Before reading part 2 of this two-part series, please read part 1.
In Part 2, I will present an informal case study that applied the concepts discussed in Part 1. I desensitized and counterconditioned my dog, Zani, to an electronic beep, starting with a lower frequency that did not scare her and gradually increasing it back to the original frequency. I did this rather than starting with a quieter beep, because the beep was originally quiet and still scared her. I also adjusted the duration as part of the process.
Case Study: Successful DS/CC using frequency and duration to modify trigger sounds
Zani was a small hound/terrier mix weighing about 20 pounds. She was fun, smart, and brave enough to stand up to the bigger dogs in our house. I got her in 2009 when she was about a year old, and she passed away in the fall of 2020 from lymphoma.
Zani was sensitive to noise, which worsened as he grew older and was diagnosed by a veterinary behaviorist with clinical phonophobia: he was scared of digital beeps, whistles, the creaking of sports shoes on slippery floors and, worst of all, the calls of birds.
One of the sounds she was afraid of was the beep of her new bathroom scale, the sound whose waveform we showed in part 1. The beep was 0.15 seconds long and 3900 Hz. This was about the same frequency as the dishwasher beep, but interestingly, it didn’t bother her. Also, our previous bathroom scale beeped at a different frequency, which she didn’t mind. But the new scale certainly did. We noticed that the dishwasher beeped 1.4 seconds, so the scale sound was much shorter. Both of these sounds were quiet, so we speculated that it wasn’t a volume issue. If volume isn’t an issue, making a quieter version doesn’t necessarily help. As we suggested in our previous post, quiet sounds can be even scarier.
I created a series of digital sounds to get Zani used to the beeping of the bathroom scale. I started with a low frequency beep that didn’t scare her, then gradually increased the frequency, the cleanest counterconditioning I know of. When I first reached the original frequency, I lengthened the sound and as a final step shortened it again. So in the desensitization sound series, I changed the characteristics of both frequency and duration to return to the original, previously scary sound.
Again, the trigger sound was 3900Hz, 0.15 seconds. Here is the list of sounds I created to desensitize closer to the original sound. I used the software Boldness Create and modify sounds.
| sound # | pitch | interval |
|---|---|---|
| 1 | 1950Hz | 0.2 seconds |
| 2 | 2112 Hertz | 0.2 seconds |
| 3 | 2275 Hertz | 0.2 seconds |
| Four | 2437 Hertz | 0.2 seconds |
| Five | 3150Hz | 0.2 seconds |
| 6 | 3900Hz | 1.0 seconds |
| 7 | 3900Hz | 0.8 seconds |
| 8 | 3900Hz | 0.6 seconds |
| 9 | 3900Hz | 0.4 seconds |
| Ten | 3900Hz | 0.2 seconds |
| 11 | 3900Hz | 0.1 seconds |
Creating a clean pairing
The graph above shows the desensitization sequence. In the counterconditioning portion, every sound was immediately followed by Zani’s favorite food.
To pair sounds and food nicely, we needed to make sure that there were no other factors that predicted her eating the food. We made sure that none of the following factors predicted her eating the food:
- Open the fridge and take out food
- Food on the counter
- My position in the house
- Setting up or powering audio speakers on/off
- My laptop presence
- time
- Set up a video camera and record
This eliminated the typical method of taking out food, turning on the camera, sitting near the speaker with the dog and a plate of food or a bag of treats, and doing multiple exposures. I tried to make every exposure a surprise.
The most time consuming part was the food. The dog usually knows we’re going to do something because we go to the fridge and get the food out or pull out a bag of treats. To make the special food unpredictable, I put it out every day for a few days and placed it on a plate on the counter (out of reach of the dog). I did this before we did the sound exposure. I put it out every day whether we were doing the exposure or not. The goal was to make the smell of the food (and the act of me putting it out) not predict that Zani would eat the food.
The other thing that took the most planning was making sure I wasn’t always sitting in front of my computer when the sound would play. Sometimes I was sitting, but usually I was moving around the room. I pressed play on my computer and sent the sound wirelessly to a speaker (these days we use a handheld device – our dog sees us carrying it and is always looking at it and poking at it). I included a few seconds of lead time at the beginning of each sound recording so that I could press “play” and then perform a random activity.
Recording with the camera mounted on a tripod might also be a big tip. I always carried a tripod with my phone attached and had to turn the video on and off randomly during the day. I was always careful to start the video quite early for real exposure, so I could trim out a few minutes of unimportant footage (that won’t be in the video).
I did a maximum of two single exposures in a day, but usually just one, and not every day.
When I first tested the base tone to see if it scared Zani, I treated it as a planned exposure: I did all of the above, exposed her to the food smells for a few days beforehand, and of course, she got a special treat after playing the test tone.
All the exposures in the video were taken in the kitchen because it was easiest to film there. I also did exposures in other parts of the house. Also, dogs often behave as if they are waiting for something to happen. That is typical “kitchen” behavior. They would hang out on their mats while I was working in the kitchen or sitting at the table. They would look at me not necessarily because they were expecting a sound exposure. I think I was pretty successful in reducing the premonition. When the sound comes on, you will see a physical reaction from the dogs, often a moderate startle.
What would you do now?
Zani’s conditioning was a success, but I’ve learned enough since then that I can reduce the risk even further. The technical part now seems pretty crude to me.
The first sounds I made for her series were an octave below the trigger, i.e. 1950Hz versus 3900Hz. Exactly half the frequency. Now I start with sounds that aren’t very mathematically related; sounds an octave apart sound similar. The goal is to start with something. do not have They are very similar.
For most dogs, I find that gradually tapering off the sound on the first repetition rather than a hard attack can be more effective than fine-tuning the duration.
As I was watching the video back, I was surprised to see that I used at least two different foods. I wouldn’t do that now. I know that all the foods I used were of great value to her. There were also some crunchy foods that I think were Lakse Kronch fish snacks.
The tricky question was how many times to repeat each sound. The science of classical conditioning tells us that the first repetition is the best for conditioning. Do you think of each sound as a single conditioning target, or the series of sounds as a unit? I kept the repetitions low, only two or three at each level. I tried not to immediately say “I love this” on the first sound. I moved forward when Zani said “This is fun. Where’s the treat?” This strategy worked, but there may have been some luck involved.
In Poppen’s study (1970), rats typically needed 5-10 repetitions of each frequency to recover from fear conditioning, and more repetitions were needed as the frequency approached the original frequency. I didn’t intentionally increase the repetitions (I hadn’t read the research at the time), but my sound series had a similar effect in that the final frequency was presented for a different duration, hence the more repetitions.
Generalization
Those of you with a good pitch memory will notice that the beep of the oven control in the DS/CC video is not the same pitch as the beep of the weighing scale. That’s the good news about pitch differences. The initial counter conditioning extended to all beeps. I ad hoc counter conditioned any high-pitched noises that occurred when I was with her. Adding this practice to the formal desensitization and counter conditioning helped Zani recover from her fear.
I continued to treat her with these sounds throughout her life to keep this pairing strong, the risks were too great to tinker with, especially since there are some sounds in life that I have no control over.
Notes
- Zani was in the care of a veterinary behaviorist and she instructed me to wait until his medication stabilized, which I did.
- I got some great advice from a professional trainer.
- Having lived with a sound-phobic dog for a long time, I knew I had to be very cautious – any contact was risky.
Another success was that Zani was happy to step on the same scale for the rest of his life, even though it was beeping, and you can see that at the end of the video embedded above.
My confusion
For the last few years, I have quietly created custom sound series for clients with sound-phobic dogs. Since my work was solely from the acoustic side, I needed the help of a licensed behavioral specialist. After some success, I stopped this service. Nearly every dog that came to me had already generalized the fear, so we always had to deal with new scary sounds. It was like a sad game of whack-a-mole. (The sounds the dogs became sensitive to were not as similar as beeps. An example is a dog who was bothered by the engine sound of a particular delivery truck became sensitive to the sound of backing up and braking.) For these dogs, conditioning one sound did not “go away” the problem with other sounds. By the way, most of the dogs were on medication.
I believe this information should be free and publicly available, which is why I’m writing these two posts. But I worry that more people will scare their dogs by messing around with sounds without the professional help of a behavioral expert (of course, people are already doing that with various “desensitization” sound apps). I’ve toyed with the idea of creating an app or membership program, but the administrative work that would require is beyond me.
My compromise would be to publish a series of “off-the-shelf” sounds (electronic beeps, explosions, dead smoke detector batteries, etc.), and gently manage them by having users sign a disclaimer before access and be convinced that they will follow good practices. If I charge a fee, it will be reasonable.
For now, I’m putting this concept in the public domain. Take care, and I hope you see great success.
Related article
References
Poppen, R. (1970). Counterconditioning of conditioned inhibition in rats. Psychology Report, 27(2), 659-671.
Copyright 2023 Eileen Anderson
