Trick #1

September 12, 2009

Over the past two days, Duchess has been perfecting her first trick – running through a tunnel when she hears the clicker.  I began training her yesterday, and today she proved that she knows her stuff!  Below are two videos:  

After establishing the association between the sound and the food during clicker training, I needed to spend another few days maintaining this behavior to ensure that it stuck.  Next, I needed to shape her association using operant conditioning.  I did this through quite a bit of trial and error, and eventually figured out the best way was to have the tunnel leading into the corner, since she would run there when she heard the sound.  I made it impossible for her to get to her food any other way, and she eventually got used to running through the tunnel.  

After some trials with the tunnel in the corner, I moved the tunnel into the middle of the box.  Soon enough, she would run through it with the prompt of my hand leading her way.  I phased this out after a few trials, and she was running through the tunnel on her own quickly enough!  

 

Day 1 - tunnel in the corner

Day 1 - tunnel in the corner

Day 2 - tunnel in the middle of the box

Day 2 - tunnel in the middle of the box

You’ll notice I used different boxes on each day – while she was clicker trained in the cardboard box, on our first day of “trick” training, she tried to jump out of it.  At this point, I felt that she was habituated to her surroundings and she did just fine working in the bigger, wooden box in the lab instead.

Advertisements

Below is a video that illustrates Duchess’ response to the clicker a bit better than the video in the previous post.  Today, I was also successful in phasing out the double click, and shaping her to respond after just one click.  It was great to see this hard work pay off! 

After thinking back, throughout this whole process of clicker training, I wanted to make a list of a few points that I think are important: 

-Water breaks (this improved her responding remarkably – and something I unfortunately didn’t even think of until a few days into training) 

-Reinforcing Duchess with the right-sized pieces of food (starting her off with crumbs made her expect nothing greater) 

-Establishing the perfect deprivation/feeding schedule (for her, it seems that she’ll work hard after about 18 hours with no food).

Today Duchess perfected her training, as illustrated in the video below:  

Goal for the rest of the week:  To phase out the double-click, and get Duchess to respond soley to the single click method.  

Today, Duchess was deprived for 17.5 hours (and her weight was back up to 181.1 grams) and worked really well through about 50 trials, eager to eat each time.  

Another video:  

The goal for the past few days:  to perfect Duchess’ clicker training and eliminate any associations she might have been making with my hand when I place the food in the box.  

This was successful for the most part, although we encountered some issues yesterday during training.  I think I had deprived Duchess of food for too long (it had been close to 22 hours – I’m a bad rat mother) and she got a little sick when we got about 15 minutes into training. If I deprive her any less than that,  she just isn’t motivated enough!  We ended our session early yesterday and I left her with 9 grams of food.  Her weight on Sunday was 183.7 grams, after 18 hours of deprivation, and her weight dropped on Monday to 177.1 grams after 21.5 hours of deprivation.

Clicker Training, Days 3 & 4

September 5, 2009

Yesterday, Duchess weighed in at 189.2 grams after about 9 hours of food deprivation.  She worked for a little while, but was clearly not motivated enough;  I decided it was necessary to deprive her of food for even longer today.  Before we began working today, Duchess weighed 183.1 grams, which made me nervous because it was down quite a bit from the day before.  

After acclimating herself to the box, I did a few trials, realizing that my hand was definitely becoming part of the association for her – at this point, the clicker noise just fades into the background.  I made a hole to administer the food through (which Duchess also gets a delight out of sticking her head through and looking around for said food), which worked a little better, although the food falls in and rolls out of the corner.  Problems, problems.  

On the food hunt

I also noticed that giving her water throughout the session has made her performance improve significantly. We did about 50+ trials today, with various sizes of food bits.  I was interested to see how she would respond to crumbs vs. bigger pieces (which are limiting in the sense that only so many trials can be done), but she certainly likes holding the bigger pieces of food in her paws and eating them slowly.

Clicker Training, Day 2

September 4, 2009

I broke up clicker training today in two parts, after realizing that Duchess had not been food-deprived for long enough the first time around.  If today were a timeline, it would look like this:

10:30AM – removed Duchess’ food 

4:00PM – clicker training, round one (Duchess weighed in at 193.9 grams and worked through 21 trials) 

8:00PM – clicker training again (Duchess weighed 190.4 grams this time, and worked through about 20 trials, took a break, then worked through only 3 more)

A few issues that make me concerned:

1.   She is not responding to the food after the first two or three trials. I am astonished that after about 9.5 hours without any food in her cage, she was still not very motivated to work as well as I had hoped.  

2. I worry that my hand is becoming part of her association. To prevent this, I will cut a hole in the corner of the box where I normally put her food, in hopes that I can just stick it through there instead.

In order to train Duchess, I am using the “double-click” method to reinforce the positive association between the clicker and the presentation of the food.  I placed the food in the bottom right corner of the box while Duchess was on the other side, unaware of my hand entering the box.  She weighed 188.1 grams and received 10 grams of food before her deprivation period.  

Goal for the next few days: to perfect her responding, then eventually fade out the double click.  

Day 1:

An important act to note is her hoarding behavior, which we discussed in class briefly last week. I was consistent in placing the food in the bottom right corner of the box while she was turned the other way, and after she found the food, she would always run up to the top left corner before consuming it.  Interestingly, after she was full and unmotivated to work for food any longer, she would hoard the food in the corner instead of eating it.  Our last trial consisted of her curling up comfortably on her pile of collected grub.  

Full and tired