Saturday, March 5, 2011

Would You LAT?

It's a big world, full of marvelous wonders to satiate hungers both common and incredible, but it's not for the weak of spirit. Something those of us with reactive dogs know all too well, when just looking at this brave new world can prove over stimulating. Luckily, there are a few tools to add to your repetoire that can make everything a little easier on the reactive dog's eyes.

One of the new techniques being used with reactive dogs is called the "Look at That" (LAT). This exercise was popularized by trainer Leslie McDevitt in her book Control UnleashedLAT is a chained behavior. What I mean is that it is made up of more than one behavior and then those behaviors are strung together. To be more specific, the dog looks at the trigger, and then looks back at the handler. To get these behaviors, a trigger is presented, and the dog is marked for looking at the trigger.

Many people seem to have trouble with timing right off the bat. They give the dog time to look at the trigger, and then they give the dog time to think about the trigger, which, of course, gives the dog time to react to the trigger. I set a two to three second time limit for looking at triggers with my own dogs. It's long enough for them to see the trigger and really notice it, but it's usually too short an amount of time for them to start acting like morons. Some dogs can handle more than three seconds, but many dogs can only handle an instant. Timing is important. Another common problem - and this is one that crops up a lot even on other exercises - is that people tend to feed the reward close to the dog. The dog sees the trigger and the handler marks and then waves the food under the dog's nose. The dog should hear the click and then look for the treat, and the treat should be given in close to the handler's body. The dog shouldn't be looking for the treat in outer space, although there is a time and place for treats that appear out of nowhere. If the dog is not responding to the marker, either he or she is too close to the trigger (try increasing distance), or the marker doesn't have a strong enough association with the reward (work on conditioning the marking with good stuff away from the trigger).

Once the dogs has the behaviors of look at trigger/look to handler figured out, it's time to put the behavior chain on cue. Personally, this is my biggest problem with the LAT exercise. I don't mind marking a dog to look at other dogs. Heck, I've done it with Rubi. I just have trouble convincing myself to tell a dog to look for her triggers, in large part because I've seen it backfire for people. I labeling this behavior creates hypervigilance. It's like walking to my patio door that looks out on an empty backyard and asking, "Do you see squirrels?" It doesn't matter if there are squirrels out there or not, the dogs are going to start obsessively looking for them. I find the same is true for many people if they say "look at that" - the handler cues before the dog sees the trigger, and the dog starts to obsessively look for the trigger. Sometimes the dog sees the trigger and looks back for a treat. Sometimes they see the trigger and are so aroused by looking for it that they react to the trigger. And sometimes, what the handler thinks is a trigger isn't a trigger at all (for example, the person thinks all dogs are triggers and cues for the white dog walking by, when really their dog only reacts to black dogs), and the dog becomes obsessed with looking for a trigger that isn't there. So I don't cue the LAT.

What I want is for Rubi to see a dog and then look back at me without any action on my part at all - an "auto watch." I started by marking Rubi for looking at other dogs, and even now I still do this if there's a dog I think is going to really upset her. But once Rubi figured out the look at dog/look at me routine, I pretty much stopped marking her for looking at the other dog. Now, I mark her for looking back at me. After she gets her reward, if she looks back at the other dog and then looks back at me, I mark and reward her again. And if she keeps looking at me for another two seconds after that, I'll mark and reward that, too. Ideally, I want to be the most interesting thing in Rubi's environment, and I reward accordingly. It's really rare that I don't reward my dogs for looking at me, and as a result, it's a pretty strong default behavior. When my dogs aren't sure what to do, they look at me - and then they usually keep looking at me until I tell them want to do. I really, really like this, particularly for my reactive dogs.

Here's a reasonably short video of Rubi and I practicing auto watches in the lobby at TCOTC. The video was taken on this day, so you know what I mean when I talk about her blonde moments (no offense to blondes, I should have probably said her "Rubi moments"). I would've turned the sound off altogether, but I wanted you to be able to hear the clicker. The video starts when the first dog appears, and ends before they all leave, so there are dogs present for the whole thing.

I did mark Rubi a few times for returning to heel position. Being still around other dogs is still very difficult for Rubi, so staying in one place is something I want to reinforce. I like that place to be heel position because it helps me to be sure we're both looking at the same thing(s). Around second eighteen, Rubi throws me a few of what I call "flickers." She twitches her head toward the other dogs, but doesn't actually look at them. This is a sign that she's becoming more comfortable with the other dogs and doesn't feel the need to watch them; however, I don't mark her for flickers. That's not the game. The game is look at other dog/look at me or just look at me. It's not to develop small head seizures to fake mom out. After she gives me those flickers, I start building duration for the amount of time she looks at me. I still keep it really easy for her, but I vary the time before she gets the treat from instantly to about four seconds, and by the end of the video, she's giving me a lot of attention. And, of course, you can do a lot with a dog once you have her attention.


  1. Thanks for clarifying the clicking/rewarding to be: "Look At Trigger, Then at Me" or "Look at Me". I was wondering about WHEN and HOW OFTEN to use this technique when I left class last Thursday, and the video :-) and your comments made things a lot clearer. I don't want to create a hypervigilant behavior or confusion when I use LAT either, and it can be confusing... You are a great example! Thanks for your patience and kind words in the CB class on Thursday nights too, they mean a lot to those of us that are just beginning dog owners but want to do the best job we can for and with our wonderful dogs. You are much appreciated :-)

  2. Thanks for the kind words, Cindy! You actually weren't the first person to come to me this week with questions about LAT, hence the blog post. I love Cosi, she's such a fun little dog, and the three of you are a blast to work with. Glad you found your way to the blog!

  3. I've always wondered about LAT ever since I read her book. I don't have a dog who has needed it so haven't obsessed over it too much but I've wondered about the reasoning of sticking with LAT after the dog seems to need it. I understand that it can really help interrupt the dog from intensely focusing on his trigger but in her book Leslie continues to play the game quite often.

    Do you think that the majority of people (not those with truly reactive/aggressive dogs)could get the same results just by clicking offered eye contact and having a high rate of reinforcement? Or do even the moderately crazy dogs really benefit from being clicked for looking at the potential or even old trigger rather then just rewarding any attempts at offered attention?

  4. I will likely keep rewarding Maus and Rubi for auto watches periodically for the rest of their lives, because it's an important tool for them for dealing with their triggers. Rubi in particular, has been punished for looking at other dogs, so un-cued LAT was helpful for desensitizing her to that. An auto watches is, among other things, a behavior that's incompatible with wailing and lunging, which is her first go-to behavior around other dogs.

    So for dogs without triggers, per say, who are not prone to going over threshold, I've never felt the need to mark for looking at distractions. I just wait for them to get bored with the distraction and then reward for when they look back at me, the same way you would train a "leave it." I used just a high rate of reinforcement for offered attention for both Allister and Piper, never using anything like LAT work, and I think they turned out pretty well. So I think it's very possible to take your average, non-reactive, goofy, hyper dog and train excellent attention without ever marking them for looking away from you.