Saturday, May 28, 2011

How Horses See: Facts and Fallacies about Equine Vision




If we could see the world through the eyes of a horse, what would it look like?  Most of us have probably wondered about this at some time – especially when our horses spook at something we can’t see, or something the horse has already seen and apparently accepted when going the other way.  Unfortunately, many of the “facts” we have been told about equine vision are anything but factual, and there has been little in the way of hard evidence to dispel the myths, assumptions, anecdotes and old wives’ tales that permeate the horse world.  The good news is that there are now a number of scientists working to unravel the mysteries of the equine eye, and their research is giving us a better understanding of how our horses see the world.
 
A leading researcher in this field is Brian Timney, PhD, of the University of Western Ontario.  A professor of psychology and associate dean in the faculty of social sciences, Timney has been studying vision for more than twenty years.  With the help of both graduate and undergraduate students at the university, Timney has undertaken a systematic study of equine vision that is shedding new light on the horse’s visual acuity, depth perception and color recognition.

Nearsighted, Farsighted or 20/20?

            Depending on who you listen to, you will hear that horses are nearsighted, farsighted, neither or both.  This is one of the many areas in which Timney found previous scientific literature to be both “sparse and contradictory”, so he set about devising a way to test the visual acuity of horses.  This was a rather tricky proposition, since horses can’t read an eye chart and tell us what they see. 

What Timney and his team came up with was a method using an apparatus that had two swinging doors on which different visual displays could be placed.  A treat could be placed behind either door, and the horse could get the treat by pushing the door with its nose.  Once the horses had learned how to do this, the researchers placed a visual display on each of the doors, one that had a pattern of black and white stripes, and one that was plain grey.  The treat was always to be found behind the striped door, though the striped door might be on the left or the right in the various trials. 

The horses quickly learned to choose the striped door with nearly 100% accuracy, and that is when the testing really began.  As Timney explains, “Over several days, we increased the difficulty of the task by making the stripes finer and finer, until eventually they were so fine that the animals could not distinguish between the stripes and the grey.”  The accuracy of the horses went down to about 50% at that point, as they were just choosing at random.

The results of this study suggest that horses are slightly nearsighted compared to a person with normal vision.  Timney rated the horses at about 20/30 on the Snellen scale, which uses 20/20 to denote normal human vision.  This means that to a horse, an object that is 20 feet away would appear about as clear as an object that is 30 feet away to a human with 20/20 vision.   As animals go, this is pretty good (cats, for example, have 20/75 and rats 20/300), and Timney believes that horses can therefore likely see most of the things that we see.

Plumbing the Depths


            Another common fallacy about equine vision is that horses have poor depth perception.  How a horse could clear a course of jumps, run at speed over all kinds of terrain, or even avoid bumping into walls if this were true defies logic, yet the myth persists.  This is largely due to the fact that the majority of the horse’s field of vision is monocular (meaning that objects are viewed with only one eye), and there is a common misconception that binocular (two eyed) vision is necessary for depth perception. 

Horses do, in fact, have a substantial field of binocular vision (55-60 degrees in front of them), but as Timney points out, even monocular vision is not necessarily flat.  “There are lots of monocular depth cues,” he states, “and anyone can judge for herself whether the world looks flat when one eye is closed.”  Humans use both monocular and binocular cues for depth perception, and Timney suspected that the same was true for horses.   To test this, he and his team once again devised a series of ingenious experiments, with some surprising results.

The first experiment tested the horse’s ability to perceive monocular depth cues, which are sometimes called pictorial, as they are the ones that allow us to recognize depth in pictures.  As Timney explains, “We set up a monocular test that takes advantage of a well-known visual illusion called the ‘Ponzo effect,’ in which the sense of depth is conveyed by a flat image when lines that appear as separate at the edge of the picture converge near the center.”  The classic Ponzo effect image used by the researchers was that of a set of receding railroad tracks with two horizontal bars superimposed upon it.  One bar was higher on the page and therefore overlapped the converging rails; the other was lower on the page and was therefore well between the two rails.  The two bars are actually the same size, but the pictorial depth cues create the illusion that the higher bar is longer. 

The team first taught the horses to choose between two pictures that had no depth cues, one that showed two horizontal bars of equal length, the other that had horizontal bars of unequal length with the longer one on top.  The horses were rewarded only for choosing the picture with unequal length bars, and they soon learned to do this consistently. Next, the researchers superimposed two sets of equal-length, horizontal bars on two different photographs.  The first lacked strong depth cues, so to the human eye, the bars appeared of equal length.  The second superimposed the horizontal bars on a photograph of railroad tracks taken specifically to illustrate the Ponzo illusion, and thus the upper bar appeared longer to the human eye. 

Shown these two images, the horses overwhelmingly chose the picture with the bars superimposed over the railroad tracks, indicating that they perceived those bars to be uneven.  “Not only could they recognize depth in pictures,” Timney says, “but they also must be susceptible to the same kinds of visual illusions that we perceive.”

Timney also wanted to test the role of binocular vision in equine depth perception.  In a series of several experiments, the researchers utilized random-dot stereograms, which are pairs of pictures that look like flat patterns of dots when viewed normally or with just one eye.  “However,” explains Timney, “if you direct one image into each eye, a shape jumps out in depth if you have stereoscopic capability.”  In human vision testing, one picture is tinted green and the other red, and colored filters over the eyes allow the images to be viewed separately (similar to those 3-D glasses you get to watch a 3-D movie). 

Timney and team devised comparable colored goggles for the horses and showed them two sets of stereograms, one with depth cues and one without.  “Much to our surprise,” he says, “they responded to the images as if they saw depth in the appropriate one.”  Though this doesn’t definitively prove that horses have stereopsis comparable to ours, Timney concludes that “Horses apparently have many of the same depth-detecting skills that we have, despite the lateral placement of they eyeballs.”

Is the Grass always Greener?


There has been research looking into whether or not horses have color vision as far back as the early 1950s, but once again, the data and conclusions had not led to any definitive answers.  Some studies indicated that horses could see certain colors but not others, while other studies seemed to show that some horses could see more colors than other horses.   Anecdotal evidence from horse owners who had animals that preferred to drink from certain colored buckets or would run away if people wore a certain colored clothing seemed to indicate that horses could see at least some colors, but no one was absolutely sure.  

In looking at many past studies in this area, Timney realized that one of the problems that rendered them so inconclusive was that they didn’t adequately control for brightness cues that might have skewed their results.  “For example,” Timney states, “an animal may be trained to discriminate a blue from a yellow light.  But if one of these appears brighter than the other, the animal may be using this difference in luminance, rather than chromatic information, to make its choice.”  

Timney and team therefore incorporated careful controls for luminance in their experiments.  “We tried to control for these brightness problems by asking the horses to choose a color on a gray background that we could vary from light to dark,” he says.  “When we did this, we found that they could always pick out a red or a blue, no matter what the background was like.  However, they could pick out yellow and green reliably only when these colors were brighter than the background; once the brightness was equal, the horses could no longer discriminate the colors as easily.”  Timney’s results tell us that horses do have some color vision, but he stresses that it will be a long process to really learn how their color vision compares to our own.

Show it to him on Both Sides


            Another well respected researcher who has done myth-shattering research into the vision of horses is Evelyn B. Hanggi, MS, PhD, of the Equine Research Foundation in Aptos, California (www.equineresearch.org).  As a life-long horseperson, Hanggi was aware of the widely held belief that horses cannot transfer information from one side of the brain to the other, and that therefore, they must see something with both eyes before they can recognize it as the same object.  As a doctor of biology and an equine cognitive behavioralist, however, Hanggi knew that this didn’t make sense.

            The two reasons most commonly given as to why horses would be incapable of interocular transfer of learning are their lateral eye placement and the misconception that horses have a minimal or nonexistent corpus callosum, the structure that connects the left and right hemispheres of the brain.  In reality, anatomical data disproves both of these assumptions.  Animals with laterally directed eyes (like horses) have more crossed optic nerve fibers than animals with frontally directed eyes (like humans), which indicates that information is transferred from one side to the other.  In addition, horses do in fact have a substantial corpus callosum.  Yet as Hanggi says, “Old beliefs persist, even when physiological evidence is presented.” 

That is why she set out to devise and carry out a behavioral experiment that would demonstrate and quantify the degree to which horses were capable of interocular transfer.  In her experiment, each horse had an eye patch fitted that completely blocked the vision of one eye, and in a series of trials, was taught to choose a specific shape from a pair of black and white images.  This was done with four pairs of images.  In each case, the horses started out choosing the correct shapes at a rate that suggested pure chance, but they soon improved as they learned which choice would be rewarded. Once the horses were correctly choosing the desired shape 90% of the time, the researchers switched the patch to the other eye.

            If the horses were incapable of interocular transfer, their success rate with the second or “testing” eye would have been the same as it was with the first or “trained” eye.  It would start off at around 50%, and then gradually improve over the series of trials until it reached the 90% goal (criterion).  Contrary to what many in the horse world would have expected, this was not at all the case.  Says Hanggi, “When the trained eye was covered, both horses reached criterion with the testing eye immediately except for one horse on one pair of images.  However, even with this pair, that horse responded at near criterion levels from the start and did reach criterion within a couple of sessions.”

With the results of Hanggi’s experiment, there can be no doubt that horses are capable of transferring information from one side of the brain to the other.  Why, then, do horses so often react in ways that appear to the contrary?  The answer may be as simple as the fact that things look different when approached from different angles.  Adds Hanggi, “We can guess forever on this but my take at this time is more based on training history.  Horses that have been exposed to all sorts of sights, sounds and smells in a variety of environments and situations do not startle nearly as much as horses that are housed in stalls, rarely turned out, isolated from other horses, stuck doing the same routine every day, not exposed to desensitization and generalization training, and not allowed to ‘be horses’.  These horses are, in my opinion, more likely to become upset or startle when something changes or looks different.” 

Seeing in the Dark


            One area that has yet to be scientifically examined is the degree to which horses have night or “scotopic” vision.  The common view, based on anecdotal evidence, is that horses have better night vision than humans.  As Hanggi says, “Simply watching the behavior of horses in dim light provides insight into this.  Whereas humans have difficulty moving around in dim light or darkness, horses do much better and many a rider has had to rely on his/her horse when caught out after dark.”  We also frequently hear that it takes horses longer to adjust to changes of light, but neither Timney nor Hanggi knows of any research to support or disprove this, either.  What we do know is that the retinas of horses have a large number of rods, which are necessary for good night vision.  Horses also have a tapetum (the structure that makes the eyes of cats shine at night), which should also help them to see a little better in very dim light.

However, as Timney explains, “The question of what is ‘better’ night vision is an ambiguous one.  Individual human rods are as sensitive as they possibly could be; a rod will react to a single quantum of light.  But we don't see details very well at night because of the way that the rods are wired up.  Many feed into a single ganglion cell, which means that the cell will respond to light very well, but will not necessarily distinguish its location.  This means that scotopic acuity is very poor.  For horses we don't know the wiring diagram, but it's not likely to be much different from humans. The story is a little more complicated than that, but I suspect that they won't be that much ‘better’ than humans.” 

As for the response to changes in illumination, Timney says, “I think that all rods contain the same kind of photopigment, and so the way that changes in response to light would be the same for all species.”  In other words, all eyes need time to adjust to light changes, and horses are not likely to need more time than we do.  Therefore, if horses seem hesitant to move from a light area to a dark one – such as they often have to when going into a trailer – it may have nothing to do with their vision.  It could be their general mistrust of the situation, a training issue, or they may have some deep instinct that tells them that predators often lurk in the shadows.

The Ongoing Quest


            Though science has taught us much about equine vision in recent years, we clearly still have a great deal to learn.  Fortunately, researchers like Timney and Hanggi are continuing to push the boundaries of knowledge in this area, and we may soon know much more.  Hanggi has recently concluded a study of her own on equine color vision, and her results will soon be available to the scientific community.  Timney’s work is also ongoing.  For now, many functions of the equine eye remain cloaked in mystery, reflecting, perhaps, the great and unfathomable mysteries of the equine soul.

Thanks for reading! If you would like to be notified when new articles are added to this blog, just click the "follow" button on the upper left part of any page.