Squid Visual Ecology

Keeping with the theme of sensory systems, I thought I'd review some newer research on squids.

While searching for recent cephalopod neurobehavioral research (which is pretty scant) to blog about, I came upon Makino and Miyazaki's study on the distribution of retinal cells in the retina of squids.  I have a soft spot for visual neuroscience that I picked up from working with my first research advisor, who works on the visual system of frogs.  In any case, this is a good paper (although it was a bit hard to get my hand on,) and I'll review it here.

The study aims to look at the distribution of retinal cells in the retinas of a variety of squid species.  This has been done in several vertebrates, with the general finding that animals have retinas that perform well for their lifestyle.  Seems pretty simple, right?  For example, fish who live in "closed" environments have dense retinal ganglion cells (RGCs) in the area of the retina that sees light from directly ahead, while oceanic fish have a strip of high-density RGCs that stretch laterally across the whole visual field.  Thus (to make a horribly crude generalization,) cave and reef dwelling fish have focused binocular vision, while oceanic fish largely lack this but have a greater ability to monitor their whole visual field, ie. for predators or food items.

In vertebrates, retinal ganglion cells are often mapped in this sort of study.  By the time RGCs exit the retina, they are carrying visual information that is already processed into the very basic components of visual perception (namely, hue and tone contrast.)  As vertebrates have complex retinas, it is also possible to map photoreceptors in vertebrate retina, or a variety of other types of cells (which might be more or less informative.)  Cephalopods, however, only have one type of visual cell in their retina - the retinal cell (or rhabdomere.)  So, the authors chose to map this.  It is useful to keep in mind that this is not directly comparable to the mapping of retinal ganglion cells in vertebrates - it could be the case that the density of visual cells in an animal's retina is not always correlated with the importance of that piece of the visual field in further levels of visual processing.  This problem is partially solved in studies on vertebrates by the use of RGCs, in which the processing of information from photoreceptors is already underway.  With cephalopods, however, there is currently no way to probe this any deeper, and so for now it remains an assumption - albeit a pretty noncontroversial one - that rhabdomere density is correlated well with the relative importance (behaviorally and neurophysiologically) of portions of the visual field.  (For more on cephalopod visual anatomy, check out my earlier post on cephalopod eyes.)

The image to the left shows cell counts (in retinal cells per mm) across the retinas of the 5 species of squid.  I added color to this image to make it easier to see the distribution of cells.  It's important to not that the colors are relative within each figure, and do not represent absolute cell density, which is shown as (difficult to read) numbers on the boundaries of regions.  Also note the scale bars, which are 10mm in every image. 

In terms of orientation, keeping things straight gets a little tricky (as it does with all cephalopods.)  Dorsal-ventral orientation is pretty easy - remember that the lens of the eye inverts the light coming through it, so that the ventral part of the retina forms the top part of the visual field and the dorsal part of the retina forms the bottom part of the visual field.  Anterior is the direction the squids' arms point in, so the anterior retina forms the posterior part of the visual field.  The posterior retina is the part that forms the anterior part of the visual field.  This is the part that is used when squids look forward to form a binocular image.

Using this data, the authors estimated the visual axes of the squids, based on the location of the highest density of photoreceptors.  The visual axis is the general point of focus, which is known to be of utmost behavioral importance in vertebrates.  When you follow a moving object with your eyes, you are keeping it in your visual axis.  The location of an animal's visual axis is key to its visual ecology - many predators have forward facing visual axes so that they can see their prey accurately, while prey species often have very laterally oriented visual axes (think of rabbits and deer) so that they can monitor more of their environment at any given time.  Thus, we'd expect that squids with different lifestyles have different visual axes, because they will be looking for food and predators in different places.

In coastal squid (E. morsei and S. lessoniana), the visual axis is directed downwards, presumably reflecting the importance of monitoring activity on the substrate that these species live on.  In oceanic squid (T. pacificus, E. luminosa, and T. rhombus,) the visual axis is directed upwards, and the eyes have a much greater density of photoreceptors overall.  I think the retinal cell density map of E. luminosa is especially interesting, because the concentration of cells on the extreme posterior edge of the retina suggests that binocular vision is disproportionately important to this species.  The authors conjecture that this eye may be specialized to detect and track bioluminescence in the open ocean, but this is purely speculation.

These findings are important because they expand our knowledge of cephalopod eyes, which are a model evolutionary system.  If we can begin understand the impact of ecology on the organization of visual systems (which is part of the emerging field of visual ecology,) we can generate a wealth of testable hypotheses about the ecological conditions that occurred during the evolution of differnt species eyes, as well as the other sorts of adaptations we might see in sensory systems as they diverge (or converge) during evolution.  It's also a nice piece of evidence that our rather basic theories about visual ecology and the structure-function relationship of the visual system are largely correct.  This is good to know, as we base an incredible amount of more complicated neuroscience research on these theories.

Thanks for reading!


Akihiko Makino, & Taeko Miyazaki (2010). Topographical distribution of visual cell nuclei in the retina in relation to the habitat of five species of decapodiformes (Cephalopoda) Journal of Mulluscan Studies, 76, 180-185 : 10.1093/mollus/eyp055

Octopus Sensory Systems: Part 1

I've been reading M. J. Wells' book "Octopus", one of the "classic" works on the octopus.  It was published in 1978, and is an essentially complete review of research on octopuses (primarily O. vulgaris, the common octopus) up until that time.  It must have been fun to write the book, because he had done much of the research he cites (he cites at least 45 of his own papers in the book.)

Wells has published a large number of experimental studies on octopuses, and I've become very interested in his series of experiments dealing with the sensory capabilities of the octopus.  In this series of posts (there will be 3 or 4, I anticipate, over the next week or two) I'll relate his findings, as well as others, about the sensory systems of the octopus.  Forgive me if I write brief posts, as I'm rather busy writing other things at the moment.

I'll jump right into the octopus visual system.  It's a good place to start, I think, because it's relatively easy for us to imagine what it's like to see, and so to think about octopus visual perception by analogy.  It's relatively harder to imagine what it's like to control eight boneless arms (just like it's hard to think about what lateral line perception is like in fish - we simply do not have the anatomy to do it.)  We'll move on to other senses in the next few posts.  For now, though: vision!  (For a brief look at the anatomy of the octopus visual system, check out my earlier post on that topic.)

Vision in octopus has been studies mostly through the technique of visual discrimination tasks.  Simply put, an octopus is taught (via food rewards for correct answers and/or electric shocks for incorrect answers) to attack one visual stimulus and not another.  The visual stimuli are then varied, and it can be discovered what stimuli the octopus can discriminate between.  If two shapes are not discriminated between, they can be said to be perceptually identical (or at least perceptually similar) to the octopus.  These experiments were widely done (well, as widely done as anything in octopus neurobehavioral research) throughout the 1950's, 60's and 70's, with some research still going (albeit at a slower pace and with shifting foci) today. 

Octopuses, it turns out, can see pretty well.  They can reliable learn to discriminate a variety of shapes, and can even tell the difference between rectangles that are identical except that one is rotated through a small angle relative to the other.  An interesting deficit showed up after a variety of experiments were carried out, however - Octopuses are unable to recognize the difference between certain mirror images!  For example, an octopus does very poorly when discriminating between two diagonally oriented rectangles, even though they are perpendicular to each other.  This result is not too surprising, as this is the case to a greater or lesser extent with many vertebrates (to whom the octopus is often compared in its sensory abilities.)  It is, however, remarkable in its extent - it seems like a big deal not to be able to make such a discrimination at all.

In any case, these experiments gave theorists of visual perception a lot of things to think about (I'll write about their theories later in the series.)  What I find interesting about the whole series of experiments, however, is that they laid the groundwork for the investigation of the role of statocysts in octopus visual perception.  Statocysts in cephalopods are like the vestibular systems in our ears: they are tiny, fluid filled chambers that contain hair cells which detect the movement of this fluid (see The Fine Structure of the Octopus Statocyst (1965) by V. C. Barber for a more complete description.)  It appears that the main function of the statocysts is to help the octopus maintain its equilibrium by detecting changes in velocity as well as the direction of gravity.  For example, an octopus whose statocysts have been surgically removed will move unsteadily and fail to make normal eye movements in response to the movement of its environment.

Under normal conditions, an octopus will keep its pupil slit horizontal, relative to the pull of gravity.  However, when the statocysts are removed, the octopus will no longer be able to do this; her pupils will be fixed relative to her body position, and will not adjust to remain horizontal.  This is a figure from Proprioception and Visual Discrimination of Orientation in the Octopus (1960) by Wells, showing this effect:

 So what happens when you test these octopuses on orientation discrimination tasks?  It turns out that they can no longer discriminate between differently-oriented bars.  This is a big deal, because it can tell us something about the way that the visual system of the octopus processes shapes.  Specifically, it reveals that there is something important about the orientation of the retina that makes it possible for the octopus to discriminate shapes.  When the octopus cannot keep its retina level, it cannot discriminate between long shapes and tall shapes.  Thus, we know that the circuitry processing visual information in the octopus is relying on the input from the eye always being in the same orientation.

This can be contrasted with the way that people (and probably other vertebrates, although I don't know this for sure) determine orientation in their visual field, which is determined partly by non-visual cues but also largely by the presence of some sort of horizon line in the visual field (see A Horizontal Bias in Human Visual Processing of Orientation and its Correspondence to the Structural Components of Natural Scenes (2004) by Hanson and Essock for a neat demonstration of this.)  If there is no available horizon line, orientation is gathered from other parts of the context, such as common or known objects.  This means that in humans, even when the orientation of retina is perturbed, the visual processing system can use what it knows about the visual environment to adapt to the new orientation.  As long as the input is otherwise normal, it doesn't matter which way the retina is turned - that is, one can still make orientation discriminations when one's eyes aren't on the level.

This might seem puzzling or maladaptive on the part of the octopus, but a little thinking about the visual ecology of humans and octopuses reveals that it's perfectly sensible.  Octopuses, living in the ocean, don't often have very clear horizon lines or reliably-oriented objects around them.  Their strongest clue about orientation is the pull of gravity.  People, living on land, have reliable horizon lines almost all the time, so they can take advantage of this to orient themselves.  In addition, octopus eyes have muscles that can rotate the eyeball around any which way relative to the body - human eyes do not.  Octopuses never needed to evolve the complex orientation-correcting systems that vertebrates did, because they can just hold their eyes level.  They can, in a sense, afford this apparent deficiency in their visual processing system, because they compensate for it easily.  If this was the case with people, every time we laid down, bent over, or leaned to one side, the world would be full of novel objects that we couldn't make sense of!

I hope this was informative.  I can't wait to write the next one, about octopus touch and proprioception.  See you next time!

Cephalopod eyes

I just wrote a big post about cephalopod eyes, and then realized that I had neglect to show any real-life pictures of cephalopod eyes.  This blog seems to be getting a bit dry, so let's take some time off and just gaze at some of our beautiful, squishy friends.  All images are from the wikimedia commons and have been under a creative commons license.

(Photo by Parent Géry)  This guy is Octopus vulgaris, also known as the common octopus.  It the most-studied octopod.  You can see the slit-shaped pupil clearly.

(Photo by Theasereje)  Here's a body shot of another O. vulgaris.  Notice how they can look at you with both eyes at the same time - they have the capability for binocular vision.  Octopuses, however, prefer monocular vision, and will always use just one eye to sight their prey during an attack (for more info, see this "Lateral asymmetry of eye use in Octopus vulgaris" by Byrne et al.)

(Photo by Elapied)  Here we have another gorgeous shot of O. Vulgaris peering out of a hideout with one eye.

(Photo from www.opencage.info)  This is an ocellated octopus, O. ocellatus.  Besides being very cute, as he peeks out from the shell, he is probably using his mostly monocular vision to monitor his whole environment for danger.

(Photo by Jens Petersen)  Here is Amphioctopus marginatus, the coconut octopus, showing us how it can focus both eyes on the same area of space, even if it usually doesn't like to.

By now, you're probably tired of octopuses.  Let me give you a break then, and venture into the world of cuttlefish and squid!

(Photo by Bernd)  This is Sepia prashadi, the hooded cuttlefish.  Cuttlefish hunt by visually stalking their prey and then shooting out their tentacles to grab it.  Thus, they need to have a binocular field of vision so that they can accurately catch prey.  This little guy's eyes are apparent on either side of his head (look for the curved, black pupil slits.)  As you can see, cuttlefish can look in front of themselves with both eyes.

(Photo by Nick Hobgood)  This is Sepia latimanus, the reef cuttlefish.  Here you have a better view of the eye.  The eyes are not closed - the pupil of cuttlefish is always a horizontal slit.

(Photo by Nick Hobgood)  This is another S. latimanus, showing a different coloration.

(Photo by Nick Hobgood)  This is Sepioteuthis lessoniana, the bigfin reef squid.  Most squid normally have mostly monocular vision, but can move their eyes towards the front of their head to have temporary binocular vision.

Euprymna scolopes (Bobtail squid in orange form).(Photo by Nick Hobgood)  This is Euprymna scolopes, the Hawaiin bobtail squid.  In this photo, you can see the cuttlefish-like pupil shape and the existence of binocular overlap.

(Photo by Michael Vecchione)  I'll leave you with the bizarre-looking eye of Helico pfefferi, the piglet squid.  I don't know anything about them, but they sure look cool.

Thanks for reading!

The Octopus Visual System

Visual neuroscience is one of my favorite sub-fields of neuroscience.  As such, I've decided to write this post about the visual system of Octopus (I will leave out squid and cuttlefish for this post, but theirs are, overall, pretty similar.)

First, let's take a look at the anatomy of the octopus eye, in comparison to a more familiar specimen (this figure is from Ogura et al., 2004):

They look pretty similar, don't they?  Notably, the both have lenses to refract light, a pupil to restrict incoming light, a retinal surface distributed along the back of the eye, and a well-differentiated optic nerve.  There are, however, a few notable differences.  Firstly, octopus eyes and human eyes focus light on the retina in different ways.  Humans bring their eyes into focus by changing the shape of our lenses using the cilliary muscle.  Octopuses, on the other hand, focus by moving the lens closer to or further away form the retina.  This is the same way that a camera focuses, and so people often refer to cephalopod eyes as "camera eyes."  Secondly, the octopus retina is very different from the mammalian retina in terms of its cellular organization.  Let's look at an octopus retina first (From Young, 1962).

Light enters the eye from the top of this image.  In terms of cellular organization, the limiting membrane (l.m.) at the top of the figure faces towards the front of the eye, with the basal membrane (bas. m.) towards the back of the retina.  The distal segments (dist.) of photoreceptors face the front of the eye, and the proximal segments (pr. s.) sit at the back of the retina, from which they send their projections through the optic nerve to the optic lobes where further processing of visual information takes place.

To compare, let's check out a few diagrams of the mammalian retina (both of these are taken from Wassle and Boycott, 1991 ; the first is their own drawing, and the second is by Ramon y Cajal, one of the early geniuses in neuroscience):

Now, you wouldn't know this from looking at the figures, but the mammalian retina is actually opposite in orientation from the cephalopod retina.  That is, light enters from the bottom of these drawings, and passes through several layers of cells before reaching the photoreceptors at the back of the retina (the photoreceptors in these images are rod-like structures at the top of the images, labeled as "OS" and "IS" (for "outer segments" and "inner segments", respectively) and as "e" and "f" in Cajal's drawing.)

You, being a smart and inquisitive reader, might just now be asking, "what is all that other stuff doing in our retinas?"  It turns out that it processes the information that photoreceptors gather about patterns of light right there in the retina.  By the time retinal ganglion cells (labeled "GLC" and "c" or "d" in the two images) leave the mammalian retina, they are already carrying information about properties of the visual environment, such as where isolated, high contract areas of light are and what the overall level of illumination in different areas of the visual field is.  This information can then be used by visual processing centers in the brain to very efficiently form a picture of the visual world, including identifying objects and movements.

At first glance, it looks like the octopus must miss out on some visual processing power because it does not have this visual processing system intrinsic to the retinal.  It makes up for this, however, in a very interesting way.  Let's look on this drawing of the organization of the octopus optic lobe - the place in the brain where the octopus optic nerve ends up (this image is from Young 1961):

Young makes the argument, in this paper, that the outer layers of the octopus optic lobes (which are nestled right under the eyes and directly innervated by the retina) are likely to have the same function of the complex layers of the mammalian retina, with the deeper layers analagous to mammalian central visual processing areas.  As far as I know, there has been no direct test of this hypothesis.  However, it is an interesting and (at least circumstantially) sensible idea, and it must be the case that octopus forms perceptions of objects and movements are some central location - after all, it does have rather good vision.  The optic lobe, after presumably making sense of what the retina is saying, sends projections all over the superoesophageal mass of the brain.

Why does this matter?  Who gives a bibble?  Well, the evolution of the eye as long been used as a sort of model evolutionary system, that is, we can probably use it to understand evolution, and to determine what principals of genetic selection drive evolution.  It's suitable for this for a few reasons: we have extant animals with a variety of types of eyes, we have extant animals with no eyes, the eye is relatively accessable and easy to study (and so the eyes of mammals and insects has already been extensively studied.)  As a discrete organ, it is also a relatively straightforward task to characterize morphological and cytoarchitectural variation between species.  In addition, the visual system is interesting to study because it is one of the few systems that we can understand and model on a fine-grained computational level.  Thus, the differences between cephalopod and human eyes are important because they can not only tell us things about how our eyes work, but help us figure out the way that evolution works.  Evolution, as we know, is important to understand because it is, quite literally, the reason we are here.

It seems that there is probably an optimal organization to any visual processing system that involves progressive layers of increasingly synthetic processing of incoming visual information (for example, the progressive abstraction of visual input in humans from light to features, and then to objects and scenes.)  This point is strengthened by the fact that insects appear to share this style of processing visual information, albeit with anatomical differences (for more information on insect visual processing, see this paper).

 I had wanted to talk about the vestibular system of cephalopods in this post as well, but it's already lengthy and I need to get going.  I'll do a post on it another time.

Thanks for reading!