The objective of instrumental learning (operant conditioning) is not to learn new behaviours. Learning new behaviours is a by-product of the process. The sole objective of instrumental learning, in our view, is to teach the dog that it has control over its circumstances and that certain behaviours on its part will bring about corresponding consequences. While the pathway (instrumental learning) leading to the dog’s preferred consequences can be laid open for the dog, a critical presumption remained – that the dog is interested in the betterment of its circumstances, though its own effort.

In Chapter 2 and 3, we covered the mesolimbic pathway where neural signals trigger the release of dopamine, causing arousal and influences behaviour to seek out pleasurable activity (for example, hunting to eat). Dopamine binds to dopaminergic receptors present in the nucleus accumbens and prefrontal cortex. Increased activity in the projections to the nucleus accumbens play a major role in reinforcement, which manifests as a sense of accomplishment / pride / general happiness, leading the dog to intensify its effort in the same manner as it had done so earlier. We also covered how activity of the medial prefrontal cortex (PFC), which promotes the detection of control, leads to the automatic inhibition of the dorsal raphe nucleus (DRN). Learning control is largely a process of instrumental learning and dogs learn to act in accordance with the outcomes produced by their actions. Through instrumental learning, alterations in the ventromedial PFC circuitry at an early agelearning how to switch off stressors aka escapable stress – creates a specific and persistent change in the prelimbic-dorsal raphe nucleus circuit that led to the inhibition of the dorsal raphe nucleus and prevented passivity in response to even inescapable stress.

The vast majority of dog trainers assumed that the dog is inherently keen to improve its existing circumstances; that it is not passive by default. These trainers will focus exclusively on creating the neural signals (marker-based training / clicker-based training), rushing head long into training new behaviours without checking back to see if the dog’s inherited genotypic and phenotypic traits are congruent with the training methods applied. They will champion the use of positive reinforcement methods loudly and proudly, both in person and virtually in the digital space. They “spit” on those who differ from their philosophy and brandish the big spear of “animal welfare” to wage war on those who understood better and deeper. True instrumental learning encompasses pre-eminently the conditioning of the ventromedial PFC circuitry leading to the suppression of the DRN and helplessness as the default. We want a dog that keeps trying its luck, learning to overcome stressors for a win starting from a young age. The creation of neural signals and spurring the dopaminergic binding in the PFC and nucleus accumbens is but secondary order. This is what we meant by building engagement and drive in the dog. At Happy Paw Ark, rushing into behaviour training is spurned. We are more interested in getting the dog to detect control, seize control and act to improve its circumstances whenever opportunity presented itself. We the humans, control the opportunities through disciplined antecedent arrangement – setting events and conditioned motivating operations – and consistent delivery of consequences.

Let us illustrate this significant topic with a Compare & Contrast table:


At Happy Paw Ark

Typical Pet Owner

Why We Do What We Do?


Food is given whenever the dog does something, e.g looks at trainer in the eye and stay still; dog ignores the bowl and walks over to the trainer Food appears at fixed timings and in a bowl, regardless of hunger level in dog Dog learns that it needs to behave in a certain way to make food appear. As the dog learns an expanding array of behaviours to make food appear, it realises the control it has over the appearance of food


Dog decides where to go; trainer is passive and acts only in the interest of the dog’s safety e.g to steer clear of glass, vehicles, cat poop etc.

Trainer is interested in the emotion state of the dog and has food ready to establish association of typically scary objects with appearance of food (counter-conditioning). Food can also be used to reinforce good behaviours such as walking away from food waste left on the ground, jumping up the park bench (overcome obstacle to get somewhere); wagging tail at a passing cyclist etc.

The walk can last 30 minutes but covers 300m. Trainer is not interested; the dog decides.

Human decides where to go; pulls the dog along or chicken-dances to entice the dog to follow. Wished that the dog is following nicely and not poking its nose into everywhere everytime.

Human is interested in arriving at a certain destination so the dog can mingle e.g dog run.

The walk has a fixed distance/destination/duration.

Dog is acting on its own accord, to benefit itself through its own sensory detections.

Dog has control over its activity.

Dog covers a little distance but comes home exhausted because of the spontaneous learning activities involved.

Playing (predation)

Different objects are used to entice the dog into a chase, grab-bite, kill-bite.

Objects are left for dead once dog bites on. The next object comes alive with trainer animating the object.

Objects are thrown into corners or confined spaces for dog to retrieve; allow dog to enter and retrieve and feel relieved plus accomplished.

No competition from the trainer; dog learns sense of accomplishment by intensifying its behaviours of chasing, biting at object and releasing the dead object.

Owners like their dogs to play fetch because it makes them looks in control and their dog is working “for” them.

Mindless and never-ending throwing of ball / kong / rope / frisbee.

Lack of consideration of purpose of play in dog’s life and the different stages of the predation sequence.

Owners compete with their dogs over the possession of the prey object – tug of war – as a game.

Dog is acting on its own instincts and refining those instincts. Trainer shapes the dog by giving different consequences to different expressions of instincts.

If the dog bites the ball and chews, lots of pets and rubs.

If the dog play-bites trainer’s hands, the play session is ended prematurely. Dog loses access to wanted consequences.

Consequence-based learning in animals gained traction in the 1960s after B.F Skinner published his Theory of Learning – when A, if B, then C; where A is Antecedent, B is the behaviour of the learner / animal and C is the consequence that corresponds to the behaviour executed by the learner / animal. Skinner also introduced the concept of schedule of reinforcement – a “time-table” stating the various frequency of reinforcing behaviours. In some cases, a behaviour might be reinforced every time it occurs. Sometimes, a behaviour might not be reinforced at all. Other times a behaviour is reinforced every third time it is performed.

Translating the ABCs of the Theory of Learning into the ubiquitous learning quadrants, we have Positive Reinforcement, Negative Reinforcement, Negative Punishment and Positive Punishment:





When all criteria of the Antecedent have been met,

Positive (Addition, +)

Negative (Subtraction, -)

Reinforcement (Intensifies)

Consequences wanted by the dog is given to the dog after a wanted behaviour is performed by the dog; dog learns that a particular variant of a behaviour brings about wanted consequences

Consequences unwanted by the dog is removed after the dog performed a wanted behaviour; dog learns that a certain behaviour switches off pressure/aversive

Punishment (Diminishes)

Consequences unwanted by the dog is given to the dog after an unwanted behaviour is performed by the dog; dog learns that a certain behaviour under certain circumstances brings about bad stuff

Consequences wanted by the dog is removed after the dog performed an unwanted behaviour; dog learns that access to the good stuff is lost when certain behaviour is performed under certain circumstances

Instinctively, it appears to be unquestionable that dog training should reside in the Positive Reinforcement (R+) quadrant only. Those trainers who espoused such ideological fantasy could not be farther from the truth. We need to examine the premises of R+ before we go further into R-, P+ and P-.

For R+ to work, there are several precursors. For a worker to be motivated to put in his normal effort, the paycheck (wanted consequence) shall be of a sufficient amount and this amount is determined by the worker, not the employer. A new migrant might be willing to put in the same effort for less compared to a rich man’s child who led a life of comfort since young. Now if the new migrant struck lottery and won a million dollars overnight, would you expect him/her to put in the same effort the next day at work? Therefore, the first criterion for R+ to be effective would be that the dog craves its “paycheck”.

If a new migrant worker craves his/her paycheck at the end of the month, and the employer dishes out a difficult task in the middle of the month, would there be stress experienced by the new migrant worker since he/she would think that this task must be completed otherwise I might lose my paycheck. On the contrary, the rich man’s child might be more lackadaisical as the paycheck at the end of the month carries very little significance. Therefore, the second criterion for R+ to be effective would be either (i) the dog is in the position of a new migrant or (ii) the paycheck is significantly enticing even if the dog is a “rich man’s child”. It would be best if both are met. Nonetheless, it must be pointed out that effective R+ involves stress. This stress emanates from the dog’s desire for the “paycheck”. If there is no desire, there is no stress but there is also no behaviour / learning because the paycheck is insignificant to the dog. The dog learns to turn off the stress by behaving correctly and quickly to reach the eventual paycheck.

How does the dog learn which behaviours turn off stress and get the paycheck? Some trainers teach by dangling the “paycheck” in front of the dog and lead the dog to perform the required behaviours i.e Luring. Other trainers prefer to let the dog knows the “paycheck” is right there in the pocket but a certain behaviour or any variation of it is required to unlock the “paycheck” from the pocket i.e differential reinforcement of successive approximation (DR+). Luring is less stressful for the dog because it can see the “paycheck” clearly and smell it and just needs to follow it. The dog learns very little other than to follow the lure. DR+ is more stressful for the dog but the dog volunteers the behaviours out of its own volition in exchange for the paycheck. Those behaviours that unlocked the pocket will be well-remembered and re-enacted with ever increasing intensity and frequency.

The key point here is that even in R+, there is stress because the dog wants the paycheck. Dog behaves in a certain way to get the paycheck and stress is reduced / removed by its own actions. There isn’t a truly fear free or stress free learning/training afterall, contrary to what the proponents of fear-free dog training are claiming.

What about R-? Traditionally R- involves performing an unknown behaviour to turn off pressure or simply not doing anything to avoid pressure (avoidance). This results in typically subdued states of mind in dogs.

The dog is suppressed and when R- is prolonged, typically slips into learned helplessness and depression because the learner (i.e. dog) chances upon the correct behaviour by luck. In Traditional R-, there is little to no sense of control as the dog falters and try again in a repetitive loop. However, in the modern R- application, the dog has already learnt the correct behaviour through R+ and the dog comes to the realisation that when it performs the correct behaviour it has learnt prior, the taught behaviour not only switches off pressure but also brings about wanted consequences. You will get eager dogs giving their heart and soul in performing behaviours. See

Those waving the banner of zero use of punishment are literally barking up the wrong tree! What they are fighting against are these compulsive trainers who tend to create subdued and helpless dogs through their training. Combined with luring, we get a dog that is dependent on handler’s cues to perform because any form of initiative will likely earn the dog an aversive stimulus. There is a place for punitive measures in dog training.

Positive punishment or negative punishment serve to diminished unwanted behaviours such as biting, growling, outright aggression involving predation or not. When one is dealing with instincts, one can only influence it through redirection or punishment. When a human sense a fist flying into his/her face, the human will blink and duck. Instincts! There is no way to get rid of instinctive responses for these are innate and not learned through instrumental learning. So we can only hope to control these innate responses through thousands if not millions of repetitions in classical conditioning or positive punishment. Like positive reinforcement, the consequences delivered in P+ must be sufficiently punitive else there is really no point in punishing the dog. So typically, when trainers need to deter the future occurrence of bites on friendlies, the punishment must be abrupt and disastrous for the dog. Usually we will “arrest” the dog and pin it down forcefully until we feel a softening of its body. This denotes a mental state change in the dog and then we allow it to perform other simple behaviours and reward it if done correctly. For those that claim punishments are not necessary in education/training/learning, they obviously have not come across sufficient genres of dogs to be able to call themselves dog trainers. These people exist in a bubble consisting of dogs damaged by artificial selection. These creatures are not truly the dogs that enjoined human communities millennials ago.

Conceptual Tools – Respondent Conditioning and Learned Helplessness

What if my puppy is beyond the sensitive period and already a juvenile? In this chapter, we share with you the learning concepts in play when we train dogs of all ages.

There are two main branches – Respondent Conditioning and Instrumental (Operant) Learning. These are not two different schools of thought, and neither can each of these be applied singularly. Conversely, dog trainers will do well to integrate these concepts in the overall training plan for their clients’ dogs.

Let’s start with Pavlov and Respondent Conditioning. Contrary to popular belief, Pavlov did not set out to learn about the dogs’ ability to associate a sound with the arrival of food. He wanted to test if different kinds of food will elicit salivation of different intensities in dogs. Through serendipity, Pavlov discovered that the dogs will start salivating as soon as the bell connected to the laboratory’s door rang, predicting the appearance of a human and therefore presumably food. The final equation for Respondent Conditioning therefore, is:


              To be more complete, there are three steps in total:


CS + UCS = UCR (repeat)


An unconditioned stimulus (UCS) elicits an instinctive (respondent) response (UCR). A hungry dog salivates when it smells the aroma emanating from the kitchen. A tap on the knee results in an involuntary jerk on one’s lower leg. Adding a conditioned stimulus aka neural signal (CS) consistently and repeatedly before the unconditioned stimulus continues to give rise to the instinctive response. Ringing a bell brings the waitress to the kitchen portlet to retrieve the next serving of food for the restaurant’s customers. Repeating the same routine a hundred times and then ringing the 101st bell will very likely result in the waitress’ appearance at the portlet even though the restaurant has closed for the evening. When the conditioned stimulus, an artificially added signal, repeatedly appears before the unconditioned stimulus, the former becomes the predictor of the latter and the instinctive response is elicited regardless if the latter is present or not. Once we have the final equation, we can state definitively that respondent conditioning has taken place. Note the choice of words though, as the whole process involves a conditioned reflex. Hardly any learning has taken place.

What Pavlov discovered by chance has multi-faceted implications for us in conditioning our dogs to life in Singapore’s urbanised environment. Habituation for instance, involved repeated exposure to the UCS at fixed intensity over a long period, to allow the dog to realise that its UCR will not remove / reduce the occurrence / intensity of the UCS. This is one of the methods to achieve learned helplessness, a concept that we need to explain clearly at this juncture.

Learned helplessness is a grossly maligned dog training concept. The diehard fans of positive-reinforcement-only training talked about learned helplessness like a racist swear word. We would like to remain objective and not apply labels to scientific concepts. Everyone has a certain degree of learned helplessness. Most humans for instance, do not wake up lamenting why the sun rises at about 7a.m every morning in Singapore. We learn to accept that as a fact of life and move on, adapting and changing our life circumstances without actively seeking out a way to destroy the sun. Night shift workers for instance, learnt to put up curtains or blinds to darken their rooms so that they could rest while the rest of us awaken for the day’s work. These night shift workers have learnt that they are helpless when it comes to the sun rising up at 7a.m but they are not helpless when it comes to re-designing their rooms. These workers are operant (operate on their environment to benefit themselves) and channelled their resources to the problem-solving aspect rather than to sit around and suffer the detrimental effects of sunlight when sleeping.

In a seminal article published in 2016 by Maier and Seligman, the duo posited that helplessness is not learned; it is the default response to prolonged aversive events, and it is mediated by the serotonergic activity of the dorsal raphe nucleus (DRN), which in turn inhibits escape responses.[1] What was missing in the shuttle box experiments conducted in the early-1960s on dogs was the perception of control – the dogs involved in the experiments were not given a chance to learn control, with the activity of the medial prefrontal cortex (PFC), which promotes the detection of control leading to the automatic inhibition of the DRN. Learning control is largely a process of instrumental learning and dogs learn to act in accordance with the outcomes produced by their actions. Through instrumental learning, alterations in the ventromedial PFC circuitry at an early agelearning how to switch off stressors aka escapable stress – creates a specific and persistent change in the prelimbic-dorsal raphe nucleus circuit that led to the inhibition of the dorsal raphe nucleus and prevented passivity in response to even inescapable stress.

Therefore, it is not that Learned Helpless is a detrimental state of being that should be avoided but the lack of instrumental learning at the sensitive period and beyond that leads to a state of passivity in dogs as default. These dogs are unable to escape stressors or perceive the possibility of control because they have always been in a passive position in relation to their environment. This state of passivity is induced by their human owners and such circumstances are prevalent among the pet owners, unlike the handlers in the working dog world where instrumental learning is a requirement for the dogs to attain their proficiencies. A very common exercise we do is to allow dogs on walks to choose the direction of the walk, using their keen sense of olfaction. Yet the most common training objective we hear from our pet dog client is “I want my dog to follow me for the walk and not pull me in all directions.” For us, the dog’s ability to walk politely alongside the human handler is to lead to the eventual arrival at an off-leash dog run where the dog can run free, following its senses. Our purpose of loose-leash walking training is not to make the owner appear in absolute control of the dog.

Going back to the concept of habituation where the dog realises that its UCR will not remove / reduce the occurrence / intensity of the UCS. Yes, the dog has reached a state of learned helplessness in relation to the aversive stimulus but is the dog able to direct its energy to alternative measures to “solve the problem” rather than sit around and suffer in silence, just like what the night shift workers would have done when they discovered that they could never destroy the sun. We had a client whose dog would panic whenever he started his car engine in the garage. The dog would rush to its bed, located about three metres away from the garage door, sat on the bed and started barking at the garage door. The circumstances worsened eventually as the dog started to “act up” whenever the client walked down the stairs, regardless if he was going to start his car or not. The dog also generalised its hatred of engine sounds to all vehicles with running engines, in motion or parked, by barking at the vehicles it encountered on walks.

Clearly the visual image of the human walking down the stairs (Cs) is a predictor of the eventuality of a loud roaring engine (UCS) and the UCR was to rush to the haven (bed) and barked at the engine sound until the vehicle pulled out of the garage. While the UCR was initially an instinctive response out of fear of the sudden loud engine noise (client drove a Nissan GTR with a 3.5 litres V8 engine), the dog would have learnt that barking at the sound appeared to have remove it. This is instrumental learning at play and the evidence was the generalisation of the barking behaviour to all vehicles encountered outside of its home. There are many solutions to this quagmire but the client insisted that he would not change his vehicle nor park it outdoors. The conclusion therefore, was to “train” his dog to accept the engine noise. There are multiple concepts involved in the behaviour modification plan but for the purpose of the current discussion, we shall examine the habituation portion only.

We got the client to record the sound of his car being started while the dog was out of the house. He was to play this recording using his home theatre system / IP camera, depending on whether he is out at work or at home, at regular intervals – once every four hours – daily, when his dog is awake and minding its own business. He shall not play it if his dog is asleep, even if it is time to play the recording. The recording shall only be switched off immediately after the dog paused its barking. The client struggled with the impeccable timing required to switch off the recording. Over time though, he caught the rhythm and was able to accurately tell when his dog will pause to take a breather. Three weeks in, the duration between the first bark and the first pause shortened from three minutes to one minute and forty seconds. We then decreased the frequency of playing the recording to once every five hours. The duration shortened again and we decreased the frequency in response. After 8 weeks of hard work, the dog merely runs to its bed and chewed its Kong. There were other measures at play too and we shall discuss these now.

The habituation process described earlier was executed concurrently with the Classical Extinction protocol. The client’s movement down the stairs cannot be the trigger for his dog to run to the bed and start barking because that would result in a severely stressed dog, or our client needs to stop inhabiting the upper floors of his home. We found out that our client is a car lover and much of his recreational time was spent tinkering with this beloved car. Therefore, he has a high frequency of walking down the stairs and entering the garage immediately, even on weekends. Classical extinction is a process of decoupling the CS and UCS. Our client had to consciously walk down the stairs and move to other places on the ground floor to break the expectation of engine noise in his dog’s mind. Should he need to move to the garage after walking down the stairs, our client first had to move to the kitchen or garden, attract the attention of his dog and call it over for a treat before moving to the garage. He was also not permitted to ignite his car engine unless his trip to the garage coincided with the pre-planned timing to play the sound recording. There was an additional condition – if our client had to ignite his engine and that it was the right timing to do so, he had to ignite the engine and then get ready to present a treat to his dog at the first pause. A chew treat that takes some time to ingest is preferred. What he is doing is Classical Counterconditioning where the CS (walking down the stairs) predicts the appearance of something pleasurable (treats) and then the UCS of the engine noise predicting the appearance of something pleasurable again (chew treats). There is an element of Differential Reinforcement of Incompatible Behaviours as well because our client is specifically delivering a pleasurable outcome the moment the dog stopped barking and it is unlikely for a dog to bark and chew its treat at the same time.

We also advised the client to move the dog’s haven to a farther corner of the house, away from the garage door. This allowed us to start tinkering with the volume of the engine noise recording. We started playing the recording at 10% of actual volume and gradually increased it over the weeks. This is known as Systematic De-sensitisation and the exposures to the noise recording were gradually intensified. Over time, the chew treat was replaced with a chew toy with some Kong mousse spread over it. The dog expressed its hatred of the engine noise by chewing the toy vehemently and getting some pleasurable outcomes at the same time. The dog also formed the habit of running off to the owner whenever he came down the stairs and waited at the kitchen counter or the entrance to the garden for a treat or two. As the habit solidify, we advised our client to be more random at dispensing treats and add in demands for certain behaviours such as roll over or play dead before treats are given. Here, the element of instrumental learning is incorporated. The dog learnt to expect control and the levers of control are in fact the alternate or incompatible behaviours vis-à-vis barking. The entire behaviour change plan would not be possible if we utilised either Respondent Conditioning or Instrumental Learning; we needed both in an integrated way.

Lastly, we would like to explain the concept of Flooding, which was not applied in the case study delineated in preceding paragraphs. Flooding is something Cesar Millan did in his past videos on The Dog Whisperer where he grabbed a small dog firmly while the female owner presented a grooming scissors to trim the facial fur. We would like to watch him do it to a Rottweiler or Tibetan Mastiff. Flooding therefore, is presenting the CS or UCS at full intensity or extreme intensity until the UCR terminates. If the dog is afraid of engine noises, lock the dog in the garage, play the engine noise recording at the maximum and tether the dog to a fixture in the garage until it stopped barking. This was what Maier and Seligman meant by objective helplessness – animal is objectively helpless with respect to an important outcome (O) such as shock offset if the probability of (O), given a response (R) is not different from the probability of (O) given the absence of that response (notR). When this is true of all responses, objective helplessness exists. Such an animal is in a state of depression.

We will cover instrumental learning in Chapter 4, together with the building of drive and engagement in our dogs. Stay tuned.


[1] Maier & Seligman, Learned Helplessness at Fifty: Insights from Neuroscience, Psychol Rev. 2016 July; 123(4): 349–367.

How to Raise Resilient Puppies – Environmental Socialisation

In the previous chapter, we identified the sensitive period for a puppy to be six to nine weeks of age and that puppies should not be removed from their mothers and littermates before 12 to 14 weeks of age. From the third week on, as the puppies gain cognition of their environment through their senses – visual, auditory, olfaction and tactile – socialisation to the environment shall begin.

The environment consists of all things inanimate and animate, from the puppies’ perspective. With a significantly larger proportion of their frontal cortex dedicated to olfaction, the way puppies and their adult versions perceive the world is very different compared to primates, including humans. When humans take a walk at a neighbourhood park, we take in the sights, then sounds and lastly smells. We will see the blooming flowers before we smell them. Even if we could smell the flowers from afar, our cognitive bias is to search for the blossoms visually until we locate them. For puppies, instead of saying “Let’s go for a walk”, we should perhaps tell ourselves “Let’s go for a sniff”.

The second principle we stick to when we start socialising our puppies to the environment is the concept of control – we want the puppy to be in control of when, where and how they interact with the inanimate and animate things in their environment. We are not leading the puppy, coaxing it to go near a vending machine or garbage trolley found along our walking route. We let the puppy take the lead instead and we follow behind, always ready to prevent mishaps or to encourage the puppy to venture further. In this regard, the leash and collar are merely safety tools, very much like a seatbelt in an automobile. They are not your joystick to manoeuvre the puppy to where you want it to be. If the puppy takes a whiff of the vending machine and decides to move back, we make space by stepping back further to allow the puppy to withdraw. We will not move away nor make cooing noises to “comfort” the puppy. We hold our ground and keep a loose leash, give the puppy space to make its own decision to venture or withdraw.

The third principle we live by, is to determine the puppy’s reinforcers as soon as possible during socialisation sessions. We will always be ready with the puppy’s meal and some high value treats. These are usually the primary reinforcers that kick off the mesolimbic pathway in the puppies; we want them to associate typically “scary” things such as zooming bicycles or noisy construction vehicles with pleasurable situations. We will also be their surrogate mother by supplying the tactile stimuli to comfort the puppies after exposure to environmental stressors – to suppress the puppy’s hypothalamic-pituitary-adrenal (HPA) axis responsiveness and trigger the hormonal homeostasis seeded by their mother as early as the neo-natal period. That said, even before the puppies opened their eyes, we would have subjected the puppies to mild levels of stress. A common method involved removing the puppies and handling them gently, placing them half a body length away from the rest of the litter and let the puppy crawl back to the litter and its mother (transition period onwards). Another method involved switching the puppies’ suckling positions in relation to the mother’s nipples. The puppies will “struggle” among themselves and achieve an equilibrium eventually. The mother in all cases, will start grooming and licking the “stressed” puppy once re-connection has been made. A stress-free puppyhood will not necessarily produce stable and operant adult dogs. In our experience, teaching the puppies to accept and overcome stress to improve their own situations produces confident, stable and operant dogs in the future.

When we say environmental socialisation, it does not mean we pluck the puppies out from the litter, away from their mother, and expose them to the world immediately. There are many preparatory exercises we conduct in a sanitised environment which we call the Operant Chamber or Operant Field, to prepare the puppies. Snuffle mats are great for nosework in juvenile or adult dogs; they are excellent tools for puppies to “struggle” and find some food. We start this game when puppies are transiting to solid food, at 6 weeks of age and beyond. As the puppies gain expertise, we will increase difficulty level by placing the snuffle mat in a wooden tray. The tray will be replaced by a box eventually and puppies would have to climb and claw their way in to get the food.

Another game we play is “focus”. We teach the puppies to target to our face and palms. Before we play focus, neural signals that indicate the appearance of pleasurable situations would have been created. Some trainers prefer the tin clickers, some prefer verbal cues that are often monosyllabic. Through respondent learning, these unique auditory cues predict the onset of feeding and stimulate the release of dopamine, causing arousal and influences behaviour to seek out pleasurable activity (hunting to eat). Dopamine binds to dopaminergic receptors present in the nucleus accumbens and prefrontal cortex. Increased activity in the projections to the nucleus accumbens play a major role in reinforcement.

Once the puppy understood the unique auditory cues, we will let the puppy roam freely in a small room with a passive human seated in the middle, on the floor. The puppy will inevitably come over to investigate the passive human. The human will make the recognised auditory cue once the puppy makes eye contact with the human and provide a small portion of food as a form of reinforcement. The game is played during mealtime when we know the puppy is hungry and eager to eat. There can be variations to this game where the puppy makes muzzle contact with a closed fist or open palm instead of eye contact with the passive human. We prefer the eye contact version as it forms the foundation for a good loose leash walk later in basic obedience.

Another game we like is “chase”. The same set up in the operant chamber is required. A passive human shows the puppy a food kibble and fling it out to perhaps a metre distance. Let the puppy chase the rolling kibble. When the puppy caught the first one, make kissy noises to get the puppy back to the human, then fling a second kibble. Repeat this with some hand feeding interspersed between repetitions. The amount of food the puppy gets from the hand must be more than what it gets from each repetition of chasing. As the puppy improves, we usually combine “focus” and “chase” together in one session; add in the recall signal to replace kissy noises. We will also migrate the games to an outdoor field that has a segment fenced up to control the movement of the puppies. In the fenced area, we will throw in some wobble boards, small size yoga balls, an empty box, and some irregular shape Kong toys. Food is hidden in the majority of these items for the puppies to scavenge. Humans will also join in and some of them can be passively sitting around while the others play a more active role as the surrogate mother delivering tactile stimuli after every “struggle” to get food.

One of the key socialisation components we do with the puppies is the introduction to human touch. We do this in a deliberate manner, preparing the puppies for the most intrusive and obnoxious of our specie – the human toddler / child – the little creature who is small in stature but very unpredictable, loud and lacks logic. These three traits evoke fear in puppies quite effectively. Before introducing real toddlers or children to the chamber, adult trainers will mimic them and expose the puppies to the degraded versions of what the little ones are capable of. An adult can squat or kneel down and as the puppy approaches, sticks out his/her hand suddenly and stroke the puppy’s head and ears. Some food kibbles or treats are given to the puppy immediately after the stroking action. The active human will also make loud screeches and the puppies will get a rude surprise. A second human will play surrogate mother once again and deliver tactile stimuli to calm the puppy, then flick some food kibbles to start a short game of “chase”. The exercise repeats and the picture of an unpredictable, loud and irrational young human is introduced to the puppy frame by frame, jigsaw piece by jigsaw piece, before going outdoors to interact with role-players and then real passer-by in the neighbourhood.

When introducing puppies to other animals, we take special care to select good role models so that the correct behaviours will be imprinted. Typical puppy classes conducted in Singapore involved introducing puppies to puppies, standing around, having a chat about each other puppies and laughing at the puppies’ antics. This is what we call the blind leading the blind. First of all, the puppies would have an authoritative figure – their mother – in their lives if they were not removed from the litter prematurely. At around eight weeks of age onwards, we will introduce several stable and confident young adults – two or three at a time – to the puppies in deliberate socialisation sessions. The idea is to let the puppies learn how to interact with the other dogs by mimicking the behaviours of the stable and confident young adults who are known to be able to interact well with one another. For short spurts of time, the young adults may mingle with the puppies and these adults will teach the puppies important behaviours such as recognising an invitation to play, knowing when to back off (adult dog will usually growl and chase puppies away) and bite inhibition.

We do not believe in introducing cats or smaller house pets to the puppies because we will eventually need to harness the puppies’ prey drive for training. Some trainers such as Grisha Stewart proposed introducing the cat or other flight-prone house pets to the puppy in a way these animals would not run away.[1] It was suggested to reinforce the puppy whenever it turned away from the cat or parrot. We prefer another method of “learned helplessness” when dealing with the issue of other house pets. This concept has gotten a bad reputation from the Positive Reinforcement purists’ propaganda machinery, but it is a grossly misunderstood concept; a degree of learned helplessness is good for the puppy. We will talk about Learned Helplessness in another chapter.

The other big component of the puppy’s environment is surfaces. It will surprise you that puppies born and bred in apartments in Singapore typically do not dare to step on turf when they get their first walk outdoors. We utilise artificial turf for approximation purposes and puppies start to have a variety of surfaces – turf, carpet, tiles, yoga mat etc – to walk on and roll over with indoors. We also incorporate noises into surfaces. In Singapore, we have drain covers that are metallic and grated. We start teaching the puppies to jump over grated drain covers as soon as they are up on all fours and prancing around. Metallic scrapping sounds are layered when they get confident to habituate the puppies to the auditory stimulus. Other surfaces to get accustomed to are sand and sponge mats (playgrounds). These are given to them outdoors and we find that a confident puppy inside will usually be more willing to take on challenges outside.

Sounds, scents and sights are significant variables in Singapore’s condensed environment. Our buildings are tall and closely located together; sound gets reflected and refracted in ways human ears cannot detect. Puppies and dogs as a specie can hear all these and the auditory stimuli can be very harassing. As early as the age of three weeks, we will start playing white noise at low volumes in the nest, interspersed with construction noise, traffic noise, popular and classical music. At about three weeks old, we will start putting blown up photos of metropolitan areas, bicycles, garbage bins etc around the nest as they can learn to recognise novel objects they have viewed previously and will subsequently exhibit less fear of the actual objects when tested at seven to eight weeks old. We will also ask for garments of the future puppy owners as well as trainers to throw it into a designated play pen for the puppies to be imprinted with the scents. As the puppies matures, we will start bringing them out to the quieter streets to be exposed to the buildings, scents along the pedestrian path (usually garbage) and infrequent vehicle, subway train or construction noises. They will be fed while being exposed to these stimuli, creating a pleasant association. As they grow older, play sessions will be conducted in suitable grass areas to cement the pleasant association. The level of stimuli will be increased gradually by careful selection of location and time of exposure.

As the reliability of their behaviours improves in the “focus” and “chase” games, we will start using these foundational behaviours to teach the puppies to focus on their handlers and return when called. Combined with tactile stimulus, these behaviours allow the puppies to switch off when potentially aversive situations arise. Going to a public dog run is a deliberate and cautious decision, bringing the most confident puppies in the litter. At the dog run, we observe the puppy’s interaction with other dogs keenly. If it shows any signs of apprehension, the recall is initiated followed by a few repetitions of the focus game. Locomotion may be incorporated in the focus game to get the puppy further away from the source of apprehension. When the puppy is visibly calmer, we let the puppy free roam again and explore its environment. If another dog is “ill-mannered” and unable to read the puppy’s cut-off signals, the human handler will distract the ill-mannered dog instead and leave the puppy with his/her teammate.

We only do public dog run sessions after our puppies have been taught properly by the stable young adults in our premises. In the same vein, the “focus” and “chase” behaviours are utilised on walks on public pathways and by around 12 to 14 weeks of age, our puppies are well-accustomed to the picture of an approaching human with hand sticking out to pat its head and rub its muzzle. We will still honour the puppy’s desire for space should it move backwards from an approaching human. An alternate route is taken, and the puppy is allowed to sniff and explore before “focus” is initiated for a short distance.

We have given a brief overview of what we do for puppy socialisation. If you have any queries, please feel free to email us at [email protected] or visit us at and submit an enquiry using the Contact Us form.


[1] Stewart, Grisha, Behavior Adjustment Training 2.0, Dogwise Publishing, 2016, p187.

How to Raise Resilient Puppies – An Ontogeny Perspective

In the last 10-15 years, the emerging field of behavioural epigenetics has revolutionised our understanding of the ways in which genes and experience – nature and nurture – interact during development to produce effects on one’s early experience. It is now known, for example, that a wide variety of environmental factors, including maternal behaviour, physical and emotional stress, and exposure to toxins, drugs and hormones – especially during the early part of life – can modify the expression of the genes that regulate the central nervous system without altering the DNA sequence. This essentially means that, while the genome still provides the design blueprint for the developing brain and nervous system, it is an adjustable blueprint that can respond dynamically to information and experience coming from the environment, and thereby alter the course of behavioural development in adaptive ways.[1] Furthermore, at least some of these changes in gene expression appear to be heritable in the sense that they can be propagated, either through germ cells or via the quality of parental behaviour, to affect later generations.[2]

It is important to stress that these changes in gene expression occur more readily at certain stages of brain development. The early growth and development of the brain is punctuated by periods of increased plasticity during which it is more susceptible to remodelling and reorganisation in response to environmental triggers.[3] These episodes of sensitivity tend to occur in similar sequence across mammal species, but the time scales involved may differ considerably from one species to another.[4] It is therefore important to specify the key stages of development that are characteristic of the canin familiaris and its nearest relatives.

According to the findings of the Bar Harbor studies and subsequent research by Champagne, Heim & Binder and Joffe, the early development of the dog can be divided into a series of six natural stages or periods:

  1. the pre-natal period,
  2. the neo-natal period,
  3. the transition period,
  4. the socialisation period,
  5. the juvenile period and
  6. the pubertal period.

The pre-natal period has been largely ignored as a developmental stage in the ontogenesis of canid behaviour and temperament. Nevertheless, numerous studies of rodents and both human and non-human primates indicate that transplacental maternal influences can exert significant long-term effects on the subsequent behaviour of the offspring. Subjecting female rats to stressful experiences during pregnancy activates the maternal hypothalamic-pituitary-adrenal (HPA) axis, resulting in the release of glucocorticoid (stress) hormones. Offspring exposed to sufficiently high levels of these maternal hormones in utero display enhanced stress sensitivity when tested later in life. Similar effects have yet to be investigated in dogs but female cubs a related canid, the blue fox, whose mothers were subjected to stressful handling during the last third of pregnancy, were found to display higher behavioural reactivity in novel test situations at 35 days post-partum than those of unhandled females.[5]

During the neo-natal period the puppy is still comparatively helpless and dependent on the mother, and adapted to the life of suckling and care-soliciting. At this age, from birth till approximately two weeks, the puppies are sensitive to tactile stimuli and certain tastes and odour, but their motor abilities are limited, and neither their eyes nor ears are open or functional. Due to the immature state of their neurosensory systems, it was originally assumed that canine neonates were largely incapable of associative learning. Subsequently, it has been shown that neonatal puppies can learn simple associations and only within the limits of their own rather specialised sensory and behavioural capacities.[6] Both Fox and Zimen reported that wolves hand-reared from birth or six days of age are more reliable and friendlier towards humans than those hand-reared from 15 days or later. The mechanism for this effect is most likely some form of olfactory imprinting during the neonatal period.

Episodes of early life stress (ELS) can also have marked long-term effects on the behavioural and physical development of mammalian neonates, including puppies. In rodents, intense or prolonged stress, such as two or more hours of separation from their mother, seems to intensify neonatal sensitivity to stressful or anxiety-provoking situations later in life. In contrast, mild to moderate stressors such as brief periods of separation from the mother and litter, or daily handling, appear to have a positive impact on the stress resilience as well as producing accelerated maturation of the nervous system, more rapid hair growth and weight gain, enhanced development of motor and problem-solving skills, and earlier opening of the eyes.[7]

Some canine studies have found similar effects. For example, puppies exposed to varied stimulation from birth to five weeks of age were found to be more confident, exploratory and socially dominant when tested later in strange situations than unstimulated controls, and puppies handled gently on a daily basis from three to 21 days after birth were calmer, more exploratory, and gave fewer distress calls in eight weeks old puppy isolation tests than littermates who were not handled.[8]

To account for these effects of ELS, recent research indicates that the effects of these mild-moderate stressors are probably mediated through the changes they induce in maternal behaviour when the pups are returned to the nest. Variation in maternal behaviour toward offspring during the neonatal period is now known to have profound and lasting effects on the development of their physiological and behavioural responses to stress.

These effects appear to be mediated by four key hormones – corticotrophin releasing hormone, arginine vasopressin, oxytocin and prolactin – all of which contribute in different ways to change the developing brain’s responsiveness to stress hormones.[9] Female dogs whose pups are subjected to experimental stressors tend to lick and groom them more than the mothers of unstressed pups, and this extra attention suppresses the puppies’ HPA responsiveness. Conversely, stressed mothers tend to engage in reduced levels of licking and grooming of their offspring which respond to this by developing higher reactivity to stress and a lower propensity to engage in parental behaviour as adults, thereby perpetuating the stress-sensitive phenotype across generations.[10]

The transition period is marked by a period of metamorphosis during which behaviours associated with the neonatal existence disappear and are replaced by those more typical of later puppyhood and adult life. The whole process takes about a week, beginning with the opening of the eyes at around 13 to 16 days and ending at approximately 18-20 days with the opening of the ear canals and the first appearance of the auditory “startle” response to sudden loud noises. There is also a sudden increase in visual cortex activity during this period all the way to maturation at six to eight weeks of age. Puppies also exhibit a number of changes in behaviour during this transitional phase. They show an ability to crawl forward and backward and begin to stand and walk, albeit clumsily. They start to eliminate outside the nest and anogenital licking by the mother is no longer required to stimulate elimination. Puppies begin showing an interest in solid food at this time. They also start to engage in social play with littermates and display social signals such as growling and tail wagging. Patterns of distress vocalisation also change. Whereas neonatal puppies yelp primarily in response to cold or hunger, a three-week old puppy will also yelp if it finds itself outside the nest in an unfamiliar environment, even if it is otherwise warm and well-fed.[11] In terms of learning and the effects of early experience, the transition period resembles a continuation of the neonatal period. Puppies’ performances on both classical and operant conditioning tasks show a steady improvement at this age, although rates of learning and the stability of the conditioned response do not reach adult levels until four to five weeks.[12]

The socialisation period in puppies was first described as a “critical period” for the formation of primary social relationships or social attachments.[13] Current evidence, however, suggested that the boundaries of this period is not always as rigid as originally conceived, and that behaviour or preferences acquired during this period can often be modified or reversed at later stages, albeit with varying degrees of difficulty. Authorities in this field now favour the term “sensitive period” when responses or preferences are acquired more readily than at other times.[14]

Primary socialisation in puppies has been found to be largely independent of associated reinforcers or punishers, although emotionally arousing stimuli, both wanted and aversive, seem to accelerate the process.[15] In the case of the domestic dog, this period enables puppies to form non-conspecific attachments for humans or other animals encountered socially. For example, puppies can recognise their littermates at four to five weeks of age and cross-fostered puppies raised throughout the socialisation periodthree to 16 weeks of age – with only kitten littermates tend to avoid interacting with strange puppies when first exposed to them. This contrasts with their kitten foster littermates who will interact playfully with strange puppies on first encounter. Such experiments demonstrate that the character of the socialisation experience not only determines the young animal’s choice of future social partners but also to some extent, defines the species to which it belongs.

The ease with which most domestic dog puppies are able to establish non-conspecific social attachments is a product of selection under domestication. In contrast to wolves, domestic dog puppies readily form such attachments if they have some human exposure before seven to eight weeks of age and their socialisation window seems to remain open until around 12-14 weeks or even longer in some individuals.[16] This difference was due primarily to delayed maturation of the HPA axis in dogs resulting in retarded onset of the normal fearful/avoidant response to unfamiliar individuals and situations. Confirming this theory, Morrow et al detected significant breed differences in the age of onset of fearful/avoidant responses in puppies, as well as correlated difference in stress responses as measured by salivary cortisol.[17] Recent studies in dogs indicate that genetic polymorphism in the oxytocin receptor gene (OXTR) may account for the differences in prosocial tendencies between dogs and wolves.[18] This meant that the reduction in fear/avoidance and enhanced prosocial tendencies are outcomes of selection for extreme socialiseablity early in the domestication of selected dog breeds.[19]

During the sensitive period for socialisation, puppies may also form attachments for particular places, a phenomenon known as localisation. This process is not significantly different from the puppies’ ability to form non-conspecific attachments – the outcome is that puppies become attached to both the living and non-living parts of its environment at this age.[20] Puppies’ cognitive abilities develop rapidly during this period. Puppies at three to five weeks of age can learn to recognise novel objects they have viewed previously only on video and will subsequently exhibit less fear of the actual objects when tested at seven to eight weeks old.

In a definitive study conducted by Freedman et al (1961), the upper and lower boundaries of the socialisation period was determined to be “2.5 to 9-13 weeks of age approximates a critical (sensitive) period for socialisation”.[21] Eight litters of beagle puppies were tested in laboratory conditions until 14 weeks of age, during which time each pup received one week of moderately intensive human testing and handling before being returned to the litter. Some pups received this week of human socialisation at two weeks, others at three, five, seven or nine weeks of age. Five “control” pups remained unsocialised until 14 weeks of age. At 14 weeks, all the pups were tested for their responses to a “passive” human handler, to being walked on a leash, and to being strapped into a physiology harness and subjected to various arousing or unpleasant stimuli. Those socialised at seven weeks obtained the most favourable scores in terms of their reactions to being tested in harness. Control pups remained uniformly fearful and intractable even after many weeks of careful handling and petting.

Scott and Fuller (1965), through behavioural observations of naïve puppies to human handlers at different ages have confirmed Freedman et al’s findings. Although initially fearful in the presence of an “active” human handler, young puppies show a rapid increase in their tendency to approach and make social contact with an unfamiliar person between ages of three to five weeks. This tendency decreases thereafter. Conversely, puppies at three to five weeks show little or no fear of a “passive” handler but they become increasingly wary or fearful of strange individuals or situations beyond this age. Scott and Fuller concluded that primary socialisation period ran from about the third to twelfth week after birth with a peak sensitivity between six and eight weeks of age. During this period of peak sensitivity, the puppies’ motivation to approach and make contact with a  stranger outweighs its natural wariness.

Fox and Stelzner (1966) identified a period at around eight weeks old when puppies are hypersensitive to distressing psychological or physical stimuli. Their heart rates and rates of distress vocalisation in strange situations also tend to show developmental peaks during this six-to-eight-week period. Naturalistic observations of free-roaming dog litters have revealed that puppies also exhibit a peak in social play at around seven to eight weeks of age which would suggest that heightened sensitivity displayed by pups at this age may be related to the rapid acquisition of social skills.

The result of the Bar Harbor studies and related investigations by Guide Dogs for the Blind, USA, gave rise to various practical recommendations regarding the husbandry and training of domestic dogs.[22] Two basic rules for producing well-balanced and well-adjusted dogs were proposed.

  • Ideal time to produce a close social relationship between a young dog and its human owner is between six to eight weeks of age, and
  • Puppies should be introduced, at least in a preliminary way, to the circumstances and conditions they are likely to encounter as adults, preferably by eight weeks and certainly no later than 12 weeks.

The proposed rules should not be taken to mean that puppies are to be removed from the litter to bond with their eventual human owner. Studies on puppy morbidity and mortality showed that early-age removal from the litter and their mothers produced exceptionally high morbidity and mortality.[23] What the rules meant therefore, was the prudent selection of dog breeders whose practices (a) allow the eventual human owner to start meeting the puppy as early as six weeks old, (b) expose the puppies to the eventual and prevalent environmental conditions before handing over the puppy to its owner after 12 weeks of age, and (c) continue to influence the human owner’s practices in socialising the puppy to its environment including other species beyond 12 weeks to at least 12 months of age. These rules can rarely be achieved by commercial breeders who supplies pet shops, whose objective is unambiguously the rapid sale of adorable puppies given the high inventory cost. Perhaps uninformed and inflexible, Singapore’s authorities have tightened rules on dog breeding recently, effectively criminalising home-breeding and giving commercial breeders the free rein to continue producing problematic puppies in the cramped and chaotic Sungai Tengah area.

The juvenile and pubertal periods runs from approximately 12 weeks to two years old, depending on breed and size of the dog. This period of the young dog’s life is probably the least studied in terms of its effects on adult behaviour, despite the fact that gonadal hormones are known to modulate adolescent brain plasticity.[24] Anecdotal evidence certain suggests that experience during the pre-adolescent and adolescent periods can exert long-term effects on behaviour.[25] There is a neophobia period between seven to nine months of age where the young dog may form lasting aversion to certain humans and/or circumstances through single-event learning, especially if the event is drastic and sudden. Recent studies of working dogs have found associations between the dogs’ experiences during this period and long-term changes in adult behaviour.[26] Poor handling by novice baiters for instance could result in irreversible aversion to biting. That said, the silver lining is that environmental enrichment during puberty was found to completely erase the negative effects of early life stress on the HPA axis.[27]


[1] Bale, T. L., Baram, T. Z., Brown, A. S. et al. (2010). Early life programming and neurodevelopmental disorders. Biological Psychiatry, 68: 314–19.

Berger, S. L., Kouzarides, T., shiekhattar, R. & Shilatifard, A. (2009). An operational definition of epigenetics. Genes and Development, 23: 781–3.

Cardoso, S. D., Teles, M. C. & Oliveira, R. F. (2015). Neurogenomic mechanisms of social plasticity. Journal of Experimental Biology, 218: 140–9. Champagne, F. A. (2008). Epigenetic mechanisms and the transgenerational effects of maternal care. Frontiers in Neuroendocrinology, 29: 386–97.

Heim, C. & Binder, E. B. (2012) Current research trends in early life stress and depression: Review of human studies on sensitive periods, gene-environment interactions, and epigenetics. Experimental Neurology, 233: 102–11.

Sweatt, D. J. (2013). The emerging field of neuroepigenetics. Neuron, 80: 624–32.

[2] Crews, D. (2011). Epigenetic modifications of brain and behavior: theory and practice. Hormones and Behavior, 59: 393–8.

Curley, J. P., Jensen, C. L., Mashoodh, R. & Champagne, F. A. (2011a). Social influences on neurobiology and behavior: epigenetic effects during development. Psychoneuroendocrinology, 36: 352–71.

Curley, J. P., Mashoodh, R. & Champagne, F. A. (2011b). Epigenetics and the origins of paternal effects. Hormones and Behavior, 59: 306–14.

Dunn, G. A., Morgan, C. P. & Bale, T. L. (2011). Sex-specificity in transgenerational epigenetic programming. Hormones and Behavior, 59: 290–6.

Lynch, K. E. & Kemp, D. J. (2014). Nature-via-nurture and unraveling causality in evolutionary genetics. Trends in Ecology & Evolution, 29.

Patchev, A. V., Rodrigues, A. J., Sousa, N., Spengler, D. & Almeida, O. F. X. (2014). The future is now: Early life events preset adult behaviour. Acta Physiologica, 210: 46–57.

[3] Meredith, R. M. (2014). Sensitive and critical periods during neurotypical and aberrant neurodevelopment: a framework for neurodevelopmental disorders. Neuroscience and Biobehavioral Reviews.

[4] Heim, C. & Binder, E. B. (2012) Current research trends in early life stress and depression: Review of human studies on sensitive periods, gene-environment interactions, and epigenetics. Experimental Neurology, 233: 102–11.

[5] Braastad, B.O., Osadchuk, L. V., Lund, G. & Bakken, M. (1998). Effects of prenatal handling stress on adrenal weight and behavior in novel situations in blue fox cubs (Alopex lagopus). Applied Animal Behaviour Science, 57: 157–69.

[6] Cornwell, A. C. & Fuller, J. L. (1960). Conditioned responses in young puppies. Journal of Comparative and Physiological Psychology, 54: 13–15.

[7] Curley, J. P., Jensen, C. L., Mashoodh, R. & Champagne, F. A. (2011a). Social influences on neurobiology and behavior: epigenetic effects during development. Psychoneuroendocrinology, 36: 352–71

[8] Gazzano, A., Mariti, C., Notari, L., Sighieri, C. & McBride, E. A. (2008). Effects of early gentling and early environment on emotional development of puppies. Applied Animal Behaviour Science, 110: 294–304.

[9] Bale et al.,2011

[10] Champagne, F. A. (2008). Epigenetic mechanisms and the transgenerational effects of maternal care. Frontiers in Neuroendocrinology, 29: 386–97.;

Foyer, P., Bjällerhag, N., Wilsson, W. & Jensen, P. (2014). Behaviour and experiences of dogs during the first year of life predict the outcome in a later temperament test. Applied Animal Behaviour Science, 155: 93–100.

[11] Fox, M. W. (1971). Behavior of Wolves, Dogs and Related Canids. New York: Harper and Row.

[12] Scott, J. P. & Fuller, J. L. (1965). Genetics and the Social Behavior of the Dog. Chicago, IL: University of Chicago Press. Scott, J. P. & Marston, M. V. (1950). Critical periods affecting the development of normal and mal-adjustive social behavior of puppies. Journal of Genetic Psychology, 77: 25–60.

[13] Scott, J. P., Stewart, J. M. & DeGhett, V. J. (1974). Critical periods in the organization of systems. Developmental Psychobiology, 7: 489–513

[14] Meredith, R. M. (2014). Sensitive and critical periods during neurotypical and aberrant neurodevelopment: a framework for neurodevelopmental disorders. Neuroscience and Biobehavioral Reviews.

[15] Scott, J. P., Stewart, J. M. & DeGhett, V. J. (1974). Critical periods in the organization of systems. Developmental Psychobiology, 7: 489–513

[16] Lord, K. (2013). Comparison of the sensory development of wolves (Canis lupus) and dogs (Canis lupus familiaris). Ethology, 119: 110–20.

[17] Morrow, M., Ottobre, J., Ottobre, A. et al. (2015). Breed-dependent differences in the onset of fear-related avoidance behavior in puppies. Journal of Veterinary Behavior. http://doi:10.1016/j.jveb.2015.03.002

[18] Arueti, M., Perach-Barzilay, N., Tsoory, M. M., Berger, B., Getter, N. & Shamay-Tsoory, S. G. (2014) When two become one: the role of oxytocin in interpersonal coordination and cooperation. Journal of Cognitive Neuroscience, 25: 1418–27

[19] Wynne, C. D. L., Udell, M. A. R. & Lord, K. A. (2008). Ontogeny’s impacts on human-dog communication. Animal Behaviour, 76: e1–4.

[20] Scott & Fuller, 1965, p112

[21] Freedman, D. G., King, J. A. & Elliot, O. (1961). Critical periods in the social development of dogs. Science, 133: 1016–17., p1017

[22] Pfaffenberger, C. J. & Scott, J. P. (1976). Early rearing and testing. In Guide Dogs for the Blind: Their Selection, Development and Training, eds. C. J. Pfaffenberger, J. P. Scott, J. L. Fuller, B. E. Ginsburg & S. W. Bielfelt. Amsterdam: Elsevier, pp. 13–37.

[23] Slabbert, J. M. & Rasa, O. A. E. (1993). The effect of early separation from the mother on pups in bonding to humans and pup health. Journal of the South African Veterinary Association, 64: 4–8.

[24] Heim, C. & Binder, E. B. (2012) Current research trends in early life stress and depression: Review of human studies on sensitive periods, gene-environment interactions, and epigenetics. Experimental Neurology, 233: 102–11.

[25] Dehasse, J. (1994). Sensory, emotional and social development of the young dog. Bulletin for Veterinary Clinical Ethology, 2: 6–29.

[26] Foyer, P., Bjällerhag, N., Wilsson, W. & Jensen, P. (2014). Behaviour and experiences of dogs during the first year of life predict the outcome in a later temperament test. Applied Animal Behaviour Science, 155: 93–100.

[27] Francis, D. D., Diorio, J., Plotsky, P. M. & Meaney, M. J. (2002). Environmental enrichment reverses the effects of maternal separation on stress reactivity. Journal of Neuroscience, 22: 7840–3.

The Essence of Marker-based Positive Reinforcement in Dog Training

Actions that lead to beneficial outcomes are more likely to be repeated than those that do not. This process, whereby the probability of a behavioural response increases as a consequence of the outcome of that response, is referred to as positive reinforcement. Intra-cranial self-stimulation (ICSS) is a simple behavioural model that distils positive reinforcement to its minimum neural elements. In ICSS paradigms, mammals make instrumental (operant) responses in order to deliver stimulation to a specific brain area. Sites containing dopamine neurons, or their ascending projections are particularly effective in eliciting this behaviour[1], and systemic administration of dopamine antagonists causes dramatic reductions in ICSS[2], strongly implicating dopamine neurons as a neural substrate. A recent study used genetically targeted channelrhodopsin-2 (ChR2) to specifically activate Ventral Tegmental Area (VTA) dopamine neurons and confirmed that dopamine neurons are indeed sufficient to drive vigorous ICSS[3], consistent with a rich literature demonstrating that VTA dopamine neurons play critical roles in learned appetitive behaviours[4].

Dopamine is an organic chemical of the catecholamine and phenethylamine families. Dopaminergic signalling is associated with reward-motivated behaviour and motor control. In the ventral tegmental area (VTA), the mesolimbic pathway projects from the prefrontal cortex to the nucleus accumbens of the amygdala, cingulate gyrus, hippocampus, and pyriform complex of the olfactory bulb.

The dopaminergic projections in the amygdala and cingulate gyrus are responsible for emotion formation and processing. In the hippocampus, the presence of dopaminergic neurons is associated with learning, working memory, and long-term memory formation. Lastly, the pyriform complex of the olfactory bulb is responsible for providing dogs with the sense of smell.

In the mesolimbic pathway, dopamine is released during pleasurable situations, causing arousal and influences behaviour (motivations) to seek out the pleasurable activity or occupation and bind to dopaminergic receptors present in the nucleus accumbens and prefrontal cortex. Increased activity in the projections to the nucleus accumbens play a major role in reinforcement and in more extreme cases with addictions.[5]

Having understood the intrinsic effects of Dopamine, we shall take a look at the extrinsic factors. Once again, we assume that the goal of reinforcement learning is to maximise future benefits. Analogous to utilities in economic theories, value functions in reinforcement learning theory refer to the estimates for the sum of future reinforcers. However, since the dog cannot predict the future changes in its environment perfectly, value functions, unlike utilities, reflect the animal’s empirical estimates for its future reinforcers. Reinforcers in the distant future are often temporally discounted so that more immediate reinforcers exert stronger influence on the animal’s behaviour. The reinforcement learning theory utilizes two different types of value functions.

First, action value function refers to the sum of future reinforcers expected for taking a particular action in a particular state of the environment. The term action is used formally: it can refer to not only a physical action, such as counter surfing in a particular location with specific limbs to obtain food (reinforcer), but also an abstract choice, such as ignoring kibbles in favour of treats during mealtime.

Second, state value function, refers to the sum of future reinforcers expected from a particular state of the animal’s environment. If the animal always chooses only one action in a given state, then its action value function would be equal to the state value function. Otherwise, the state value function would correspond to the average of action value functions weighted by the probability of taking each action in a given state.

Neural signals related to action value functions would be useful in choosing a particular action, especially if such signals are observed before the execution of a motor response. Neural activity related to state value functions may play more evaluative roles. In particular, during decision making, the state value function changes from the weighted average of action values for alternative choices to the action value function for the chosen action. The latter is often referred to as chosen value.[6]

In positive reinforcement training for dogs, the term “positive” denotes the addition of a reinforcer wanted by the dog contingent on the performance of a wanted behaviour, immediately after the performance of the wanted behaviour by the dog. There are several assumptions that must be in play for true positive reinforcement training to be effective:

  1. The training dog’s mesolimbic pathway is normal and its empirical estimates of future reinforcers arising from its participation in human-directed activities are pleasurable;
  2. The dopaminergic projections in the amygdala and cingulate gyrus are normal and forms pleasurable emotional association between the training activities and future reinforcers;
  3. The dopaminergic neurons in the hippocampus are functioning normally and form pleasurable working memory from the chosen action value function;
  4. Through repetitions in a prescribed state value function, the hippocampus forms long-term pleasurable memory of the chosen action value function;
  5. The human trainer is consistent in his/her use of neural signals (marker) during training;
  6. The human trainer sets up the training environment such that the state value functions guide the dog towards the chosen value.

It is therefore a fallacy that positive reinforcement training can be applied to dogs which are defensive beyond their threshold, where the dog can be observed to be cowering in a corner of the room, tail tucked between its hind legs, ears folded down and eyes wide open (pupil dilation). Such a dog is incapable of feeling pleasured as cortisol has exerted its effects on the dog.

Similarly, a dog with an extreme level of predatory instinct is not susceptible to positive reinforcement training. Predation is a stressful activity and cortisol levels during predation, especially if parts of the predation process are incorporated in the training to bring about the wanted behaviour, are significantly high. An example of this would be bite training where the dog is frustrated on purpose to produce the Bark and Hold behaviour. A skilled trainer however, knows how to work the predatory instincts at a sub-threshold level to produce the wanted behaviour. However, prey-based training can rarely be called positive reinforcement training as a training often involves negative punishment and negative reinforcement, topics we will cover in subsequent posts.

Trainers who simply click their clickers and feed the dogs all kinds of treats lack the depth of understanding in positive reinforcement training. Good trainers will ask pertinent questions on the dog’s routine of life, dietary habits, predatory instincts amid others. They will also observe the dog for its social and defence drives to prepare a customised training programme that is suited to the dog’s unique temperament.

A dog that has been overfed on a routine basis will not perceive food treats as pleasurable future reinforcers. A dog that has been restrained on a daily basis will participate gamely in an active training session where it gets to sprint, sniff and roll on the lawn. There is no positive reinforcement when the future reinforcer is not deemed pleasurable or when pleasure is derived from non-human directed activities. A skilled trainer will teach dog owners the formation of neural signals at the onset of positive reinforcement training. He/She will also coach the dog owner on the arrangement of state value functions (antecedent arrangement) that will evolve along with the dog’s level of learned appetitive behaviours.