Wednesday, 6 September 2017

The Best Way to Train Cats is With Food

Using food alone is the quickest way to train cats to touch a target, according to this pilot study.

How to train cats, like this beautiful white cat with blue eyes
Photo: Esin Deniz (Shutterstock)


You can train cats to go up to a target and touch it with their nose. This in itself will be news to many people, but researchers at Massey University have investigated the best way to train cats to do this. It involves food.

There’s a lot of interest in training cats at the moment, not necessarily to perform obedience behaviours like sit and stay, but to help them in their daily lives. You can teach your cat to like going in their cat carrier so trips to the vet don’t have to begin with you getting scratched-up arms. And you can use positive reinforcement to help teach your cat where they are allowed to scratch (along with provision of the right scratching post, of course).

Erin Willson et al picked the behaviour of touching a red wand target with the nose, and set about training 9 cats to do this. They divided them into three groups: one that was rewarded with food alone, one that used a bridging stimulus (a beep followed by the food reward), and one that used a secondary reinforcer only (a beep – previously associated with food – but no food).

The first two of these conditions will be familiar to dog trainers who use positive reinforcement, since they equate to the use of food only or to click-plus-treat. The last condition may have some of you thinking back to an interesting talk by Simon Gadbois at SPARCS about the clicker and the emotions of seeking vs liking (you can read a nice summary and discussion on Patricia McConnell’s blog).

The scientists concluded that both food alone and the bridging stimulus (beep plus food) worked, but that food alone was faster. The secondary reinforcer only (beep but no food) did not work. In fact cats in this group began scratching and biting the experimenter.

This is only a small study so there weren’t really enough cats to draw firm conclusions about training methods.  Nonetheless the results are very interesting, and it is really nice to see cat training getting the attention of researchers.

The study used a Treat & Train, which is an automatic food dispenser. The red wand target comes with the machine.

12 cats from the university’s feline unit took part in the study. They were aged 2 to 12 years.

Use food to train cats, like this calico cat sitting pretty for a treat
Photo: Kristi Blokhin (Shutterstock)


3 of them took part in what is called an extinction procedure. First they were taught that the beep from the Treat and Train machine meant food was about to arrive. The next day they heard the beeps without any food arriving, to see how long it would take for their response to the beep to extinguish (in other words, until they stopped approaching for food). The median response was 11 trials. This was important information for one of the conditions in the experiment.

9 cats were trained to nose-touch the target using a standardized plan. This was a shaping procedure, so cats were initially rewarded for just looking at the target, then for getting progressively closer until eventually they were expected to touch it with their nose.

The cats were divided into three groups. A beep is meaningless to a cat, so two of the groups (the bridge group and the secondary reinforcer group) were taught to associate the beep from the machine with food. Food is a primary reinforcer because it naturally has value to cats.

The food-only group got to hang out with the machine without any beeping, so that time with the machine would not be a factor.

During the training sessions, the food-only group was rewarded with food from the Treat and Train whenever they performed the correct behaviour. The machine was set up not to beep.

For the bridge group, the beep of the machine was used as a bridge, something that marks the right behaviour and fills the time until the arrival of food. In this condition, the beep is always followed by food. When the cat performed the correct behaviour, the machine beeped, and then food arrived.

For the last group, the beep was used as a secondary reinforcer. In other words, when cats performed the right behaviour, they heard the beep but did not get food. These cats were given some additional beep-food pairings to maintain this association and prevent its extinction (that’s why the first part of the experiment was important).

All of the cats in the food-only group and the bridging stimulus group (beep plus food) learned the behaviour. The group reinforced with food-only was faster at learning the task, but took the same amount of trials as the bridge group.

None of the cats in the secondary reinforcer (beep-only) group learned it. As mentioned above, the cats in this group began to scratch and bite the experimenter. Perhaps they were frustrated that they could not figure out how to get the food. (This reminds me of the study of the Eureka effect in dogs, where dogs became reluctant to enter the experimental area when they could not make the reward happen).

So if you are planning to train a cat, you should use food. (Incidentally, food is also important when training dogs).

The experimenters used a piece of Hill’s kibble as the reward.  If you’re training a cat at home, you might find other kinds of food more motivating; see my interview with Sarah Ellis on The Trainable Cat for some ideas.

The results of this study are broadly in line with Chiandetti et al’s (2016) dog training study, which found no difference between use of food only, clicker-plus-food, or verbal-marker-plus food (that study did not test a secondary-reinforcer-only option).

The cats were assigned to groups in the order they happened to participate, and it turned out three older cats were assigned to the secondary-reinforcer-only group. We don’t know if age and gender of the cats would make any difference to trainability and this would be another topic for future research.

The authors conclude,
“the use of a primary reinforcer, alone, or a bridging stimulus (followed by a primary reinforcer) appeared to be efficacious for training cats to perform a novel task. However, the primary reinforcer, alone, may be a more time efficient method. The use of a secondary reinforcer, alone, may not be efficacious.”

Incidentally, learning to touch a target with the nose may seem like a trick, but it has its uses. Some people train their dogs to touch a target (such as their hand) and hold in place. It’s called a stationing behaviour because it keeps the dog still at a station, and can be useful during veterinary examinations.

This is a fascinating study and I hope to see lots more research on cat training in the future.

Have you ever tried to train a cat? If so, how did it go?


Further Reading on Cat Training


If you want to know more about how to train cats, you might like these books.

The Trainable Cat: A Practical Guide to Making Life Happier for You and Your Cat by John Bradshaw and Sarah Ellis. (This is a must-read for all cat owners).
Clicker Training for Cats by Karen Pryor - also available as part of a kit: Karen Pryor, Getting Started: Clicker Training for Cats Kit.
Cat Training in 10 Minutes by Miriam Fields Babineau.
Trick Training for Cats by Christine Hauschild.

Reference
Willson, E. K., Stratton, R. B., Bolwell, C. F., & Stafford, K. J. (2017). Comparison of positive reinforcement training in cats: a pilot study. Journal of Veterinary Behavior: Clinical Applications and Research.

Companion Animal Psychology is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com.

3 comments:

  1. My first cat was finicky about food. Maybe that's why I had so-so results with training her. I've always said that she complied only out of love. ;-)

    Two of my current cats are food-motivated and have been easy to train. My third is shy and on a prescription diet. I have only so-so results with training her.

    Yes, it's very nice to see studies about training cats! We can hope this is a sign of positive changes to come.

    ReplyDelete
  2. Tucked away in the article is the fact that only 12 of 36 cats were even sufficiently food motivated to train. That's a lesson to learn. The other tip is that younger cats (seemed) to learn quicker.

    ReplyDelete
  3. In my experience working with kittens, some of them are way more motivated by play, at least in the beginning of the training programm.
    Following my programm of Kitten Kindergarten for a few years now, I think that the most important step is teaching a kitten that it is teached by humans, so they learn to cooperate with humans.

    ReplyDelete

Companion Animal Psychology is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.ca and Amazon.co.uk. (privacy policy)

Amazon