When I started reading Kathy Sdao’s book – which is as much, if not more her personal journey through life as it is a book for dog trainers -, my first impression had nothing to do with training animals: I thought that, should I ever choose to believe in some transcendental entity, I’d like this entity to be like Kathy’s god. That god actually sounded like a god I could live with. A god that loved his creatures unconditionally and provided for them not because they were being good or despite their badness, but because they were, period. A god that was okay with Kathy’s referring to humans as just one animal species among other animal species and agreed with her that plenty in life was free. Even though I had set out to read a dog training book, Plenty in Life is Free turned out to be a book I enjoyed for all kinds of reasons – style, stickiness and anecdotes about curly hair girls.
Having seen Kathy’s videos on youtube as well, I was looking forward to her seminar in Austria. I would probably have missed it if Christine Schragl hadn’t pointed it out to me on the BAT yahoo list – thank you, Christine!
It turned out to be the best seminar I’ve been to in a long time. It inspired a number of articles I’m planning to write and provided new training insights for me. It gave me an idea for my own classes, too: I got Kathy’s permission to borrow her pineapple idea when honoring someone’s amusing contribution to class discussions. A pineapple? Yes, a pineapple. You’ll have to attend one of Kathy’s seminars to find out what it reinforces.
Anyways, here’s a couple impressions from the seminar, as well as a little Pavlovian background I read up on: *)
Cue discrimination test
One of the practical exercises we did was a cue discrimination test. We used different body postures, closed our eyes, changed the distance between dog and handler, used a word that rhymed with the cue, exchanged the cue’s vowel etc. to find out whether our dogs still understood their cues and to ask ourselves whether we wanted them to understand: did we want our dog to down, even if we said “clown”?
Bulldog Lilo plays the cue discrimination game: she knows a hand signal that means lift her paw. But can she also do it when her handler kneels on a chair rather than standing in front of Lilo?
After all that hard work, Lilo’s mum needs a break!
The Chihuahua plays Cue Discrimination – does she understand a down when her mum doesn’t cue her with her hole body, but only her hand?
Working with dog-reactive dogs: classical counter-conditioning, example 1
Client: adult Rhodesian Ridgeback, dog-reactive (and generally nervous)
Decoy: “bomb-proof” curly coated retriever with an experienced handler
Suggested approach: management; classical counter-conditioning and desensitization
Conditioned stimulus (CS): strange dog
Conditioned response (CR): aggression
decoy dog (CS) —-> aggression (CR)
Unconditioned stimulus (US) to be added: tug toy
Unconditioned response (UR) to US: joy
tug toy (US) —-> joy (UR)
Combining the two:
decoy dog (CS) + tug toy (US) —-> joy (UR)
(Note that the US must appear after the dog has noticed the CS!)
decoy dog (CS) —-> joy (new CR)
This dog-reactive Ridgeback is used to scanning the environment for other dogs that appear unexpectedly – a stressful life. In this set up, Christine appears with the decoy, her Curly Coated Retriever, from a distance below the Ridgeback’s threshold. At the Ridgeback’s choice point (the moment she spots the retriever), her mum waits for the Ridgeback to notice the other dog, then rewards with a tug toy.
This is so the Ridgeback’s emotion triggered by the CS will eventually change from aggression (CR: “Shit, scary dog!”) to joy (new CR: “Yeah, tug time!”). The sight of a dog becomes a classically conditioned stimulus (equal to the bell in Pavlov’s experiment) meaning tug time is coming.
In a set up like this, it is important that the US (in this case the tug toy) comes after the dog has noticed her trigger (Pavlov would call this delayed or trace conditioning). If the unconditioned stimulus happens simultaneously as or before the dog notices her trigger, the conditioning will not work! If the Ridgeback’s mum had noticed the Retriever first and immediately (i.e. before her dog had seen him) pulled out the tug toy, mum would at best have distracted her dog and at worst have poisoned the toy (i.e. made the toy unattractive/scary).
Combining classical counter-conditioning with desensitization
The most effective way to help the Ridgeback is to combine classical counter-conditioning with desensitization. Her training plan for the next weeks should not only include set ups like the one we did this weekend, but also elements of desensitization. That is to say, the intensity of the stimulus will be gradually increased by means of, for example:
– decreasing the distance from the suddenly appearing decoy
– increasing the duration of the decoy’s appearance
– chainging the decoy’s walking direction and speed
– practicing set ups in challenging environments.
It’s important to only increase one criterion at a time and never put the Ridgeback over threshold. When deciding whether the Ridgeback is ready for us to raise criteria, we’re not looking for the absence of anxiety, but for the presence of joy upon perceiving the decoy dog (CS).
Contingency speeds up the training process
The power of contingency tends to be underrated. However, experiments show that contingency affects Pavlovian learning on two levels: on the level of the CS (trigger, strange dog) and on the level of the US (reward, tug toy).
The level of the unconditioned stimulus (reward, tug toy):
during the training period, this special tug toy should always and only happen after the Ridgeback has seen a dog.
The level of the conditioned stimulus (trigger, strange dog):
strange dogs should only appear in the Ridgeback’s environment when they will be followed by the tug toy.
Management
It will take a while for the sight of a dog to be generalized to all kinds of locations, dogs and trigger intensities and become an alternatively conditioned stimulus meaning tug time. Therefore, for the next 6 weeks, the Ridgeback’s mum will also manage her dog’s environment and avoid walks where she’ll unexpectedly encounter strange dogs at close distances. At the same time, she will practice set ups until she has created a reliable happy emotional response (new CR) to the sight of strange dogs (CS).
It is important a reactive dog’s environment be well managed while she learns: we want to build an alternative neural pathway that leads to happy reaction. We do this by means of the set ups. However, at the same time, we have to prevent the old neural pathway that leads to an upset and reactive dog from being used – by means of management. Once the new neural pathway is strong, the Ridgeback will be able to encounter other dogs on walks without getting upset.
Why is it important that the Ridgeback isn’t surprised by strange dogs that take her over threshold outside of training sessions?
In the context of explaining what to do and what not to do when training a dog how to stay home alone, Jean Donaldson (The Culture Clash) uses a jungle metaphor that can also help us visualize what goes on in a dog-reactive dog: imagine the canine brain like a jungle. Our set ups are the machete by means of which we build the pathway that leads to joy (CS: “Yeah, tug time!”). In order for this newly created path to be attractive for the neurones to travel, we have to make the old path (leading to aggression; CS: “Shit, scary dog!”) less attractive. Only if this old path ceases to be taken by the neurones will it start to be overgrown by jungle plants. Therefore, we might have to use a road block (i.e. management) to prevent access to the old path until the new path is well established and the old one naturally made inaccessible by banana plants and fern.
Or, in more scientific terms and in the words of Robert A. Rescorla: “when the likelihood of a US is the same in the presence and absence of the CS […], there is little evidence of conditioning at all. […] [C]onditioning depends not on the contiguity between the CS and the US but rather on the information that the CS provides about the US.”
When (not) to use classical counter-conditioning with clients
Note that classical counter-conditioning will not build interaction skills. Rather, it helps the dog relax even when there’s other dogs to be met: eventually, seeing a dog in the distance will be a sign that it’s tug time. However, this is not to get her used to interacting with other dogs, but to peacefully coexisting with them/passing them on walks. Should a client wish to further her dog’s interaction skills, we wouldn’t use classical counter-conditioning but a different approach, or a different approach in addition to classical counter-conditioning.
Working with dog-reactive dogs: classical counter-conditioning, example 2
The young German Wirehaired Pointer, Ellie, is scared of other dogs. Kathy explains the set up to her handler: the decoy, Border Collie Kodiak, will be walked past her at a distance below Ellie’s threshold. At Ellie’s choice point (the moment she spots the decoy), Ellie will be fed. This is to teach her that the sight of a strange dog means food is coming, hence changing her emotional reaction, like in the set up with the Ridgeback.
Client: young German wirehaired pointer, dog-reactive (mild fear to unfamiliar dogs)
Decoy: Border Collie Kodiak with an experienced handler (first set up); two Cocker Spaniels with an experienced handler (second set up)
Suggested approach: classical counter-conditioning and desensitization
Conditioned stimulus (CS): strange dog
Conditioned response (CR): fear
decoy dog (CS) —-> fear (CR)
Unconditioned stimulus (US) to be added: food
Unconditioned response (UR) to US: joy
food (US) —-> joy (UR)
Combining the two:
decoy dog (CS) + food (US) —-> joy (UR)
(Note that the US must appear after the dog has noticed the CS!)
decoy dog (CS) —-> joy (new CR)
Ellie does very well, first with Kodiak walking past her, then Kodiak walking faster past her, then directly at her, and later with two cocker spaniels walking past her. The new CR is already setting in: Ellie spots the decoys and looks expectantly at her handler: bring on the treats!
Ellie is a foster dog. In order to make it easier to rehome her, her foster mum will work on her on a similar training program as the Ridgeback’s mum, with the only difference that Ellie’s reward is food.
Lunch break means play time!
Sunday night
Gudrun and Kathy draw the winners from the chocolate game, and everyone – including the fake Doberman that was used to work with the Chihuahua – poses one last time for the group picture …
… and something for the office wall.
Thanks to everyone who was involved in a pawesome weekend that went by way too fast!
*) Should you notice mistakes involving behaviorological terminology, misunderstandings or ambiguities in my explanation, please point them out to me! Phoebe Flausch and I love our science, but we’re still learning and always appreciate feedback and constructive criticism.