Behaviour Adjustment Training

A while ago I came across a training technique/system called BAT (Behaviour Adjustment Training) which is to be used in cases of fear, reactivity and aggression to any given trigger.

This system was developed by Grisha Stewart of Ahisma Dog Training (amongst others).

The aim of BAT is to teach the dog to be able to cope with and interact in environments and social situations which have previously triggered fear, anxiety, reactivity or aggression without reverting to previously reinforced fearful, panicked, reactive or aggressive responses. The idea is to condition and reinforce a new calm reaction to the trigger (much like more general counter-conditioning).

In the following video Grisha explains more:

Organic socialization: BAT for fear and aggression

The following article also explains the basic principles of BAT training:

BAT Article

Grisha is holding a two day course in the UK in March, which will focus on BAT training for fear and aggression and is suited to behaviourists/trainers as well as dog owners.

I'm unable to go as we have another dog course on that week but thought I'd post it here just in case anyone else is interested.

More info HERE

The Dogs Playing

Just a quick vid from this morning's walk of the guys mucking around and playing :-D

Playing 

Awwww :)

WTG Dash

On weds as usual we had training. As you may have read in previous posts I've been working on arousal and stimulus control work with him as whilst he focuses perfectly when it's his turn to run, in between he gets very excited and overaroused watching all of the other dogs run.
We've been working on mat work and teaching a high variable reinforcement schedule (for duration and distance) to this behaviour. ATM he's successful at a VR schedule of about 60 at home. We've also been working on sheepball games to teach focus which he is really getting the hang of. Because of the type of WSD that he is, he instinctively works a real distance away and has a very strong eye which meant that he was less likely to flank to more subtle body cues from me. I've been teaching him to work in a bit closer (for the purpose of using the game in an agility setting) by working with him initially on the full length of his double ended lead on a small concrete basketball court near us. I've been clicking and reinforcing him for offering me behaviours such as facing me (12 oclock position) and the ball at my feet, right flanking (which he was offering me more) and I've gradually been able to capture him flanking to the left and click and reward it. Where he was offering me a really strong "eye"/stare behaviour, I've also been able to teach him to lie down and "watch" his (sheep)ball and remain in that position with the ball in front of him whilst I walk behind him or around the court. He'll now only move when I release him. Having been successful at the basketball court I've been able to integrate this into our walks with him offlead. I had to lower the reinforcement criteria quite a lot to start with, as previous to starting to do any sheepball training I had always, without thinking about it, rewarded him for offering me distance by throwing his toy when he faced me at a distance. This meant that I had to do a few sessions with him on his double ended lead, before progressing to working him offlead and marking and rewarding any flanking and the other behaviours as I did in the initial stages. In the initial stages on walks and prior, I often needed another equal or higher value toy to "swap" so that he would offer me the previous one that he had been working until reinforced. We've managed to phase this out on walks, purely through repetition of the game so that he'll now just offer me the original one once he has been released to the ball or reinforced. He still needs the additional toy at training though as it is a more exciting environment that I am yet to proof to.

In any case, this week at training I was really chuffed with him. He worked his socks off and focused really well on his mat and without even looking at the other dogs running. He was also able to work on tricks whilst we were waiting for our turn outside the school (previously he hadn't really been able to focus on this as he was so fixated on the other dogs running and the equipment). About half way through we were even able to do sheepball work offlead in the school with other dogs working around us. Total focus!

Just goes to show that the phrase "Great dog, shame about the handler!" is so true!

Dog Whisperer comes to the UK :( Let's emigrate!!!

 I found out yesterday that Cesar Millan, a "trainer" (not the word I'd use) popular with some, who features in a programme called  the Dog Whisperer on Nat Geo Wild has filmed a UK series, the first episode of which, aired on Nat Geo Wild last night.

Bad times! :(

I've watched several episodes from various US series (although rarely a peaceful viewing and usually involves me throwing the remote at the telly in exasperation!). 

For those of you that aren't familiar with his shows, I wouldn't bother watching them. In all honesty they are just quite distressing to watch. Some people get sucked in to it, listening to the utter rubbish that the guy spews out continually. Turn the volume off and watch the dogs; I personally find it quite upsetting to watch what the poor things have to go through and how he totally ignores every calming signal and sign of appeasement or fear (TBH I don't even think he could recognise them in all fairness).

What Cesar preaches is pack/dominance theory; which as we all know has been debunked time and time again. 

An exploration of the theory, makes it clear how the use of dominance based interpretations of behaviour and techniques are so damaging:

The theory in itself claims to be based on observations of wolf "packs" in the wild. However these studies took place at a time when there was little or no research on the behaviour of canine/wolf (canis lupus) or domestic canines (canis familiaris) for that matter. These observations have since been identified as being highly flawed (scientists have found that wolves basically live in relatively small familial groups- the young leaving once mature to create their own social group. Not too dissimilar to humans in some ways!)

So forgetting for a moment that the supposed evidence in favour of dominance/pack theory is flawed: the theory itself seems to essentially claim that dogs live in packs ( of presumably unrelated individuals). In this "pack" there will be one "leader" or "alpha" dog which will assume the dominant position within the social group. At times youngsters may attempt to rise up in the ranks and take this position.
Now those of you that have multi-dog households are likely to have seen evidence of some kind of social hierachy amongst your dogs. For instance one dog may be slightly more strongwilled. Nevertheless to my mind, whilst there is certainly evidence to support these ideas, I know of various friends and family members who have several dogs who have found that one of their dogs may be strongwilled/bossy indoors whilst outside another of the dogs is the more outgoing in this way (just an example). Similarly some dogs show no such tendancies. Surely in such a cases there is no "alpha". To me this suggests an alternative idea; dogs are individuals & therefore their relationships with other dogs need to be examined on a very individual level. This explains why one dog might in a bossy manner with one dog and then play nicely with another or act very "respectfully".

To my mind, you cannot simply label dogs as "submissive" or "dominant" as many dominance enthusiasts, such as Millan insist. It is not that simple. In this sense the dominance approach seems to be very simplistic and painfully reductionist. 

All dogs are individuals and should be treated as such. As should every individual dog to dog relationship and interaction. 

Personally the idea of dogs being opportunists to me however does seems like a plausable explanation for the way that dogs will behave differently with different dogs. They do what they can get away with. They're not asserting their authority or dominance but they are behaving in the way that is accepted and/or reinforced in that situation/environment. This will vary from individual to individual.

The following video goes some way to explaining why the term "alpha" is incorrect:

HERE

Interspecies dominance (between dogs and humans)- again this concept is debunked time and time again (for all of the reasons above and more!). Dominance enthusiasts will harp on about a dog that sees himself as the pack leader (as the alpha) and therefore sees his humans as inferior. For this assumption to be valid, dogs would surely have to have some kind of perception of self; how else would they be able to perceive rank. I question why dogs would even want or indeed have any need to assert their authority on their human companions. Nevertheless the following extract from Canine Dominance Revisited by David Appleby explains the relevance of self-perception in dogs further:

 "However, scientists believe that dogs do not have a sense of self so it could be more accurate to say that it behaves with no inhibition and it is uninhibited behaviour that we interpret as dominant behaviour. Owners often describe a belief that their dog is trying to increase its status over them. This would require that their dog has a capacity for forward planning and to know how its behaviour affects the feelings and thoughts of others, which we believe they are not capable of. So the notion that the dog behaves with lesser or greater inhibition according to who it is interacting with and the value of the resource in question may be a better way of describing what is going on." 

In my opinion behaviours that dominance enthusiasts such as Millan attribute to dominance and status seeking ; take territorial behaviour (reactivity) for example, may be better explained as the dog responding to environmental cues about it's safety (often the owner needs to step in let the dog know that they'll handle the situation and take the weight off their shoulders). Similarly over zealous behaviours are more likely to be due to the dog not having being taught what IS acceptable behaviour rather than what isn't (by reinforcing the right stuff :-D).


 To me dominance theory seems like a very humanized theory. By choosing to misinterpret dog behaviour and describing them in terms of dominance I believe we are anthropomorphizing their behaviour. This is an injustice to our dogs. They are so more complex than what dominance/pack theory and Cesar Millan would have us believe.


In addition to Millan's insistence on dominance/pack theory, the other main issue I have with the show is the methods used. In short Millan painfully misdiagnoses their behaviour. It seems like in every show, every dog is "dominant" (you do get the occasional one that he describes as submissive and fearful and he'll then go and flood the poor thing in any case!). Once he has used his hideous methods on them (jabbing, grabbing, kicking, shocking, prong collars, checking, hitting,strangling, alpha rolling, pinning and thoroughly intimidating) the dogs will either:

• Be in a hysterical state where they react to very low level triggers or misdirect aggression (this is of course, after the poor things have tried every calming signal and avoidance measure in the book! IGNORED by him of course). These dogs he labels as DOMINANT (you see the poor owners' faces fall at this point)
• Totally shut down and show every appeasement signal in the book (which he naturally ignores or is unable to identify); he then classes them as "calm submissive!" (Brilliant! NOT)

Below is an example of a dog that become hyper-sensitive to triggers. Cesar floods him by putting him in a situation of close proximity to a feared trigger: the collie- staring at him from 2:50 onwards. Shadow just about copes with this, until Cesar kicks him at 2:55 which leads to Shadow lunging and biting Cesar's arm. The footage that follows makes me feel disgusted. Cesar begins to strangle him on the lead (he is wearing an Illusion collar (Millan's wife's own creation- it's horrible)- had previously been wearing a prong collar :( ). Then he eventually alpha rolls him, holding him around the upper part of the neck. 
Shadow turns blue!
A second example here of a bull dog that appears to have a resource guarding issue (he tries to wrestle with the dog to get the hose off it and gets bitten! DUH!). The clip is in Spanish though (I couldn't find one in English).
Bulldog


The following clips also reveals how show producers and Millan himself  deliberately set the dogs up to fail and rehearse the undesired behaviour (they appear to find it funny too :()
DW provoking the dog


So if you've got to here, well done!!! I went on a bit of rant there, which I didn't intend to but I guess I wanted to highlight just how damaging the show and the methods that Cesar Millan uses are. 


I wish that the UK would follow Italy's example on this one:
http://apbc.org.uk/blog/CM_Italy


In any case I certainly won't be watching the new series. 

(Rant over honest! :-) )

Great website for those with fearful or anxious dogs

Despite a rather messy few days, which I think my liver has just about recovered from (:-P), I thought I'd better update the blog.
We've been working hard on our mat work as well as a new trick which is a headshake that we are still working on shaping. All fun stuff. :)

My poor BC has been acting like a bit of a stress head recently (he can be quite easily stressed generally). One particular thing that he doesn't seem to like recently is his harness being put on- he'll run into another room, with his ears flat and tail low and look at it with wide eyes, showing all calming signals possible and being very appeasementish. He'll come to me for cheese and let me pop it on him (and then he run to the front door ready to go)- but he definitely isn't 100% happy with it. Alot of signs of conflict there. Think he wants to go out but isn't 100% with the harness. He has no anxiety about going out, as he'll happy run to me when I get his lead out and will go and sit by the door. He'll also fetch his lead (I taught this as a trick for fun). He used to show similar behaviours with his old harness, which I realised was uncomfortable for him to wear as it rubbed. This one certainly doesn't so my feeling is that this is rehearsed behaviour? 

I'm going to have a go at reconditioning his emotional response to it appearing and then gradually being put on. The harness will become a predictor for lots of cheese. I'll eventually raise the reinforcement criteria once he realises that harness appearance= food so that he has to offer me his neck/head to pop it on with, before he gets his cheese. I found a great video, featuring Jean Donaldson and her dog Buffy, which shows how she taught her dog Buffy to LOVE wearing her gentle leader headcollar.

"If you're good I'll let you wear the Gentle Leader!"

Watching and looking for the video, reminded me of a fantastic website that I found a while back which has been designed for people with fearful or anxious dogs, to give them ideas and resources to help their dogs. 

Just thought I'd share the link in case it's of interest to anyone.

www.fearfuldogs.com

Going to get my camera out and film some clicker work and post it on here at some point.

ETA: Our first harness clicker session from earlier today. What a transformation! It took him several minutes to decide to come towards the harness (I put it on the floor). At first I clicked him for just looking at it. Then I raised the criteria to sniffing and targeting of any sort (holding with mouth or just nose touching)- that way I could be 100% sure that he knew what the reinforcement was for. I free shaped the nose target behaviour through the loop of the harness! He even started offering me duration! All in one session! Clever collie :D
(I couldn't hold the clicker whilst filming so am just using a verbal marker in the vid)

VIDEO HERE

:)

First post and it's about pigeons?! Sorta

So I set up this blog a while back but never really got round to posting much on it. Now that I've got a bit of extra free time I thought I'd better get my bum in gear!
I've recently been doing a bit of research and reading on teaching default type behaviours for arousal. Main reason being I've noticed that my BC who I do agility with has been getting more and more aroused at agility training, especially in more enclosed indoor type areas. Now arousal is good in one way. If he wasn't aroused I wouldn't get much sense out of him when I wanted to run him in agility. However when he's got to the point that he will bark often as he awaits his turn to run, is watching all the other dogs run a bit too intently for my liking and is so aroused that he isn't able to properly focus on doing tricks (that I originally used as a distraction to keep him busy) or focus; I think it's time to be doing some work on this. I can currently keep him focussed by constantly throwing food on to the floor for him to "find"; if I don't feed him at this rate he goes over threshold again and cannot focus. I've been reading the fantastic book, Control Unleashed by Leslie McDevitt, in which she suggests teaching placement cues and building value on a special mat or bed where the dog can lay for duration and remain in control of their arousal. I've been working on clicker training placement cues for the mat, which he has cracked now. The main criterion that I am yet to reinforce is serious duration. I've done a bit of research on methods and plans for clicker training and proofing more serious durations to a behaviour and the key concept that came up was the 300 peck method. I done some work already with this method to teach my BC loose lead walking and I recently wrote a short article on it, as much as anything to make sense of it for myself. 
So here it is:

300 Peck Pigeon- the key to teaching duration and distance.

Hope the title got your attention! I expect you are thinking “I thought this was about training dogs?” Well it is. The 300 peck pigeon method is what many trainers will use to teach dogs to perform behaviours for serious durations and distances.

Ordinarily when we are clicker training a new behaviour that requires a moderate duration, we will only raise the duration criteria for a click and reward once the dog can perform the behaviour itself to a hand or verbal cue. To do so we simply click the dog for performing the behaviour for a second more and we repeat this a few times. Then we raise the criteria and ask for 2 or 3 seconds and reward and so on. This technique is similar in many ways to the 300 peck technique; however where they differ is that the 300 peck technique deals with long durations (and distances), meaning that by the end of the process a 300 peck trained behaviour may only need reinforcing after 10 minutes! Sounds cool doesn’t it?

The 300 peck method gets its name from an experiment in which a scientist was investigating how to build long duration variable schedules. She taught pigeons to peck a bar repeatedly hundreds of times (300 on average to be exact) before they were reinforced; meaning that the pigeons had a VR schedule of 300. So a VR schedule of 300 is impressive right, but how did she do it? The key to her success in teaching this behaviour was that she very gradually extended the number of pecks they had to give her before they were reinforced. You may wonder when we would want our dogs to perform behaviour for such a huge duration. The answer is, when you think about it, everyday! Take heelwork for example. You don’t want to have to be constantly reinforcing the dog after a couple of footsteps. Apart from anything else it means the dog doesn’t get any time to act like a dog. What you want is a dog that will happily walk on a loose lead by your side, taking in all the sights, sniffing and engaging with his surroundings (and you) in a relaxed controlled way. Similarly those of you who do any kind of competitive obedience work with your dog will have taught your dog to perform long duration and distance stays. Agility trainers will also teach their dogs a long distance and/or duration start line wait. These behaviours can both benefit from training with the 300 peck approach.

So how does it work? As I already mentioned the secret to creating duration and distance is to very slowly increase the time or distance before the dog is reinforced. According to the 300 peck method we do so in small steps of one until we reach 300. You may choose to count out loud if this helps you. To take for example loose lead walking; the only criteria we have for reinforcement is that the lead must be slack whilst we are counting for the dog to be clicked and treated. So in essence the dog can be sniffing, looking elsewhere or looking at you- that doesn’t matter. So long as the lead is slack when we reach our first “count” the dog can be reinforced. If then, the lead isn’t slack; we re-set our counter and start over again at 1. So the dog maintains a loose lead for 1 step on a walk- click, treat. The dog maintains a loose lead for 2 steps- click, treat. The dog maintains a loose lead for 3 steps- click, treat and so on until you reach 300.

Patience is the key with this method. At first you may have to reset the count several times before you can reach 10 steps but this isn’t a bad thing. Remember after all that clicker trained dogs learn just as much from withdrawal of the click as they do for constant reinforcement. The dog thinks “Oh… Why didn’t I get my click there?” and they’ll work on offering the right criteria. Once the dog grasps the idea of the 300 peck game you may be able to progress quicker.

 

Similarly with a 300 peck trained stay/wait behaviour you may choose to do this in seconds or steps. Depending on whether you are teaching a long distance/duration stay or just duration. For agility for example you ideally want the dog to wait on the start-line once you have set him up no matter where you put yourself on the course. You need distance so that, especially with fast dogs you can buy yourself time to be well ahead of him before he even starts; and duration so that he’ll wait and not break until you release him. When you are teaching this behaviour, impulse/ stimulus control comes into play especially if you are expecting the dog to wait for duration and distance in a high arousal environment (like a show). This is something I am working on with my dog, Dash (I’ll hopefully add some videos of our work on this at some point).  You basically want a dog that’ll stay where you’ve asked him until released even if a bomb were to go off behind him! That’s what we want to aim for anyway (although on a practical note we definitely don’t want to be doing stays in minefields!). To teach this provided your dog can do a moderate distance wait/stay with low or no other environmental variables (distractions), in order to help the dog succeed at each stage, we keep the criteria low. If we are teaching the dog a combined long distance and duration stay this means we are only going to mark and reward one criteria type at a time (so perhaps duration first and then distance). So as with the loose lead walking, you will first click and reward the dog for 1 second. Then raise the criteria to 2 seconds, click and reward. Then 3 seconds, click and reward and so on until you reach 300. If the dog breaks at any point, you re-set the count and start back at 1. Once you have a dog that will stay for long durations you want to make sure you have properly proofed the duration criteria so that he will happily perform a long duration stay in the park, at home, in the garden, at your training field etc. Then you can start to 300 peck train the distance criteria in the same way as described above. One step away, click and treat. Two steps, click and treat. Three steps, click and treat and continue until you reach 300. If the dog slips up at any point, re-set the counter. And as with the duration criteria you can proof the distance criteria (separately initially) and then eventually (and only when the dog has reached a VR schedule of 300 in each) along with the duration criteria in a variety of settings and then add distractions. When you add distractions (i.e. a ball being thrown at a distance), you need to not only start with relatively low level distractions but also raise the distance and duration criteria for the click. It would be unrealistic and setting the dog up to fail for example if you asked the dog to stay and only waited to reinforce a VR schedule of 300. It ain’t gonna happen! Whenever you change one criterion/variable in training you need to keep the others constant or decrease them. So if you were to introduce a ball being rolled along the floor whilst the dog is asked to stay (assuming the dog is easily aroused by movement) you would need to lower the distance criteria to one step and the duration to a second for the dog to be clicked and then increase the criteria accordingly. When proofing distractions it is best to work on duration and distance separately and then combine once the dogs have really successful at both.

The 300 peck method is great for teaching really reliable distance and duration behaviours. Even if the process seems like a bit of a struggle at first, patience is the key. It’s a fantastic tool in the dog training toolkit!

Link to the original article:

https://sites.google.com/site/lucysdogblog/300-peck-pigeon

So what I need to do is use the 300 peck method to add duration to the mat placement cue. As stated above, he'll only get clicked and reinforced for 1 second, then for 2 seconds, then for 3 seconds. If he moves from the mat I'll restart the counter. Once I've proofed duration, the plan would be to add distance and proof against distractions so that I can have him on his mat whilst I walk through a course at training, even if there are other dogs working at the other end of the arena. That's what we are aiming for anyway!