Keiko's Journey

On the 9th May 2018, I made a little video of my practice with Keiko using an emergency stop. I was really proud of her and so shared it on some of the training pages I frequent. The response was huge, it was shared in Canada, Australia, France, it went global very quickly and I was overwhelmed by trainers worldwide praising it. There are probably 100’s of videos of dogs doing the same thing, so what made this one different? Keiko has been trained using positive reinforcement. She showed exactly the type of behaviour ecollar, shock and adversive trainers say cannot be achieved by positive training. I have been asked a lot of questions, how did we get there? How long did it take? How did we proof it? So I’ve combined some of my answers about the training here in a more readable (I hope) format.

The What

The emergency stop is a new cue for her that we have been practising for about a month. First during recalls, building it up to stopping her halfway to her dinner bowl. The training was going really well and the day before I had successfully tested it stopping her in her tracks as she chased a bird. Ace! Also on the 8th (my birthday) I had picked up a shiny new phone, bought due to the excellent camera reviews. So the morning (middle of the night for some) of the 9th we headed out really early to test the new camera on the sunrise. On the way back home, I decided to test to see if I could capture on camera Keiko’s emergency stop as she chased some birds. It didn’t quite go to plan, as we crossed the bridge, the noise of my shoes on the ironwork scared them off. I worked instead on recall from easy distractions (something worth a good sniff) and some sit and down proofing. I release Keiko from a sit and saw her body language change, she had spotted something- rabbits! A real test of her emergency stop!


Let me explain, in 2016 Keiko was 14 months old, she gleefully and proudly caught a baby rabbit :'( while the bunny seemed physically intact it was in shock. I tried to return it to the burrow but it tipped over, as if its balance had gone, most likely concussed. I couldn't leave it alone and vulnerable so took it home (up my top for darkness and body heat) and called the SSPCA but unfortunately it died before they arrived.

Although this is a completely natural behaviour for a hunting breed dog, it’s not a necessary or wanted in a pet dog, with Akitas especially having such a bad reputation, I had to do something about it. So it set me off on a course to manage dealing with my high prey drive pooch.

The How


I started by keeping her on the lead (management) around prey, I didn’t want her practising the chasing behaviour that is reinforcing in itself. So I used lots of exposure to the trigger and classical conditioning to desensitise her to bunnies, so every time she saw one I would mark and reward her for simply looking at the rabbit. This is where I get a lot of people questioning the logic- ‘Rewarding a behaviour you don’t want?’ that does not make sense, does it, how can that be right?

It works by creating a conditioned response. Classical conditioning comes from Pavlov, maybe the new Bill & Ted film will feature him as ‘Dude with a bell’. His experiments showed that we can link things together to make positive associations. He would ring a bell before feeding the dogs, initially the ringing meant nothing- the dogs would either ignore it or bark at the noise, this was their unconditioned response. They soon began to realise the bell meant that food was coming, they made that association BELL = FOOD, and so the conditioned response when the bell rung was attentive salivating dogs. With Keiko I was replacing the bell with rabbits, her unconditioned response was to stare and stalk them anyway, I wanted to associate rabbits with something else- treats and fun with me.

It didn't take that long until every time she spotted one she would look back to me or away (disengagement) and that then became the moment I would mark for reward. It took around six months before I extended the trust and let her off leash again, BUT with high supervision, rewarding her attention on me in a high distraction situations.

If she became fixated with a hard stare I would get her attention back by either using a kissy sound, calling her name or ask for a 'leave'. As time moved on I stopped interfering when she fixated, I trusted that she knew she what I wanted of her, it became her choice and any voluntary disengagement from fixation was highly praised. I was asked about rewards ‘What is better than chasing rabbits?’ The rewards I used were praise, love, and a bit of kibble (she has a sensitive tummy so other treats were out the question). Keiko and I are a team, we train together, we play together, building a good relationship with your dog is the key to becoming better than a rabbit. When you communicate clearly you build up a partnership of trust.

Her breakfast comes out with me every morning and I use it as reinforcement for any voluntary nice behaviour, attention on me, nice heelwork, some tricks and games we play (touch, spin, bounce, weave) or training we do while out. Whatever is left goes in her bowl when we get home.

So problem fixed?

I dont' believe you can 'fix' prey drive, it's an inbuilt natural instinct. I'm also not sure we should be trying to eliminate it all together either. It's a welfare right that animals should be allowed to behave naturally. We do have to balance that right with keeping them safe and fitting in with our modern lives together. Their ancestors didn't have to deal with busy roads after all. We just have to find other outlets and ways of making their natural insticts work for us. For example, Squirrels are another challenge for us, I have started working on this one now too. We play a game 'Did you see' so when she spots something that might hit her prey drive instinct, I get all excited say 'Did you see' and start dishing out rewards (again just classical conditioning). Instead of chasing the squirrel I will give her something to chase instead.

So in conclusion....

As I said in the beginning the stop cue is a new one I have been teaching her, but impulse control exercises, the classical conditioning, the exposure to triggers to desensitise her, it's all training that has come together nicely over a period of time. Of course it helps hugely that she is an awesome dog :)

I do still question why she chose to chase that morning, even after all the work we had put in, some people will see it as a failure rather than a triumph of training. In other videos I’ve made with her she voluntarily disengages from the stalking and staring and in turn I hugely praise her doing so, and she did this time too, but I didn’t mark it as I needed her to be fully engaged and to run to effectively test the emergency stop. I’d like to think Keiko read my relaxed body language as permission to go, but we can’t read each others minds just yet. Either way I am very proud of all she has achieved.

The Video

This is the video that caused a fuss. Nothing spectacular, simply my dog being an Akita and listening to a pre-conditioned antecedant

Training Articles

The links below lead to the training pages. these pages are designed to help you train your dog based on positive reinforcement and clicker training