Apparently there is an article comparing the merits of these in a current/recent (?) Equine Vet Journal. And conclusion is that desensitisation is the most effective.

I don't know what criteria they used but I wouldn't agree with that and would really like to see the article. Has anyone come across it??

Thanks
Catherine

Thanks for that Catherine, sounds intriguing. I'll get my vet student friend to find it and photocopy it for me.

I wonder if they did the counter conditioning procedure without desnsitising first? We had a discussion on just that on the Southampton Uni course. Depending on what is used to counter condidition, the procedure can be just too arousing. And given that frustration can give way to fear due to functional similarities... (ref. is Chapter 10 'Fear and Frustration'in 'The psychology of fear and stress' by Jeffrey Alan Gray)

Guess i shouldn't make presumptions without reading the paper first though!

All the best

Jenni

I'd like to see this too- Suzanne can't remember is this one you publish?! Jenni if you get electronic format please can you email to me? Hope you're well!
Helen

Guess i shouldn't make presumptions without reading the paper first though!

Tee hee - I'm struggling with that one too!!

Nope, don't think its published by Elsevier, sadly.

I have a copy sitting next to me! (I'm in the James Herriot Library of Glasgow Uni Vet School right now...)


Training methods for horses: habituation to a frightening stimulus

J.W.Christensen, M.Rundgren & K.Olsson

(Danish Institute of Agricultural Sciences)

Equine Vet J. (2006) 38: 439-443

Summary

Reasons for performing study: Responses of horses in
frightening situations are important for both equine and
human safety. Considerable scientific interest has been
shown in development of reactivity tests, but little effort has
been dedicated to the development of appropriate training
methods for reducing fearfulness.

Objectives: To inestigate which of 3 different training methods
(habituation, desensitisation and counter-conditioning) was
most effective in teaching horses to react calmly in a
potentially frightening situation.

Hypotheses: 1) Horses are able to generalise about the test
stimulus such that, once familiar withe the test stimulus in one
situation, it appears less frightening and elicits a reduced
response even when the stimulus intensity is increased or the
stimulus is presented differently; and 2) alternative methods
such a desensitisation and counter-conditioning would be
more efficient that a classic habituation approach.

Methods: Twenty-seven naive 2-year-old Danish Warmblood
stallions were trained according to 3 different methods, based
on classical learning theory: 1) horses (n=9) were exposed to
the full stimulus (a moving, white nylon bag, 1.2 x 0.75 m) in 5
daily training sessions until they met a prededined habituation
criterion (habituation); 2) horses (n=9) were introduced
gradually to the stimulus and habituated to each step before
the full stimulus was applied (desensitisation); 3) horses (n=9)
were trained to associate the stimulus with a positive reward
before being exposed to the full stimulus (counter-
conditioning). Each horse received 5 training sessions of 3 min
per day. Heart rate and behavioural responses were recorded.

Results: Horses trained with the desensitisation method
showed fewer flight responses in total and needed fewe
training sessions to learn to react calmly to test stimuli.
Variations in heart rate persisted even when behavioural
responses had ceased. In addition, all horses on the
desensitisation method eventually habituated to the test
stimulus whereas some horses on the other methods did not.

Conclusions and potential relevance: Desensitisation appeared
to be the most effective training method for horses in
frightening situations. Further research is needed in order to
investigate the role of positive reinforcement, such as
offering food, in the training of horses.



Now I am going to read the rest of the paper!

Cheers Francis!

Do the authors reference the following text at all:

Domjam, M. (2003) The principles of learning and behaviour 5th Ed. Thompson and Wadsworth

This is the book that informed us at Southampton of the pitfalls of implementing counter-conditioning before desensitisation is complete. Have to confess to not actually having read it, only having it read out to me and the rest of the group during an intensive debate about how to rehabilitate a German Shepherd who was afraid of cars! (I bought the other learning theory book on the list!).

The idea is that desensitisation is performed to get the animal to a more neutral state in the context of the scarey stimulus. But as desensitisation does not generalise well counter-conditioning is then implemented to train the animal to generalise.

It's good to see some formal study out there to test this as what I've outlines above comes from theory rather than tried and tested theory that the Danish researchers were doing. The follow on study would be to test the effects of each treatment in a novel location and to add treatment 4, desensitisation followed by counter-conditioning?

All the best

Jenni

All the best

Jenni

You're welcome, Jenni!

Just after I posted the above I remembered I had to go and buy some haylage, and so far I have only glanced through the paper.

Anyway... They don't cite Domjam, but here's what they say about counter-conditioning:

During the first training session, the usual feed container was removed and a bag similar to the test stimulus was scattered with feed and placed with the folded test bag, encouring the horses to approach the test stimulus and feed there. Whenever a horse willingly fed on a bag a second step was introduced, with the full stimulus applied while the horse was feeding next to it. Once habituated to this step, the usual feed container was reintroduced and the horse exposed to the full stimulus while feeding from the feed container as in [the other two methods].
...
Counter-conditioning [...] appeared from our results to be the training method that required the highest number of training sessions. Feeding from frightening objects is a method often applied in police horse training [...]. However, police horse are typically left with a task in an enclosure, e.g. food is placed in the middle of a large piece of plastic and the horse has to step on the plastic in order to get the food. In the way the horse is able to train itself and has plenty of time to solve the problem (obtain the food). Gough (1999) succeeded in reducing fear responses towards clippers while feeding during a 2 week period. However, the study was carried out in only a small number of animals with no control group, and it therefore possible that the reduction in response may have been caused by general habituation. Further research is necessary in order to identify possible benefits of the use of positive reinforcement and counter-conditioning in relation to fear reactions in horses, as well as whether habituation during feeding is applicable to the nonfeeding context.

That's pretty much all there was specific to CC.

Finally getting to my backlog of threads!

Francis thank you so much for posting all of that, it was really interesting to see what they actually did.

I have a slightly different take on this from Jenni because I hadn't heard about the need for systematic desensitisation before counter-conditioning. In fact, what really struck me about this article was that it just didn't sound like what I would call counter-conditioning. And I think that the reason for this comes down to the difference between working with science alone, and working with feel, backed up with a working understanding of some of the science.

Based on what the paper said about CC, it just doesn't sound like the horses were counter-conditioned to the stimulus. Yes the trainers fed the horses in the presence of the stimulus but that is not the same thing. For a stimulus to be truly counter-conditioned surely the horse needs to see it as a positive thing? And in my experience, when horses view something as a positive stimulus they make a special effort to get to it. The horses in this experiment were just habituating to the stimulus in the presence of food which is not the same thing. They weren't actively seeking the stimulus as they would have been if the CC was complete.

When I used CC to teach Jak to walk over a tarpaulin I didn't just feed him on it or near it until he could tolerate it. I shaped it using rewards at each stage so that he could associate every step along the way with something positive. And I didn't just rely on my assumption that he was finding it positive. I actually allowed him to actively tell me that he wanted to walk across the tarpaulin because he knew it represented treats or scratches.

Thinking about it I suppose I do use a combination of CC and SD and never CC by itself. Afterall, in order for a stimulus to be perceived as positive, rather than negative as previously, it has to proceed through a neutral phase. (I think! Kind of thinking aloud here...) Had just never thought about it in this vocabulary before.

I don't pretent that CC is a quick process - it definitely isn't. And so if by "most effective" the authors of this work meant "quickest" then yes, I would probably agree that SD is quicker (although also runs the risk of partial flooding if you are not working with sufficient "feel" to gauge the size of the steps in the process). But I would also say that it's not quicker by very much and the lessons taught by CC are much longer-lasting. I only need to get the tarpaulin out once a year or so and Jak will still deliberately go and stand on it to earn a treat (even if, or maybe particularly if, I'm actually trying to work on a different, more challenging exercise!)

Catherine