What's Something Universally Mildly Aversive?

[corvus]

I need a brainstorm. You should all be excited, because this is about the only time I'm ever going to ask for punishment ideas!

I need something that is universally aversive, but only mildly so. Enough to create a very mild avoidance reaction, but one that is resistant to extinction/habituation. It has to be environmental, as in can't be associated with the handler. Tall order.

As an example, something that tastes bitter.

[Jaxx'sBuddy]

what about something that smells bad or that shoots out a stream of compressed air or water when the animal moves past it

[Staranais]

I need a brainstorm. You should all be excited, because this is about the only time I'm ever going to ask for punishment ideas!

LOL.

Very low stim e collar? Fits all your criteria nicely, if you can borrow one.

[SkySoaringMagpie]

What for?

[Kavik]

Sharp, sudden noise?

[corvus]

E-collar is out for various reasons such as cost and ethics. The problem is a curly one. I need to be able to teach any dog that one stimuli means good stuff and another means bad stuff. There are several twists. The "good" and "bad" has to be pretty standard so it will be experienced much the same by all dogs. The "good" and "bad" need to be delivered automatically so we take variable timing out of the equation. I suspect it has to be something that has no other cues associated with it. I don't want dogs to be able to back out of the "bad" before experiencing it because they are using a more reliable cue than the signal I give them. This is my problem with bitter food. I feel like the dogs are going to learn they can trigger the "bad" but avoid experiencing it because they will be able to smell if it's going to taste yuck and back out. Unless I can find something to spike it with that tastes bad but doesn't have much of a smell.

A jet of air will probably not be standard enough. I'm concerned that a loud noise or bad smell would be habituated too quite quickly.

I could possibly use something that tastes SO revolting that after the first taste the smell itself becomes quite aversive....

[Staranais]

E-collar is out for various reasons such as cost and ethics.

Fair enough, I don't see any ethical issue myself, I think e-collar can be one of the gentlest & most controllable aversives, but I imagine the cost would be prohibitive if you couldn't borrow one.

So, why don't you just use a conditioned negative punisher? That would be fairly simple to do. & probably kinder than a nasty taste or scary sound.

& what exactly are you trying to teach the dog to do? That would surely affect which aversives were likely to be useful.

My dog apparently finds the smell of wasabi a little aversive after stealing a mouthful at one point. She also finds the sound of a water pistol loading to be very aversive after she got a few surprise squirts for cat chasing.

[corvus]

E-collar is out for various reasons such as cost and ethics.

Fair enough, I don't see any ethical issue myself, I think e-collar can be one of the gentlest & most controllable aversives, but I imagine the cost would be prohibitive if you couldn't borrow one.

It's for a scientific study, so it has to be approved by an ethics committee, which means I would have to explain why I haven't tried a less controversial aversive. And then people have to agree to let me put them on their dogs. And then I have to teach the dogs not to be collar wise. And on top of all that, I'm then in the dubious position of being a dog welfare researcher using e-collars, which is not going to go down well with the people that will be peer reviewing my work. Anyway, it's a bit complicated and I would need a bunch of collars, which I can't afford.

So, why don't you just use a conditioned negative punisher? That would be fairly simple to do. & probably kinder than a nasty taste or scary sound.

Ultimately, the signal for "something bad" is supposed to be something of a conditioned punisher. The dog should see/hear the signal and go "that means something bad - I'm gonna avoid the place where the bad stuff comes from." I'm trying to keep it as simple as possible because an extra layer of training means more time with the dog, longer to reach criteria for testing, and another possible confounding factor.

ETA I should point out that the place where bad stuff comes from is the same as the place where good stuff comes from. The dog has to make the choice whether to approach that station depending on whether they think it will result in good stuff or bad stuff.

Edited August 7, 2010 by corvus

[Staranais]

Fair enough, I don't see any ethical issue myself, I think e-collar can be one of the gentlest & most controllable aversives, but I imagine the cost would be prohibitive if you couldn't borrow one.

It's for a scientific study, so it has to be approved by an ethics committee, which means I would have to explain why I haven't tried a less controversial aversive. And then people have to agree to let me put them on their dogs. And then I have to teach the dogs not to be collar wise. And on top of all that, I'm then in the dubious position of being a dog welfare researcher using e-collars, which is not going to go down well with the people that will be peer reviewing my work. Anyway, it's a bit complicated and I would need a bunch of collars, which I can't afford.

Ah, controversial I can certainly understand.

I suspect you could do a conditioned P- by repeatedly removing something nice from the dog just after giving a signal. Perhaps Aidan would have some good ideas how to go about doing that?

But possibly a conditioned P+ may be easier to train. In which case I guess you'll end up with a dog that feels mild anxiety when it hears the conditioned sound. Is that the idea?

[corvus]

I think mild anxiety would be ideal because it's unlikely to deteriorate over time and would induce the dog to avoid the station rather than come over just to see if something good came out. The tricky thing is I'm trying to measure a bias that is adaptive. So the consequence for an incorrect judgement should be a tiny bit bad. If it's neutral, then the dog can afford to be optimistic because it can only be good or neutral.

But having said that, in emotional terms it's possible that "neutral" is sometimes negative. E.g. a neutral expression on another animal might be bad news.

ETA The original idea was food versus, say, a rock or piece of metal. Something that is entirely disappointing could be aversive enough.

Edited August 7, 2010 by corvus

[corvus]

We moved to liquid rewards because they are easier to manipulate and standardise in relation to the weight of the dog. Oo, maybe parrafin oil or something like that. Although dogs will lap up parrafin because it doesn't taste bad, just not at all.

[Erny]

Not sure exactly what you're trying to do (ie what behaviour you'll be subjecting to aversion) but when I adopted my first RR (Avatar) girl (bhcs) she had a habit of getting up on my bed when I wasn't around. I didn't really want to have to always close the bedroom door after me, but because it was a waterbed and her nails were more like talons (never could get them down as short as I think they should have been), I needed to do something to teach her.

For her - she absolutely abhored squeakies. So I purchased a number of cheap ones from the Reject shop and I set them up under my doona cover. When she got up on the bed, she'd squeak them. Two or three times of this was enough for her to never get up on my bed again.

Most dogs love squeakies, so I'm not suggesting this. But some sort of aversive noise that sets off automatically when you're not around might work. It depends on the dogs used though.

If you use sound aversion, be careful - with the wrong dog and the wrong noise, you risk sensitisation which could have a carry over affect to other things in the dog's normal life.

Not to make it an argument, but I hate that your overseers oppose equipment like the e-collar so much that all who come after them aren't comfortable to use them in their presence for fear of vilification (and not great learning for all their students who themselves will become tomorrow's teachers) which can be taught well without risking pairing or sensitisation, yet might agree to subject dogs to things such as sound aversion.

Edited August 7, 2010 by Erny

[Aidan3]

Not to make it an argument, but I hate that your overseers oppose equipment like the e-collar so much that all who come after them aren't comfortable to use them in their presence for fear of vilification (and not great learning for all their students who themselves will become tomorrow's teachers) which can be taught well without risking pairing or sensitisation, yet might agree to subject dogs to things such as sound aversion.

It would depend on the committee, the subjects and the purposes of the study but I would be very surprised if sound aversion were accepted. Shock is used a lot but it depends on the context and as I understand it these are volunteer pet dogs so there are too many issues. Corvus, is your ethical committee accessible? Someone from the committee might make a suggestion, they would have plenty of background to draw from. Water spray would be my suggestion. Alternatively, I've not met many dogs who like the taste of orange, and I can't see anyone getting upset about giving dogs pieces of orange or orange juice.

[corvus]

Erny, it's not quite as you have painted it. From a purely scientific standpoint I might say "Gee, it's a shame we can't use shock because it would be perfect", but as far as I'm concerned it would be foolish to approach it from a purely scientific standpoint because I don't live and operate in a vacuum. This is why we have ethics committees, and it's why they have to include scientists, veterinarians, lay people, and welfare representatives. Personally, I would not want to conduct behavioural research that was not considered ethical by any one of those groups of people. It would be damaging to my scientific standing, and as a scientist that's about all you've got. This is where we are at now and it's more important that my methods are ethically acceptable to the majoriy of society RIGHT NOW than that I use the easiest form of aversive.

As for academics in the field, I think that they know very well how useful something like an e-collar may be, but that doesn't change the fact that they are essentially the face of science when working with the public and their dogs. I certainly wouldn't want to be seen to be advocating the use of e-collars whether they are useful or not, because what is useful for a trained professional can be downright cruel in the hands of the average dog owner. The issue is certainly political, but it's not as simple as not liking a tool and therefore frowning at anyone who wants to use it. IMO, as long as I'm wearing the hat of an expert I have a responsibility to consider what my every word and action might mean for the welfare of dogs in general. I'm never going to lie, but boy am I going to be careful about directly or indirectly supporting methods or tools that I believe require a careful and knowledgeable approach.

Aidan, I haven't put in for approval yet, but I believe there are people we can ask before submitting the forms who can tell us if it's likely to be accepted and give us suggestions. Otherwise it gets knocked back with comments and we adjust the protocol and re-submit. The former is much less time consuming, though! Orange juice might work. If any dog will like it Erik will, seeing as he's a fiend for fruit and vegies, so I can try him. I'm not so sure about water spray, as I can imagine it being considerably more aversive to some dogs than others.

Edited August 8, 2010 by corvus

[Dogs rock]

Does the aversive have to be instantaneous? If not, perhaps you could have the same food in both choices, but perhaps use some ipecac syrup on the 'aversive' choice (ipecac syrup can be used to induce vomitting). Of course, I understand this may be tricky with ethics, but if you could perhaps find out how much ipecac to use to induce nausea but not vomitting they might accept it. I don't know much about the syrup myself, so of course you'd have to look into it to check all of the effects it has etc, or else perhaps there is something else that can induce short-term nausea.

Then you would just need to have something 'different' about the two choices for the dog to make the association (such as a bucket present next to one but not the other, as a random example). Is the choice offered simultaneously (eg Y-maze) or separately? (This is not at all relevant to finding something aversive, I'm just curious). How many training trials? How many 'real' trials?

I think generally it would be hard to find anything that is universally aversive that doesn't work on the dog's natural instincts and how they relate to survival. Particularly if you are using pet dogs (ie from different backgrounds), the amount of variation in aversiveness of sprays, sounds, odours, tastes etc would no doubt be vast. However, if you use something that relates directly to survival (eg 'makes me feel sick, not going there again') I think you are more likely to find a 'universal' aversive. Maybe...

[JulesP]

Victoria Stilwell (Its Me or the Dog) used to use a loud noise as an aversive. I seem to remember her having a gadget that let off a loud siren when a dog went near it. Think it was for a dog that was counter surfing.

There are also high frequency sound gadgets to keep possums away.

[Staranais]

Erny, it's not quite as you have painted it. From a purely scientific standpoint I might say "Gee, it's a shame we can't use shock because it would be perfect", but as far as I'm concerned it would be foolish to approach it from a purely scientific standpoint because I don't live and operate in a vacuum. This is why we have ethics committees, and it's why they have to include scientists, veterinarians, lay people, and welfare representatives. Personally, I would not want to conduct behavioural research that was not considered ethical by any one of those groups of people. It would be damaging to my scientific standing, and as a scientist that's about all you've got. This is where we are at now and it's more important that my methods are ethically acceptable to the majoriy of society RIGHT NOW than that I use the easiest form of aversive.

I can see both of your points.

I agree with Erny that it's a damn shame that experts are perpetuating the idea that e collars are somehow inherently nastier than other types of aversive. It all too easily becomes a vicious and illogical cycle. "The ethics committee don't let us use e collars because they're so nasty" - "how do we know they're nasty" - "because the ethics committee don't allow us to use them". I see this attitude in vet school too, which is perhaps why it really bugs me. I'd rather see the experts leading the way & showing how these devices can be used ethically, instead of bowing to uninformed public pressure and putting them on the banned list.

However, I agree with Corvus that it makes sense not to use a controversial aversive in her study, in case it overshadows the point of her work (which isn't anything to do with e collars).

Back to the point, I don't think using a loud sound for this is very sensible, even if the ethics committee will allow it. I'd worry about you creating sound sensitivities or phobias in dogs that are already sensitive, or that are in a fear period. Those can generalise to other common noises, and take some work to get rid of.

I don't like the ipecac idea either, really (no offence intended to the poster who suggested it). Causing a dog to feel nausea (even if it doesn't actually vomit) is IMO not ethically defensible, since it's such a nasty feeling. A dog with nausea looks utterly miserable. I'd personally much rather be mildly stimmed by an e collar, even shocked, than made to feel nausea.

A bad taste, or a jet of air or water, seems much kinder to me & less likely to cause long term issues for the dogs. Not sure if every dog would find these aversive, but you could always try it out & scratch the dogs that didn't find it aversive?

Edited August 8, 2010 by Staranais

[corvus]

Does the aversive have to be instantaneous?

Well, not technically, but technically I don't have to use an aversive at all. It's a matter of reducing the training time, though. I think if the aversive is instantaneous I have the advantage of being able to use a milder aversive and probably cutting down on the training time and insuring against the wrong signal being learnt. For example, I wouldn't want them to learn to associate the food that made them feel sick with the apparatus itself instead of the signal, because the apparatus delivers the rewards as well.

Is the choice offered simultaneously (eg Y-maze) or separately? (This is not at all relevant to finding something aversive, I'm just curious).

I plan to offer them separately, in a way. Once they can discriminate between the signal for rewards and the signal for aversives I give them a new signal somewhere in between to see how they interpret it.

How many training trials?

To be determined by the pilot study.

How many 'real' trials?

Depends on whether I can solve a couple of problems. Possibly as little as one, but I hope I can solve some habituation/learning problems and make that "as many as you like". Or at least more like half a dozen. I hope to follow some puppies through to adulthood over a period of about 2 1/2 years and retest them several times.

I think generally it would be hard to find anything that is universally aversive that doesn't work on the dog's natural instincts and how they relate to survival. Particularly if you are using pet dogs (ie from different backgrounds), the amount of variation in aversiveness of sprays, sounds, odours, tastes etc would no doubt be vast. However, if you use something that relates directly to survival (eg 'makes me feel sick, not going there again') I think you are more likely to find a 'universal' aversive. Maybe...

With rats they often use citric acid and sugar because no rat likes citric acid and all rats love a sucrose solution. There are tastes that I think would be universally aversive to a dog without making them sick. Like the bitter spray you can buy to deter puppies from chewing items.

[Aidan3]

I don't know about ethologists and veterinary scientists, but shock is used extensively in psychology and animal behaviour studies. We do know that shock can be nasty, that's why it works. Whether it causes distress or not depends on a couple of factors, how much is used and how it is used, not the fact that it is used. Ethics is a science in itself, going back to the Nuremberg trials. Nothing is ruled for no reason or for circular reasoning. I asked the chair of the ethics committee at UTas whether Watson's "Little Albert" or the Milgram experiments would pass ethics today, she told me "possibly, it depends", whereas I have asked the same question of lecturers and they have said a flat "no". (I must sound like a bit of a deviant, eh?)

[Aidan3]

...and all rats love a sucrose solution.

... unless they are suffering depression, which usually only happens when we shock them and they can't do anything to avoid it.

(ETA: several times, not just the once)

Edited August 8, 2010 by Aidan2