Thursday, September 1, 2011

Kathy Sdao Seminar: Good Foundations

Pretty much all dog training is based on some form of conditioning, either operant or classical. Since classical techniques tend to be less flashy (read: tedious and boring), most trainers choose to focus on operant ones. Kathy even admitted that while classical conditioning is a bit like watching paint dry, it's so incredibly powerful that it deserves to be the star once in awhile. And of course, Kathy did an amazing job giving classical conditioning the billing it deserves.

Both types of conditioning are built up over repetitions. The difference is that operant conditioning is about consequences, while classical conditioning is not. Operant conditioning seeks to influence the dog's behavior by pairing his actions with rewards like food. The dog only gets the food if he behaves in a certain way. By contrast, classical conditioning doesn't care about what the dog is doing at all- its only goal is to create a direct association between two stimuli. There is no behavioral criteria, and the dog gets the food no matter what he does or doesn't do.

Instead, classical conditioning focuses on the dog's feelings and his reflexes by repeatedly pairing a neutral stimulus with something that elicits a strong reaction from the dog. For example, if everytime a dog hears a bell he gets some food, he'll soon learn to expect some food whenever he hears a ringing sound. Classical conditioning can also be used to change a previously learned association, typically from bad to good. This is called counter-conditioning.

This form of conditioning is incredibly useful when working with fearful, reactive, anxious, or aggressive dogs, but is often poorly understood or glossed over. It does take a bit of time to do it right, and people often try to cut corners, or skip over it entirely. It's a shame, really, because if you can change a dog's feelings about something, the behavior will often change as a result. After all, if children are no longer scary, there's no need to lunge and growl at them, right?

The problem with classical conditioning is that it really doesn't hold up well for the long-term. Yes, you can create associations pretty easily, and you can even change associations with a bit more work, but classical conditioning is incredibly fickle. Unlike operant conditioning, where behavior can be maintained with intermittment rewards, classical conditioning can extinguish quickly if the association is not maintained.

This means that classical conditioning is not a holistic plan. It's simply too exhausting to continually maintain the pairing of scary thing=good thing. The person will get tired, or will screw up. It's also impractical to constantly have treats (or other good things) on you. Ultimately, Kathy said, you'll need to switch to an operant technique, whether that's CAT or BAT or Control Unleashed or whatever.

So why bother with classical conditioning at all? Why not just cut straight to the chase and use an operant strategy from the start? The biggest reason to do this is because most dogs with fear/reactivity/anxiety/aggression issues are okay... until they're not. And usually, once they're “not,” they've gone over threshold, and are now in that fight-or-flight mode where they can no longer think. Then it's too late. You can't teach the dog anything. You can't reinforce an acceptable behavior because the dog isn't physiologically capable of learning anymore. Operant conditioning becomes impossible.

So, you start at the beginning. You pair the sight of a trigger with good things, over and over again, no matter what the dog is doing, so that the dog no longer sees the trigger as a scary thing, but rather, as a thing which results in chicken. You'll know this association has happened when the dog gets demanding- he'll see the trigger, and instead of reacting, he'll be obnoxiously nudging your hand or staring at your chicken pocket. And since he wants that chicken so badly, it's pretty easy to ask him to do something to earn it.

In other words, what classical conditioning does is give you the foundation on which to build those operant behaviors. It helps your dog relax enough to think. It gives you enough time to intervene. It gives you the space you need to begin training the behaviors you want instead of constantly focusing on the ones you don't.

Of course, it's not as easy as I make it sound. If it was, Kathy wouldn't have been able to spend a full day discussing classical conditioning. Indeed, Kathy shared some really interesting- and important- tips to make the most of classical conditioning. If you're going to go through all that work, you might as well do it right! I'll share that info with you in the next post.

In the meantime, I'd love to hear from other people with reactive dogs. Did you do classical conditioning as a foundation? What was your experience like?

7 comments:

OntarioSheltie said...

Hi Crystal,

Thank you for explaining the difference between classical and operant conditioning and how it affects our dogs. I know that I use both of these techniques with Toby but I can never remember the difference (and will probably forget again and refer back to this post.).

When it comes to classical conditioning, I've noticed exactly what you describe. Toby, who is reactive to noises on walks, will digress if he is not walked every day (classical conditioning). Toby's operant conditioning is still a work in progress as well, but over all I'm seeing good results.

My other sheltie Ginny had some fear issues as well for a while, when she was stung by a bee while playing in grass. Operant conditioning is helping her over come her fear and she's almost back to normal now. I spend more time with her in the grass, which has some classical conditioning aspects, but she also gets treats in the grass and gets played with in the grass, so she's begining to associate grass with rewards and fun, instead of pain.

Kristine said...

Thanks for explaining this so well. I knew what the two types of conditioning were but I don't think I understood the subtle differences before. For one thing, I didn't really realize until now that how we initially treated our dog's reactivity was with classical conditioning.

I have to say, it definitely worked. It gave me the confidence to walk my dog again, which I hadn't had for a long time. By treating her every time we saw a person or a dog, I started to see that it would be possible to walk a whole block without my dog barking and lunging. That alone was worth wonders! Once I was feeling better about things we were both able to advance further to an operant style of training.

I think it really depends, though, on what works for the handler and the dog. People say it is inconvenient to have treats on them at all times, but I have never found that. I don't leave the house with my dog - actually, I don't leave the house period - without some sort of reward. It's become habit and I don't find it annoying at all. I like to be prepared. Even though she has overcome most of her fears and doesn't really need the reward, having the food available still gives me confidence. It may be possible I am using it as a crutch but I am not sure that is really something I should worry about.

Crystal (Thompson) Barrera said...

OntarioSheltie- you are quite welcome. It IS difficult to keep these concepts straight. I remember the first time I read about them, I was horribly confused. Since then, I've had the benefit of a ton of seminars, so it's easier.

Kristine- you are also welcome. Kathy's seminar did a REALLY nice job of explaining some of the nuances. I knew the basics, but she blew me away. I will not come anywhere close to doing as good a job as she did.

Like you, I never go anywhere without treats. It doesn't seem like a huge deal, but the other day, I had Maisy in a crate and I was about 20 feet away. In the meantime, another dog ran up to her crate, and I couldn't get a treat to her! Yikes! So, I definitely think switching to an operant style at some point is useful... (In that situation, I bridged the gap with praise. Thank goodness, praise has been classically conditioned to predict good things!)

Kirsten (peacefuldog) said...

Ooooh, this is such a good post! Thank you for explaining this so clearly!

I have used classical conditioning with my fear reactive dog Lamar, with good results. I've been doing it for about two years and have not really transitioned to operant with him, unless you count the fact that when he looks at another dog, I say his name and he looks at me for his treat. But we haven't really asked him for anything beyond looking at me. Maybe it would get even stronger if I did?

With my other aroused-reactive foster dog, I'm still trying to get to the point where he is below threshold enough that he'll even eat when he sees another dog, let alone do a behavior!
http://peacefuldog.blogspot.com/2011/07/fozzies-second-reactive-dog-class.html

This post helps a lot--I'm going to think about it more and maybe figure out better ways to work with both my boys. Thanks again.

Crystal (Thompson) Barrera said...

Kirsten, that TOTALLY counts as operant conditioning. Have you ever tried Look at That (from Leslie McDevitt's Control Unleashed)? That's my personal favorite operant strategy.

Sadie Hart said...

For Toby, my dog reactive dog, our foundation work was primarily CC. I credit that to his ability to cope more than any other technique we've done. Add that to my increased confidence when we were out and about (because I knew what to do for a change and I could see how it was helping him), and it made it so much easier to keep improving. We still had our bad moments... can't always control the universe around you, no matter how hard ya try!

Kirsten (peacefuldog) said...

Hey Crystal-yes, we have worked a lot with Control Unleashed and LAT! That's when Lamar started looking at me.

And now that I think about it, Lamar is definitely in an "operant-capable" frame of mind now when we see other dogs. It took a lot of classical conditioning to get us here, but I think he could do other behaviors if I asked him to. Thanks for the great discussion!