The pigeon in the machine – How machine learning can aid the study of animal behaviour

What if someone told you that pigeons can tell the difference between artwork painted by Picasso and Monet? Or that they can discriminate between different facial expressions in humans and have been able to identify malignant versus benign breast cancer. You may be surprised and like me, want to find out how this humble bird can achieve these challenging tasks. Pigeons have good visual memory. However, researchers do not select birds that have specific degrees in arts or radiography to help distinguish between artwork and tumours, they have to provide the pigeons with examples of the data they ultimately had to categorise. The pigeons were given time to learn the differences between the data through several training trials in which they were rewarded for making the correct decision and then were given the opportunity to test what they had learned on data that they had not seen before. This type of task is known as classification since the data is being separated into categories. 

There are many tasks in the research of animal behaviour that require the discrimination of data. Since we can’t all keep and train pigeons to carry out tasks such as identifying a bear from a deer in camera traps we need a faster way to carry out these tasks. This is where machine learning comes in. Machine learning is a branch of artificial intelligence that can be applied to solve many different tasks such as speech recognition, driving cars, facial expression recognition and is even behind netflix algorithms that tailor the content they suggest based on our watching data. Put simply, they are computer algorithms (mathematical procedures) that learn patterns in data they are provided with to make predictions about new data. They come in different flavours, for example there are algorithms that classify data, placing labels on them in the same way pigeons classified paintings. Others cluster data ‘deciding’’ a place to separate the input data based on their similarities and differences, for example, when a dog scent marks or not based on the accelerometer data placed on the dog’s pelvis. Or regression where you would like to predict an outcome based on data you have by exploring the strength of the mathematical relationship between the input and output data. For example, predicting the diving behaviour of seabirds just from features of GPS data such as the altitude and coverage data (the number of signals present and missed over specific time).

Identifying the difference between an image of a dog or cat may seem like an easy task for us since we use a holistic way of identifying these animals but computer algorithms ‘see’ this data differently and extract information (features) from the data. These could be pixel colours, edges of the image such as the side of the face, or more abstract data structures that we can only dream about. Data can be provided as raw data and the algorithm extracts what it deems important information and carry out the task without your input. These methods are termed unsupervised and you have little knowledge what it is about the data the algorithm learns structure from but are useful in carrying out a task quickly and using patterns in data we might not notice. If you had more ideas about the informative parts of the data you would like to evaluate, use a supervised approach. Here you can select and label which parts of the data go into the algorithm and this technique lends to certain research questions that have more context since you are interested in what the model learns from rather than just its output. For example we know these fur seals and these sea lions are foraging and grooming because of these informative features of accelerometer data.

Once we have decided what type of machine learning algorithm to use, like the pigeon, we provide it with some example data to learn from. The pigeon required 144 trials over the length of 15 days with food reinforcement when they made the correct distinction between malignant and benign tumours. Similarly the algorithm will need several iterations of learning the pattern in the data, hopefully hours and not days!, before it is applied to new data that it has not hasn’t seen before. The pigeon reached an accuracy of 85% when it was asked to label the tumours of new x-rays. Models that reach an accuracy similar to the pigeon on test data are deemed good and you can then use it to start exploring your data. However if the researchers also wanted the pigeon to help diagnose other cancers, they would be at a loss since the pigeon learned that only one type of cancer exists. This is an example of over-fitting, algorithms can become too restricted if they are provided with training data with little variation. Models are only as good as the example data they are provided with. A model that applies what it has learned well to new data is one that has been trained on data that reflects its diversity. If you wanted a model to identify a dog, you wouldn’t provide it with examples of just Labradors but rather an example of all dog breeds.

Why use machine learning to study animal behaviour? Machine learning is being used and applied in studies of animal behaviour and cognition. Often in these studies decisions have to be made about how to capture the behaviour of interest, describe it and analyse it. Some behaviours, such as the motion of mouse whiskers, may be imperceivable or at least challenging for a human observer to notice. I doubt that a person would be able to identify the different species of flying insect pests from a video of them but a group of researchers were able to train an model to do so which would help in targeting methods to protect crops. Machine learning also enables the automation of time consuming tasks such as identifying species from video data, can enable researchers to collect and analyse more data in shorter time frames and this might be useful if they want to respond to population declines rapidly with conservation efforts. Or labelling the behaviour of animals in videos automatically, reducing human error which is inevitable after staring at a screen for hours on end. We must not forget that these models learn from the data we provide them and do not see data in its context the same way that we do. Therefore, they can not replace how we as humans approach and solve problems but there is no doubt that they offer exciting new ways to delve deeper into phenomena in the animal kingdom.

Next time you come across the humble pigeon strolling across the pavement just take a minute to consider what it notices and might remember in its environment. Maybe your local pigeon even recognises you.

References:

Bidder et al (2020). Monitoring canid scent marking in space and time using a biologging and machine learning approach. Scientific Reports. https://doi.org/10.1038/s41598-019-57198-w

Browning et al (2017). Predicting animal behaviour using deep learning: GPS data
alone accurately predict diving in seabirds. Methods in Ecology and Evolution. DOI: 10.1111/2041-210X.12926

Jan-Harm et al (2020). WhiskEras: A New Algorithm for Accurate Whisker Tracking. Frontiers in Cellular Neuroscience. https://doi.org/10.3389/fncel.2020.588445

Janelle Shane (2019). You Look Like a Thing and I Love You. https://www.janelleshane.com/book-you-look-like-a-thing

Kirkeby et al (2021). Advances in automatic identification of flying insects using optical sensors and machine learning. Scientific Reports. https://doi.org/10.1038/s41598-021-81005-0

Ladds et al (2016). Seeing It All: Evaluating Supervised Machine Learning Methods for the Classification of Diverse Otariid Behaviours. PlosOne.
https://doi.org/10.1371/journal.pone.0166898

Levenson et al (2015). Pigeons (Columba livia) as Trainable Observers of Pathology and Radiology Breast Cancer Images. PlosOne https://doi.org/10.1371/journal.pone.0141357

Watanabe et al (1995). Pigeons’ discrimination of paintings by Monet and Picasso. Journal of the Experimental Analysis of Behaviour. 10.1901/jeab.1995.63-165

My female dog pees like a male dog…Or does she?

Drawing by Bea Pagano

If you were asked to picture a dog scent marking you might visualise a dog standing near a lamppost with their back leg raised. You would be correct as this posture helps dogs to be more specific about what they leave a mark on. However, like me, you might presume this dog is male. I recently noticed that my female dog has been marking up the side of objects in a way I previously thought only male dogs did. As someone who has been interested in dog behaviour, I realised my lack of knowledge about a behaviour I see so often on my dog walks. I rushed home and did some research straight away. What I found out confirmed why I had not considered female dog scent marking. For a long time it was understood that male dogs scent mark via urination as a way to share information about themselves such as their social status in order to compete for mates, whereby females only urinate to eliminate. This understanding is being updated thanks to more research from the female dog’s perspective.

Urination in female dogs is actually quite fascinating. They do not have one specific stance and there are individuals who differ in the way they urinate. Some may squat, others elaborate it with a handstand transition and other times, they will raise a back leg to help target objects. This also depends on whether they are scent marking or not. The difference between scent marking and urinating to eliminate is quite obvious. Most of the time scent marking occurs when the urine is directed towards an object or place, sometimes while or just after the dog has sniffed the area and usually only with a few drops of urine. Studies that focused on female dogs found that even non-oestrus dogs who had been spayed, scent mark by the squat leg-lift and they do so more often as they get older. This extended upon previous research that had concluded that female dogs scent mark as a means to solely advertise reproductive status. 

Female dogs using different postures while urinating. Taken from Wirant and McGuire 2004

Early studies exploring the meaning of scent marks in dogs found that male dogs presented with the urine of different dogs were more interested in the urine of females that were on heat than those that were not. Concluding that scent marking was mostly used to assess and compete for mates fit with an outdated concept at the centre of sexual selection theory. That is, that males are competitive for access to mates and females are choosy and at the same time, passive. If we only investigate phenomena through one-sided perspectives, we will never understand it in a nuanced and whole way. For example, singing in songbirds was considered a male only behaviour linked to competition for mates before a recent study put this concept to bed. In fact females also sing in 71% of 229 species of songbirds suggesting its importance for both male and female birds in competition.

Scent colours the world of our canine companions. We owe it to them to at least attempt to understand their smelly world. Scent marking is similar to leaving a post on a social media platform, it allows indirect communication of information about yourself without the risk of a physical altercation. Are females and males that different? Maybe in the physical and anatomical constraints, resulting in different postures while urinating. However it appears that scent marking is used by both to share information about themselves. The information dogs share on their “nosebook” may be of social status, age, sex, health and reproductive status and maintaining territory boundaries, although this was studied in free-ranging dogs that happened to live in packs. Dogs even recognise their own urine and that of familiar household dogs. Both male and females counter-mark, scent mark over or close to another dog’s scent mark but males seem to mark more often in one sitting than females. There are also individual variations in scent marking behaviour which begs us to question the generalisations that we make in explaining animal behaviour. For example, male dogs show variation in how high they leave their scent mark, smaller dogs tend to mark higher than larger dogs, maybe as a way to deceive others about their size. Does this also happen in females? To my knowledge, it has not been studied yet. 

The picture of why and how dogs scent mark is being filled in by more representative colours. To continue to better understand a phenomena we must explore it through all perspectives involved which may help counteract biases in our interpretations of data we see around us. So does my female dog pee like a male dog? No, she just pees like a dog.

Dogs are human behaviour experts

Whenever I drive around with my dog, be it to visit friends or go for a walk far from home, she anticipates when I am about to stop and park. My first thought was that she recognises areas we are going to or notices my car slowing down before parking. However, even when I go to a new place she will start squeaking before I even turn the corner to the street I aim to park on. It can’t be the slowing down of the car since she won’t squeak in traffic or at traffic lights. I have come to the conclusion that she must be perceiving something about my behaviour. Maybe the way I breathe, or the way I look around. Maybe I make subtle changes in the way I drive the car or give off a certain smell when I feel lost or am looking for the correct streets to drive down. It is almost like she is one step ahead of me, predicting my behaviour before I even move. This got me thinking about what else dogs understand about humans. What do they study in us that helps them to navigate their human dominated world? What does my dog see when she watches me and does she take note of my behaviour? 

It’s not a stretch to assume that dogs notice our behaviour. They rely on us for a lot. To receive food, go for a walk, stay dry, warm, clean, groomed and sometimes for company. Even free-ranging dogs, which the majority of dogs in the world are, interact with humans on a daily basis. For these dogs knowing which humans to approach and avoid is beneficial since humans mostly initiate both positive and negative interactions with them such as providing food or chasing them away. Our behaviour is relevant to dogs and they can understand our actions and intentions thanks to observing and experiencing life with us. 

Dogs have developed the abilities to understand human gaze and gestures. They know when their owner is looking at them or not, increasing attention grabbing behaviours such as whining when humans look in their direction rather than away into the distance. Dogs will also produce more facial movements, raising their eyebrows more often, when their owner is facing them than looking away regardless of if there is food present or not. They can also follow a pointing gesture moving to the place humans point them to. You can test this out by standing in front of your dog with a bowl of food placed at either side of you and point at the bowl you want your dog to go to. Even free-ranging dogs can do this, they have also grown up around humans and have learned what pointing entails. These abilities have been considered special in dogs because of their long history of domestication however they are not alone in being able to follow human pointing gestures. It seems that just growing up around humans is enough for some non-domesticated species to also learn to follow human gestures. For example bats( three captive bats) and wolves raised in human environments can follow human pointing gestures. Just like any other expert, dogs need to train in and experience their subject over time.  

Apart from understanding our intentional gestures, dogs are aware of parts of ourselves that we might signal unintentionally. Our emotions can be expressed by our behaviour, body postures, facial expressions and the scent we give off. Being able to recognise other’s emotions can help an animal to act appropriately in social interactions. In order to study whether dogs can recognise human emotions researchers can train dogs to distinguish between cues such as happy or sad faces to see if they can learn the difference. However if you are training a dog to tell the difference between a happy and angry face for a food reward, even if the dog distinguishes between the two expressions it’s hard to know if they categorise them based on appearance alone or on the emotional meaning of the expressions. A group of researchers came up with a way to explore if dogs were categorising angry and happy faces based on their emotions by first training dogs to distinguish between the expressions of only half of a human’s face. They then tested whether the dogs could transfer what they had learned to do, either touch the happy or angry face, on completely different sections of the face such as just the upper, lower, left or right half of the face. Dogs were able to distinguish between the two facial expressions suggesting they understood they were of two different emotions.

Another way to explore if dogs can recognise emotions in humans is to present them with emotional cues from humans and see how they are affected by them based on their behaviour and physiology. Researchers presented dogs with the smell of humans in a fearful or happy emotional state. These scent samples were taken by placing cotton cloths in the armpits of male participants while they watched fear inducing or happy videos. The dogs that were exposed to fearful human scent were more stressed themselves, based on the behaviours they were exhibiting than dogs that were exposed to the scent of happy humans. They also had higher heart rate which shows more physiological arousal expected when in a fight or flight state. However, the scents might be more meaningful to dogs if they were of people they know since it is unlikely dogs will just experience male humans in their everyday lives. Dogs can also integrate these multiple cues indicating that they can perceive some emotional meaning from them. Seventeen dogs have demonstrated this through an experiment whereby they were presented with faces of happy, angry or aggressive humans alongside a vocalisation that matched or didn’t  match the emotion. Dogs looked at the images with the matching emotional vocalisation longer than ones that did not. 

Dog’s abilities to study and decode human behaviour have been harnessed by us. We rely on them just as much as they rely on us. Dogs can be trained to alert their owners to low blood sugar levels, blood pressure, seizures  and anxiety as well as other tasks related to changes in human scent and behaviour. We shouldn’t forget that our canine companions are acutely tuned into even the subtlest of our behaviours. It appears that the interest in studying each other is mutual.

Do dogs know who they are?

Image: https://pixabay.com/  Michelle Badenhorst

A dog and a human are sitting opposite each other with two screens placed between them, one opaque and the other clear. Behind each of the screens, on the dog’s side and visible to the dog, is a toy. The human asks the dog to bring them a toy and the dog chooses the one that is visible to the human, the one that is behind the clear screen rather than the one behind the opaque screen. Surely the human must be giving some information to the dog about which toy they want them to bring by gazing at it? Well, this does not seem to be the case. What happens if there are two clear screens, or the human sits on the same side as the dog with a view of both toys? In this scenario the dog randomly brings one of the toys not favouring one over another.

When I came across this study showing that dogs are aware of the perspective of a human, I started to wonder how dogs perceived themselves. Are they aware of themselves, what they know, how they fit into their physical and social environment? In other terms, do they have a sense of self? These are questions about cognition, how you gain and process (make sense of) information from the environment and meta-cognition, being aware of your own perceptions and knowledge. Knowing what we know and what we do not know. These cognitive abilities are challenging to study in animals other than humans since they are unable to describe to us how they experience their sense of self. Instead, we focus on studying specific aspects of cognition and combine pieces of knowledge gained together into a puzzle that might give us a glimpse into the minds of other animals.

How would you or I be aware of how we look at this moment? We would probably look at our reflection in a mirror. This is what many scientists have used to test whether animals can recognise themselves. They usually place a mark on an area of an animal’s body that they cannot easily see and often when the animal is asleep or sedated so they don’t feel it. They then present the animal with a mirror and conclude that they recognise themselves in the reflection or not based on their behaviour. This experiment has been carried out on several different animals such as ants, fish, birds and apes. Many have failed the test, meaning they did not look at, investigate and touch the new mark on their bodies through the mirror (although apes and ants have responded to their reflections). You may not be surprised that dogs fail as well. If we think about it, the mirror test has a human centred viewpoint. It assumes that animals primarily use vision to perceive themselves, others and their environment and assumes they understand the properties of reflection. 

Therefore, we owe it to animals to explore cognitive abilities through a more biologically meaningful way by focusing on what we know about how the animal senses its environment. This was the main argument by researchers looking into whether dogs recognise themselves from the smell of their own urine. Those who have dogs will know that a good portion of dog walk involves standing and waiting for your canine friend to get the latest updates from ‘nose-book’ and maybe even leave their own post for others to find. The researchers gave 12 dogs the opportunity to smell their own urine and that of another dog. To no surprise dogs preferred an unknown dog’s urine to theirs, investigating it for longer. It may be that the unknown dog’s urine is interesting to them whereas their own scent was nothing new. Dogs were also presented with their own urine that had been modified, analogous to the mirror test when a mark is placed on an animal’s body. They added anise essential oils into the dog’s own urine and presented it alongside a canister containing just the essential oil. Dogs investigated their modified scent for longer than the essential oil smell. It is likely that the dogs recognised their own scent and found it interesting or even strange that it smelled different. Maybe the dogs were just interested in the new scent of the essential oil? If this was the case, they would have investigated it for longer or the same amount of time as their modified scent. In this study dogs also investigated the smell of a familiar dog in a similar way as their own. What can we conclude from these findings? We can say that dogs perceive familiar and unfamiliar scents, but we cannot know what they perceive, or what meaning they gain from them.

Another form of self-awareness is knowing how much space your body takes up. Being aware of their size might be beneficial to animals if they need to hide from predators or prey, avoid falling off or over objects, or getting stuck. This question was explored in a simple way by providing dogs with different sized openings which they could pass through to gain access to a food reward or their owner placed on the other side. Some dogs were exposed to large holes or small holes first then the hole size changed to a midpoint between the two extremes. The amount of time it took dogs to start moving towards and to reach the different sized holes gives some insight into their perception of whether they would fit through or not. Dogs took longer to approach the small than the large hole suggesting some hesitation about whether they would be able to fit through the opening or not. Dogs appear to have some awareness of the size of their physical bodies and their interaction with physical objects in the environment. What about being aware of their own behaviour and actions? This has been investigated in dogs through studying their memory.

Episodic memory is remembering an event that took place at a certain time and place. This type of memory has been studied in bird species, such as jays, that naturally hide the food they have collected. The birds remember where they hid something, what it was and how long ago. For example, scrub jays would only reclaim perishable food such as wax worms from the place they had buried it after a short amount of time had passed compared to longer periods of time as it would be likely that the food  had decomposed by then . A group of researchers at Eötvös Loránd University, Budapest explored this memory ability in dogs. They wanted to see if dogs could remember an action that they themselves had carried out after some time had passed. They trained dogs to learn that ‘repeat’ meant repeat the behaviour they had just carried out, either a trick or a spontaneous behaviour such as drinking water or lying down. The dogs were able to perform their last behaviour when asked to repeat it after 20 second, 1 minute and 1-hour delays. They were best at repeating their last behaviour after a shorter time delay suggesting their memory declined with time, and therefore evidence that they were relying on remembering their own actions.

If dogs are aware of their own behaviours then maybe they are aware of their own knowledge. Knowing what you know and don’t know is a form of meta-cognition. Animals may seek more information when they are uncertain about their knowledge, such as the location of their food. Therefore, experiments are set up to provide the animal with an option to find out more information before making decisions when they are uncertain of the correct answer. Dogs were given a choice between two options. They were presented with two identical V shaped fences that had a 2 centimetre gap in them. Rewards such as food or toys were always only placed behind one of the two fences at any one time. Dogs could check by smelling or peering through the gap before deciding which fence to run behind when they were uncertain that the treat or toy was located there. If they made a mistake by choosing the wrong fence, they would not have access to the reward. The researchers wanted to see if dogs were aware of their knowledge about which fence had a reward placed behind it by comparing their checking behaviour. When dogs observed a person hide a reward behind one of the two fences they checked in the gap less often than in conditions when they were not able to observe the hiding process. Instead of always checking before deciding on the fence, dogs were responding to the certainty of their knowledge about the food’s location. It is worth noting that even when they observed the hiding of the reward, dogs sometimes checked as if to make sure they were correct.

Experiment setup for exploring meta-cognition in dogs (Belger and Brauer 2018)

What does the sense of self look like to dogs? It is probably different to how we perceive ourselves. We share cognitive abilities with animals but which ones, how many and how prominent depends on the species. We have our own puzzle of cognitive abilities that fit with the way we live our lives. To some extent dogs know who they are physically in space, from their own scent and may be aware of their own knowledge and can remember their own actions. This is not surprising since they are social animals, primarily sense their environment through smell and communicate and live closely with another species. In the meantime they will continue to amaze us with their abilities, sparking the creation of nifty ways to explore more pieces to a puzzle that will hopefully give us more insight into the minds of our furry friends. 

Seeing the world through their eyes

How do dogs experience their world? We can approach answering this question by breaking it down into the social and physical factors that can impact on a dog’s perspective of their environment and place in it. Dog’s can perceive their environment through smell, vision, sound and touch. They will also have different ‘experiences’ or interactions with their environments. For example, a small dog like a dachshund will probably interact with humans differently than a large dog like a Rottweiler. Small dogs might be used to being picked up by humans whereas I rarely see owners carrying their Rottweilers around. A passer-by may approach and even try to interact with the dachshund but Rottweilers might experience more fearful and staring behaviours from them. All these social interactions with humans may also shape how a dog experiences its world. To be clear, the majority of dogs worldwide do not live in houses with humans and their experience of humans and their environment will also differ. These factors can all add up to help us understand what is important and noticed by dogs in their environments. For example, I look at my chair and to me it is meaningful as an object to sit on. My dog, who is too large to sit on a chair, might see it as an object in which they can beg me for affection. Therefore, the chair has a different meaning or value to me compared to my dog.

We often say that dogs are mainly smelling animals and humans are mostly visual. There is some truth to this, dogs have 10,000 – 100,000 better ability to detect odour than humans and can detect cancer, drugs and even viruses. However, we also forget that they are predators who are built to chase moving prey. Dogs evolved alongside us, a visual creature and have surprised us with their abilities to distinguish their owner from photographs alone and even picking out happy and angry faces from images.

Visual acuity is the ability to detect details in an image and dogs in general have 3 – 8 times worse visual acuity than us. It can be measured behaviourally by comparing the ability of a dog (or human) to see alternating black stripes of different spacing as separate. There are differences between breeds in their visual acuity. Sight hounds such as whippets and greyhounds generally (there are individual differences) can perceive details in the periphery of their visual field since they use their sight to follow moving prey on the horizon. Whereas brachycephalic dogs, those with short skulls such as pugs, can perceive details in a small concentrated part of their visual field. Like us, their eyes are more central facing and they have a high density of  photoreceptors cells in the centre of their retina. 

To understand these cells let’s first look at the structure of the eye (at least in vertebrates). The eye is made up of a lens that refracts light as it enters through the pupil and focuses on the back of the eye, the retina. On the retina you will find specialised cells called photoreceptors that will transform light into electrical impulses that are sent via the optic nerve to the brain for image processing. These cells are called rods and cones, they vary in the amount of light that they require for them to trigger electrical impulses. Rods need very little light, hence they are usually useful in low light conditions whereas cones have a higher threshold that are used mostly in bright light conditions. Cones are the ones you use to see colour. 

The type, proportion and distribution of the rods and cones on the retina can tell you a lot about an animal’s visual perception. More rods to cones would be found in an animal that is active in the early hours of the day whereas more cones means an animal is mostly active in the day. Humans have around 5 % cones in the centre of the retina whereas dogs around 3% suggesting they have a more dim light sensitive vision. Furthermore, dogs only have 2 types of cone cells that allow them to perceive wavelengths of light on the blue and yellow spectrum (similar to red- green colour blindness in humans) whereas humans have three cone types that enable use to discriminate blue, green and red colours. You can have a go at seeing the world through your dog’s vision here.

Representation of human vs dog colour vision: Hirskyj-Douglas et al 2017

Dogs in general have worse eyesight in bright light conditions than us but slightly better than us in low light levels. Why is this? They have adaptations to help with low light levels: large pupils to allow more light to enter the eye to hit the retina, a tapetum lucidum which is a mirror like tissue that sits behind the retina and reflects  light back into the eye to increase the light available for the photoreceptors to pick up on (shared by many other crepuscular and nocturnal vertebrates). You can see this structure when you use a flashlight to find your dog on your nightly walks and you are rewarded with the reflection of two bright eyes staring back at you. Dogs also have many rod cells that respond to low light levels compared to cone cells. This means that during bright light their vision decreases since rod cells can become bleached leading to temporary blindness. Consider the feeling of turning a light on in the middle of the night, you may squint as you adjust to the sudden bright conditions. This happens to other animals as well, think of deer or rabbits becoming dazzled in car headlights, they will freeze since their vision is interrupted. In fact, for these dark-adapted eyes, short wavelengths of light such as blue light are worse than long wavelengths including red light at bleaching rod cells. Maybe something to consider next time you buy a car with xenon headlights (which emits short wavelengths) or halogen headlights.

Ever wondered why your dog isn’t interested in watching TV? Or maybe you have a mutt that watches TV? But do they perceive the images the same way that you do? This depends on two things, the refresh rate of the screen and the critical flicker fusion frequency (CFF) threshold of the observer. Visual processing is limited by the number of stimuli that can be processed at any one time. Light is emitted in pulses every second and above a certain frequency threshold the image can look like it is flickering or still. Different animals have different thresholds for processing flicker frequencies. If the frame rate is below your threshold the image appears to flicker but if it is above, then the image appears still. TV screens usually have a frame rate between 60-100 Hz ( that’s images 60-100 times per second) depending on the technology used. For example LCD screens have a high flicker frequency of up to 200 Hz whereas you might notice flickering in older TVs that use cathode-ray tubes.  Humans have a lower threshold compared to dogs of around 50-60 Hz whereas dogs have 70-80 Hz. It is very likely that images on TV screens appear to flicker to dogs.

Research into dogs’ perception of videos on a screen have been carried out and usually focus on the dog’s behaviour: how long they gaze at a screen and eye tracking. A simple and effective study gave dogs three screens to view different videos on at the same time and noted how long they looked at each. The dogs viewed nothing most of the time which could be because they were over stimulated by all the screens. There was some preference for viewing videos that had dogs and humans in (around 18 seconds gazing time) but their attention span seemed very short since they switched between screens 66 times. Maybe they were distracted by other movements or the images on the screens flickered and did not hold their attention for too long. This study only used 2 dogs so we need to be careful when interpreting the findings, it would be inappropriate to say that these 2 dogs represent all domestic dogs. However, it is an interesting and simple study that you could try out with your mutt at home. And playing back videos with different subject matters is a simple way to see what is meaningful to them based on what they notice.

Apart from the obvious curiosity about how our canine companions experience the world they share with us, why is it important to explore whether dogs can see images on a screen? Displays such as touch screens are often used in studies that look into an animal’s cognition, how they process information in their environments. Researchers will train an animal to interact with a touch screen and then present them with tasks such as distinguishing between different shapes. It is always important to check if your methods are suitable for the animal you are studying, There is some evidence that dogs can use touch screens effectively. Medical alert dogs can be trained to touch icons on a touch screen with their nose or paws to ‘call’ for assistance when asked to by their human. There may also be individual variation in whether the images they view appear to flicker or not. Some humans can see flickering of 60 Hz screens and others can not, and some are affected by headaches and nausea when under fluorescent lights while others are not. Your dog might enjoy watching TV. They just may not be seeing the same show as you.


If you want to see how interested your dog is in watching or interacting with screens you can carry out your own study comparing the number of times and how long they gaze at different videos. Or if you have a tablet and don’t mind it being roughly handled there are plenty of dog and cat game apps you could try out with your furry friend. Leave a comment about your dog’s screen watching habits.


References: 

Barber,A,L,A., Ratcliffe,V.F.,Guo,K.,Wilkinson,A.,Mills,D.S., and Montealegre-Z,F. (2020) Functional Performance of the Visual System in Dogs and Humans: A Comparative Perspective. Comparative Cognition and Behaviour Reviews 15

Byosiere,S.,Chouinard,P.,A.,Howell,T.,J., and Bennett,P.,(2018) What do dogs (Canis familiaris) see? A review of vision in dogs and implications for cognition research. Psychonomic Bulletin and Review 25: 1798-1813

Byrne,C.,Zeagler,C., Freil.L.,Rapoport,A.,and Jackson,M.M.,(2018) Dogs using touchscreens in the home: a case study for the assistance dogs operating emergency notification systems. Proceedings of the Fifth International Conference on Animal-Computer Interaction 12: 1-10

Hirskyj-Douglas.I., Read,J.C.,and Cassidy,B. (2017) A dog centred approach to the analysis of dogs’ interactions with media on TV screens. International Journal of Human-Computer Studies 98:208-220

Lind,O., Milton,I., Andersson,E.,Jensen,P., and Roth,L.S.V. (2017) High visual acuity revealed in dogs. PLoS ONE 12(12)

Create your website with WordPress.com
Get started