cats and the antelopes, both developed systems for high speed chase and run-away and thus some of the fastest runners are either predators; like leopards or preys like the antelopes. What food (and energy one gets from it) also ensures who outnumbers whom in the arms race (the tiger wins!). The responses may not be symmetric, while Great Cats may develop claws and teethes, the antelope may develop antler ( though antler evolved more as species specific displays to attract opposite sex).

2. Snakes are one of the predatory species for mammals. Earlier snakes relied on Boa constriction method to kill the preys, but evolved venom about 60 mn years ago as their second weapon. Mammals reacted by either detecting them (in close range) by sniffing, or developing venom resistance etc.

3. Primates leading to Humans reacted by detecting motion (via MT and other motion detecting brain areas), color and other relevant visual stimuli to predict and detect the snake’s presence at close ranges and take appropriate areas.

4. The increased encephalisation (dependent on processing of more visual stimulus and reacting to it) was dependent on a previous adaptation related to fruit eating and abundant availability of glucose in brain.

5. The features of human vision like orbital convergence (leading to depth perception and 3D vision) are tuned for such snake -detection mechanisms.

6. The koniocellular pathway is crucially involved (among other tasks) in pre-attentional visual detection of fearful stimuli, including snakes and the evolution of this system points to snake-primate arms race pressures and how the primates adapted.

7. The Parvocellular pathway is also implicated in the study (as details and color are important for snake detection). Although the magnocellular is not , but I believe movement is also very crucial as snakes have a typical motion.

Lastly, while the analogy of the snake and the apple is quite relevant in the Christian mythology context, the snake is a revered creature in many mythologies (dragon in Chinese for example) and we in India celebrated Naag Panchami – a day when snakes are fed milk- a couple of days back.

Some parting notes:

1. In experiments with monkeys and humans it has become apparent that we have specialized fear associations for snakes. For example a young monkey, which sees another monkey as reacting in a frightened manner to say a plastic snake, would by even a singular exposure to such a display of fear, clear to have fearful associations with say the plastic snake. This association can be even when the observed behavior is seen on TV (and is recorded and not happening in real-time) Like the disgust reactions and avoidance-of-just-before-taken-food in response to a single vomit, it seems the avoidance learning for snakes is also built-in and can be triggered even by one exposure and by observational learning. Thus, there is strong evidence that we have specialized circuits for responding to snakes. It makes merit to assume that we should have for detecting too.

2. In Indian philosophy, one perennial question, focused on differentiating reality from illusion is differentiating snake for the rope. the rope in dark gives illusion of a snake, but we need to enhance our perceptions and awareness to realize that the fear of the snake is illusory and that the feared object is only a rope. This example, which is in ancient texts, is evidence of the importance of snake detection from prehistoric times.

Endgame: Can one identify from which book this drawing of boa constrictor and elephant is inspired?

4 (or more) cone vision : Tetrachromancy in Human Females vis-a-vis birds.

Cognitive Daily has a posting related to Human Female Tetrachromancy that refers to some old article on the subject. An interesting and must-read article on the web by Ryan Sutherland in detail explains the rationale as to how four cone receptors may arise due to X-chromosome related procedures. Also It is instructive to note here that if the additional cone that has shifted from Red(Long cone) towards green (the red-shifted) or from Green(The Medium cone) towards the red (green-shifted) has shifted to a considerable extent, then it may assume the role of phantom Yellow and thus lead to some radical re-wiring of the optical system in brain whereby Red(L) and Green(M) do not have to combine to give Yellow that can be considered along with output from Blue (S) cone to give rise to Blue-Yellow opponent process. In this case a simple consideration of output from Blue (S) and Shifted-red/shifted-green (Yellow) would give rise to the Blue-Yellow opponent process. I don’t think such radical shifts are possible or would lead to such radical rewiring, but post-mortem analysis of Tetrachromat women may shed some light. Even if such a shift does occur , it may not lead to any change in the number of hues that could be distinguished, though the colors may appear more colorful and saturated.

Of further interest is the shift from red away from green side towards the ultraviolet. This shift may indeed give rise to ability to perceive Hues differently and to see some infra-red not normally visible to trichromatic humans.

coming back to different dimensions of vision, it is interesting to note that dogs (like most other mammals) have dichromatic vision and utilize the blue-yellow opponent process.

Cats utilize the same trichromatic color mechanisms as humans, but their total perceivable color range is sort of ‘contracted’ i.e. they don’t see some of the human Red and some of the human Blue.

Bees have also trichromatic vision, but apparently their cones lie in UV, Blue and green. Thus they are unable to see human red but able to see beyond Violet (the UV). Maybe the genes coding blue lie on X chromosomes for Bees (instead of the red-green genes of humans and yellow of dogs) and its breakup into two (just like the breakup of mammalian yellow is hypothesized to have resulted in human Red-Green) has resulted in some infra-blue and Ultra-violet cones in the bees.

Further most birds (and some fish and turtles) have tetrachromatic vision with 4 cones : one in UV, one in Blue, one in green/yellow and the other in red. Thus, if humans do want to have a tetrachromatic vision a better way forward would be the split of blue cone in infra-blue and UV cones. That would really give us the capacity to for example view the human-white feathers of some birds as actually ‘colored’/shining’ in UV (as they reflect UV). For more details on comparative chromatic vision information please visit this excellent page on comparative chromatic vision. Also some evolutionary rationale for chromatic vision (and UV in particular) can be found here.

Endgame: Would introduction of a UV cone lead to radical changes in perception of the blue (blue-indigo-violet) end of the spectrum, just like splitting of Yellow into Red and Green led to totally new colors on the original Yellow part of the spectrum?

The green dot illusion and Opponent Process Theory

Mind Hacks has an interesting article mentioning green-dot illusion. The Green Dot illusion is possible because of the opponent process theory of color perception.

As an aside, for an excellent account on Opponent process theory and how many observable normal and abnormal behaviors may be realized as gated dipole opponent processes please read an article by Grossberg on the same.

As per this theory, as applicable to color-processing (the herring theory), the higher level processing and perception of colors happens as an outcome of 3 opponent processes. Two of these are chromatic processes : one involving red and green opponent process and the other involving blue and yellow. One supposedly “achromatic” opponent process utilizing black and white ‘colors’ is also involved. Thus, while the Hue of any perceived color may be determined by the value of the red green and blue – yellow opponent process; its Saturation (or the grayness or ‘impurity’) may be determined by adding the black-white opponent process value to determine the grayness of the stimuli while some other input (in the ‘luminance’ channel/ magnocellular channel of LGN) may be used for determining the Value or luminance or ‘brightness’ (refer HSV or HSL models of colors).

It is instructive to note that the red-green opponent process is realized by subtracting the output of Medium (green) cone from Long (red) cones and thus the R minus G signal should lead to either excitation of ‘red color perception’ and inhibition of ‘green color perception” or vice versa. Thus, depending on the signal strength and polarity, later processing by neurons would happen as opponent processes, with 1) if red is being perceived then inhibit green-perception and vice versa. Also the blue-yellow opponent process is realized by first summing the Long (red) and Medium( green) cone outputs to create a yellow ( R+G) signal and then subtracting this from the Short (or Blue) cones to give a final B minus Y signal. Again depending on the strength of this signal, either ‘blue color perception’ is encouraged and ‘yellow’ color perception is discouraged or vice versa. When later the B minus Y and R minus G signals are analyzed (and possibly aggregated), one can determine the Hue of the color depending on the relative strength of the 2 signals.

An account of how all hues can be realized using this opponent processing is explained beautifully at this site and I also include a graph from that site for illustration of how all hues (in the humanly visible spectrum) can be realized using these opponent process. That said, there still remains the issue of perception of non-spectrum colors like purple, olive green , brown etc., these have been partly addressed in my earlier post on this matter.

To sum up, the moving green dot illusion works because red and green are opponent processes. When pink (which may be conceived as low-saturation red) dots are present in the visual field, then for that portion of visual filed, Red Qualia is exaggerated and Green Qualia inhibited further down the visual pathways. Prolonged presence of Red stimuli ensures that there is no need to keep inhibiting ‘green qualia’ as habituation happens and as the Red signal is strong and continuous one so the need to inhibit Green does not arise. If one refers to the gated dipole opponent process theory of Grossberg, then it is apparent that due to the gating of the dipole, when the RED stimuli disappears from the on-channel then the ‘off channel’ (corresponding to Green qualia) would result in a sudden rebound and thus momentarily the Green qualia would be perceived. Here it is instructive to note that the signal is R minus G i.e Red is the presence of signal and green the absence (or below threshold or negative signal). Thus the green dot illusion would become more stronger if pure red is used and a similar illusion can be produced by moving yellow dot when blue dots are involved.

V1 and imagination

Small Grey Matters has a post related to an experimental finding that there is an activation in V1- the striate cortex- when the subjects make motor responses to an earlier presented visual stimuli (this is the delayed response situation as in the post/ experiment). Also, this activation is not present in higher visual cortical areas and thus is a result of bottom-up processes. One speculation as to the presence of this activation about the same time as the motor response is that when making the response one needs to ‘bring back to memory’ or imagine the earlier presented stimuli (or the no-stimuli screen) and that bringing such image back to mind is necessary for the subjects to decide whether the stimuli was present or not.

Thus, a particular mechanism for explaining this activation could be that it is related to imagining the earlier-presented stimuli and is distinguished from the actual visual experience by lack of activation in higher visual areas. The ‘imaginative center’ of the brain may send inhibitory signals to the higher visual cortical areas so that this appears as imagination and not as actual hallucinatory delayed visual stimuli.

Just speculation, but speculations that could be verified if supporting experiments are conducted by someone.

Color vision continued: What role do rods play in color vision, if any and how many dimensions/ variables we need.

There is a very descriptive and helpful book eye, brain and vision on Hravard’s site and I was going through the chapter on color vision. It is posited there that color blindness occurs if one of the 3 cone pigments are not present and consequently one is not able to distinguish white light from a monochromatic light of certain wavelength. It is also posited that for color vision only 3 types of receptors are required (and are present in the form of 3 types of cones in the retina). Now here is some experimental work that I would like done for this experiment. What happens to someone who lacks the green pigment and who is exposed to light in the wavelength of light between the non-overlapping visual fields of blue cone and red cone. As per the arguments in the book, that should lead to total loss of color (and actual colorblindness as opposed to color-defectiveness for that range of colors) and thus ability to use only rods and thus get a black and white view of world for those wavelengths. Is that really so, as per color blind people with the green cone not present?

The other thought that passed while reading the article is that it uses projection of 3 types of monochromatic light with same intensities as the metaphor of choice while describing how the brain processes color. Unfortunately as we know, the blue color cone does not overlap with red color cones and this metaphor may not be right. Even, with this metaphor it strikes one as to how black is perceived, because the picture that is shown of 7 colors (including white) produced depends on a dim room in which the 3 lights are projected and the rods that would be useful in producing this black color are integral to the experimental setup of demonstrating the tri-color sufficiency of explaining the color vision. I , personally believe that rods do have a role in color perception and color perception may more involve the CMYK model than the RGB. This also brings the ‘image formation’ metaphor over the ‘laser beam’ model. Also, at the same time, due to Kline-bottle associations I may even venture forth and propose that in reality 6 types of colors/ color detecting devices may be required to fully apprehend the colors and we may still be in the process of evolving/ detecting such pigments. Maybe the rods themselves of nocturnal animals like wild cats may throw some light. Total armchair speculation!

Interestingly, the author of the above book concedes that Brown color is a bit difficult to explain, though purple can be easily explained or be intuitive. As per this article on color naming universals which references the article Berlin and Kay (1969) published under the title ‘Basic Color Terms, their Universality and Evolution’. the brown appears in stage VI of a language evolution, where apparently as per my initial eight fold developmental model, a qualitatively different sort of leap needs to be taken. The original Harvard’s book excerpt from “Eye, brain and vision” takes recourse to Herring theory of opponent processes, specifically that of red and yellow mixing to give orange and that when seen through black contrast giving appearance to brown. Thus for brown to be explained,, the 2 extreme edges of blue-yellow dimensions and red-green dimensions have to mix spatially at a point and then this has to be seen in contrast to another extreme of black-white dimension. Seems a complicated explanation and involves taking recourse to brains excitatory and inhibitory processes to provide explanation. I might revisit this later if some more suitable explanations in terms of some other inherent property of cooler like using both the hue, saturation, value and R,G,B model may explain brown. While HSV explains purple (in the sense of it being complement of green and actually lying in the region that sort of make ultra-violet and infra-red meet), it is surprising why it is not one of the words that are found while going from stage V to stage VI of language evolutions.

Endgame: Is CMYK actually CMYKW model, with white of paper acting as background essential for the CMYK to work in reproducing images?