EXPERIMENT 1: ATTENTIONAL CAPTURE AND SUSTAINED ATTENTION
In experiment 1 we tested whether dogs’ attentional capture and their sustained attention differed by age in two different contexts, Event 1 comprised of a moving object, and Event 2 a moving human and object. Previous research on attention in monkeys using a touch screen by Baxter and Voytko (1996) determined that attentional capture was preserved in aged rhesus monkeys. Zeamer et al. (2011) compared sustained attention in healthy young and aged rhesus monkeys, using a continuous performance task (individuals were trained to respond to one of three stimuli by touching a screen). Results showed that aged animals made significantly more errors than young animals. This task took many trials to learn before testing could take place. Therefore, for this experiment we simplified the sustained attention test by removing the need for a trained behavioral response to indicate attention. Instead we measured dogs’ attention to two stimuli, as indicated by time spent with the head (used as a proxy for gaze direction) directed toward the stimuli.
Previous studies focusing on measures of attention to novelty in dogs and rats found that exploratory behavior varied significantly with age; with older subjects showing the lowest levels of sustained attention (Soffié et al., 1992; Handa et al., 1996; Siwak et al., 2001; Rosado et al., 2012). Therefore, based on the previous research cited above, we predicted that dogs would show no age differences in attentional capture, and sustained attention to the two stimuli was expected to decline with age.
At the beginning of the experiment, the owners entered the experimental room with their dog on a leash. A hook on the wall next to a window allowed dogs to be tethered in one location. The owners attached their dogs to the 1.5 m leash on the hook, and sat down on a chair facing away from the dog toward the window. They started to fill in a questionnaire on an iPad. Owners were instructed to ignore their dog and the actions of the experimenter, and to be quiet and still. All owners followed guidelines, and did not attempt to interact with their dogs. Two conditions were presented in a counterbalanced order to each dog, Events 1 and 2.
Event 1: After the dog and owner were in position, the experimenter pulled a fishing line, which was attached to a small orange plastic watering can (child’s toy) placed in the center of the experimental room. The line ran through a metal hoop in the ceiling in the testing room, allowing the object to be manipulated by the experimenter from outside the room. The object was moved up and down in front of the dog (but the dog was prevented from approaching it by the leash) for approximately 1 min (Figure ). After this time the experimenter fixed the toy to the ceiling and a tone indicated that the owner and dog should leave the room.
Event 2: After the dog and owner were in position, the experimenter entered the testing room, closed the door, walked to the wall opposite the dog, and proceeded to walk up and down the length of the wall (6 m) pretending to paint the wall with a roller with her back to the dog. The experimenter removed her shoes before the test, and walked as quietly as possible. At no point did the experimenter gain eye contact with the dog. After 1 min the experimenter left the room, and a tone indicated that the owner and dog should leave the room.
We used the latency to orientation [LO; measured from the first detectable movement of the toy/door handle up to the point where the dogs gaze (head and nose) was centered upon the stimulus (toy/door opening/human entering)] as a measure of attentional capture, and the average gaze (AG)-bout duration (total duration looking time divided by frequency of looks), and the percentage of total looking time (PTLT) as measures of sustained attention. Dogs that were already orientated to the stimuli when the stimuli were first presented, were excluded from the LO analysis (Event 1: N = 24, Event 2: N = 13). A randomly chosen set of 20 dogs were double coded independently by two coders, and inter-observer reliability for LO, AG, and PTLT was excellent (r > 0.89, p <0.001 for each variable).
Latency to orientation was inverse-transformed, AG was log-transformed, and PTLT was square-transformed to attain homogeneity of variances, and additionally we fitted a variance structure which allowed for variance to differ between the two conditions (constant variance). Data was analyzed using linear mixed effects models (LMMs; Pinheiro and Bates, 2000) with condition (Event 1 vs. Event 2), age and experiment order (Event 1 first vs. Event 2 first) as fixed effects and dog identity as a random factor. Additionally, the potentially confounding variables sex and neuter status were included as fixed effects. After testing for age effects we then re-ran the model with training score and current training hours as fixed effects and dog identity as a random factor. We included the two-way interaction between condition and age, training score or current training hours respectively to test whether any effects may be restricted to one condition.
To examine whether dogs attentional performance was consistent across different contexts the relationship between PTLT at Event 1 stimulus and PTLT at Event 2 stimulus was calculated, using a Spearman’s rank correlation test.
Dogs’ LO to the stimulus was on average 0.57 s (range = 0.1–3.5 s, SD = 0.38 s). The relationship between age and LO was best described by a quadratic function [LMM, F(1,141) = 4.97, p = 0.01, Figure ]. When using age group as a predictor no significant age differences or interactions were found (p = 0.28). There was no significant difference in LO to Event 1 vs. Event 2 stimuli. The removal of two outliers did not change the results.
Percentage total looking time was significantly higher for Event 2 than for Event 1 (Event 1 = 66.17 ± 22.13; Event 2 = 90.43 ± 10.86; LMM, F(1,140) = 221.01, p <0.001). There was a significant interaction between condition and age in months [LMM, F(1,140) = 5.35, p = 0.02, Figure ]. PTLT decreased with age in Event 1 (Spearman’s rho = -1.98, p = 0.02) but not in Event 2 (Spearman’s rho = 0.042, p = 0.62). When comparing age groups no significant age differences or interactions were found. When three outliers were removed all reported results remained significant.
Average gaze-bout length was longer in Event 2 than in Event 1 [Event 1 = 12.51 ± 11.70; Event 2 = 42.82 ± 30.89; LMM, F(1,141) = 289.03, p < 0.001]. There were no significant effects of age on AG.
With both variables (PTLT and AG), attention paid to Event 1 was significantly positively correlated with attention paid to Event 2 (PTLT: Spearman’s rho = 0.224, p = 0.010; AG: Spearman’s rho = 0.270, p = 0.001). These results remained significant after removing outliers.
Training score and current training hours had no significant effects on any of the variables measured.
When examining dogs’ attentional capture abilities across age a significant quadratic relationship was found. Age differences can possibly be explained by a slight sensory motor decline in the aged dogs, and a heightened sensitivity to sound/movement in the middle aged dogs. However since latencies to orientation did not differ in Event 1 and Event 2 conditions, and response latencies in the senior age group were not significantly different from the other age groups, the observed relationship was minimally effected by age. Regardless of the original orientation of the dog, all dogs very quickly orientated to the stimuli in Event 1 and 2, and we conclude that the physiological condition of the dog minimally affected its ability to orientate its gaze to the stimuli.
Measures of sustained attention were expected to decline with age in both conditions. However, only attention to the Event 1 stimulus showed a significant reduction with age in accordance with our predictions. The novel stimulus and strange movement of the inanimate object generally caused a startle response in the dogs, and an increase in frequency of looks to the stimulus compared to Event 2. The older dogs showed a decrease in overall looking time compared to young dogs, which could be explained by a life-long learning process to reduce reaction to novel external stimuli, such as moving objects (children, cars, bicycles, etc.). Dogs learn to attend selectively, which helps them to focus their attention on relevant stimuli (for example the owner), whilst ignoring irrelevant occurrences (Lindsay, 2001). We found no age effect on attention paid to Event 2, which may be due a ceiling effect (almost half of the dogs paid attention to the stimuli for over 95% of the time). Therefore, the interaction found between age and stimulus type may be an artifact of the ceiling effect. Future studies will need to determine whether sustained attention toward a social type stimulus might also decrease with age, for example by increasing the duration of presentation of the stimulus in Event 2. Here we can conclude that even senior dogs are capable of high levels of sustained attention over 1 min if the stimulus is of high relevance to them.
Percentage total looking time and AG-bout duration was found to be higher in Event 2 (experimenter painting the wall) than in Event 1 (moving plastic watering can). One possible explanation for this difference is that the size of the stimuli caused a bias in attentional allocation. The type of movement (vertical vs. lateral), the distance of the stimuli from the dog, and the novelty of the stimulus could also have influenced the dogs’ attention. Previous studies have indicated that dogs prefer to attend to novel objects over familiar ones (Kaulfuss and Mills, 2008) and also to novel human faces when compared to familiar faces (Racca et al., 2010), therefore we might have expected dogs to attend to Event 1 and 2 similarly. A main difference between the two event situations was that Event 1 contained a non-social stimulus and Event 2 a social stimulus. It seems likely that positive experiences with the experimenter gained in the previous tests of the test battery could have motivated the dogs to attend to her, over the novel non-social object. Horn et al. (2013) found that the nature of past interactions with a human specifies the dogs’ relationship with them, and increases attention to that person. Positive reinforcement during previous training experiences has been found to be highly correlated with levels of attention (Lindsay, 2001). Therefore reinforcement of attention in one situation should improve attending to the same stimulus in different contexts.
In sum, by the age of 6 months, Border collie attentional capture and sustained attentional abilities were already at adult levels, which is comparable to the finding of similar tests in human subjects (Berardi et al., 2001; Michael et al., 2013). Nevertheless, individual differences occurred consistently across the different contexts (i.e., dogs which looked longer at the Event 1 stimulus also looked longer at the Event 2 stimulus) which could be a consequence of an underlying personality trait.
Most of us would amusingly assume our hairy little human’s attention span is measured in seconds. But you’ll laugh when you hear the true length. And how you can push it to peak levels… Look around your own group of human friends and you’ll see that the ability to focus on one task is a trial for some and a joy for others. And dogs are exactly the same. While some geek out on training, others need a feast of treats to get that brain to switch on. But the age where their focus is at peak levels isn’t even close to when most training takes place. In fact, puppy school has been dismissed for years before a dog’s attention span peaks between 3 and 6 years! In fact, it’s almost time for the 5 year school reunion before all the other distractions take a back seat to well…sitting. And staying. It’s thought a more calm, controlled brain wave pattern, along with a ‘been there, done that’ approach to previously distracting sights and sounds plays a huge part in this. So how long is that attention span? Well, while some dogs can be measured at just 2 seconds (relatable), the average sits at 27 seconds to hold one single thought. So what can you do to lengthen it? It’s all about working with their beautiful mind rather than against it. Try your training in quieter environments (yes the dog park is impossible!) and use that period in the 20 minutes after naps where their brain is at it’s most fresh and free of fatigue. And of course, treats do help with the motivation levels. Supplements like L-theanine (from green tea…and found in my “School Snacks” treats) have been shown to improve concentration and focus to keep that brain firing like the doggy dux they really are… Well, most of the time
As in humans, attention is best in young adult to early middle-aged dogs
It is becoming more common for scientists to look for similarities in the psychological processes shared by dogs and humans. This includes similarities in mental abilities (click here for an example), emotional processes (click here for an example), of humans and dogs and even the similarities of their brain mechanisms (click here for an example). Now a new study allows us to compare dogs and people on another fundamental cognitive process, namely attention. The reason why attention is important is that it interacts with so many other mental processes such as consciousness, perception, cognition, working memory and learning. Put simply, you cant perceive something if you are not paying attention to it, and if you fail to perceive something you cant learn anything about it and so forth. Many psychologists have suggested that various age-related changes in the learning and problem-solving abilities in humans have to do with changes in attention over the life span. If there are similar changes in attention processes in dogs it could also help to explain age changes in canine mental abilities.
In a research report accepted for publication in the journal Frontiers in Psychology, a team of researchers headed by Lisa Wallis at the University of Veterinary Medicine Vienna explored how attention changes over a dogs life. The idea was to see if selective attention has the same age-related pattern as is found in humans where the peak performance runs from young adulthood through early middle age (20 to 39 years).
This was a big study involving 145 Border Collies aged 6 months to 14 years. The authors claim that only Border Collies were used in order to eliminate any possible breed related differences, but one wonders what other breed of dogs they might consider given the fact that the testing was done in something called the Clever Dog Lab and research has suggested that the Border Collie may be the most clever dog in dogdom (click here for an example).
The first pair of tests attempted to determine how rapidly dogs of various ages pay attention to objects or humans. The object test involved a childs plastic toy watering can. It was attached to a string which went through a loop fixed to the ceiling. At some point the experimenter (who was out of sight) tugged on the string to cause the plastic object to bounce around. Two measures were taken, first how quickly dog began to pay attention to the object (technically called attentional capture), and secondly how long the dog continued paying attention to the object (technically called sustained attention). The person test involved an individual who was known to the dog. She entered the room and pretended to paint the wall with a paint roller.
If this experiment were conducted using human beings of different ages as test subjects, the expected results would be only small differences in attentional capture but large differences in sustained attention. That is exactly the pattern that was found here for the dogs. While age did not affect how quickly the dogs first looked at the object or the person, it did make a difference in how long the dogs continued to look at the target. In general the more senior dogs did not sustain their attention as long as the younger ones. Wallis summarizes her findings by saying “We found that older dogs – like older human beings – demonstrated a certain calmness. They were less affected by new items in the environment and thus showed less interest than younger dogs.” However what the dogs were asked to pay attention did make a difference. Regardless of the dogs age, the experimenters found that more attention was directed to the human than to the erratically moving object.
The second experiment involved selective attention. Once again a very simple test was used, but here one that required the dog to shift attention from one task to another. This test involved the experimenter standing in front of the dog and tossing a bit of food either to her right or left. Obviously a dog will pay attention to the food and move away from the person to get it. Now the experimenter waited until the dog shifted its attention and made eye contact with her. Once eye contact was made a clicker sounded and a piece of food was tossed off to the side shifting the dogs attention away from the experimenters face and onto the treat itself. The data involved measures of the amount of time it took for the dog snatch the bit of food and then look up into the experimenters face again. Under these test conditions the performance of the dogs varied with age and the pattern mimics the findings in humans. It was the young adult to middle-aged dogs (3 to 6 years for canines) who reacted the most rapidly.
While it is understandable that the older dogs might have slower mental processing and therefore might have more difficulty shifting their attention, that leaves us with the question as to why are the younger dogs not reacting more quickly? According to the experimenters one reason is probably their general lack of experience. A second answer is that dogs like humans go through a difficult phase during their adolescence (which is 1 to 2 years in dogs), and this is partly due to hormonal changes in their bodies.
There is a quirk in the data here and that has to do with the fact that in the selective attention task there is something to be learned, namely that the optimal strategy is to pay attention to the food and quickly grab it, then to rapidly shift attention to the experimenters face to trigger a click and another piece of food. Over the 20 trials all of the dogs, from six months to 14 years, showed an ability to learn to do this— and that includes the geriatric dogs. However when the researchers looked at the rate of learning, the fastest improvements were for the adolescent dogs. Wallis sums this up saying “Thus, dogs in puberty have great potential for learning and therefore trainability.” This is similar to what anyone who has ever lived with young human teenagers has probably experienced. They are capable of learning quickly and efficiently if only you can get their attention — which is not always easy. This also appears to be the case in dogs.
Copyright SC Psychological Enterprises Ltd. May not be reprinted or reposted without permission
Data from: Lisa J.Wallis, Friederike Range, Corsin A. Müller, Samuel Serisier, Ludwig Huber, and Zsófia Virányi (2014). Lifespan development of attentiveness in domestic dogs: drawing parallels with humans. Frontiers in Psychology, 5, doi: 10.3389/fpsyg.2014.00071
Get the help you need from a therapist near you–a FREE service from Psychology Today. City or Zip
