We have now officially finished our experiment. The last few days of our data collection showed that giving the rats an activity to do between bar presses did actually help them wait. Below are graphs comparing the phases of our experiment, as well as how each rat day each day of the phase.
Phase 1 (no Sophie, Lisa only) – Rat must wait 6 seconds between bar presses and there is no light involved as a cue
Phase 2 (both rats) – Rat must wait 6 seconds between bar presses, light is now involved as a cue
Phase 3 (both rats) – Rat must wait 10 seconds between bar presses, light is involved as a cue and box is present as waiting activity
As you can see from the graphs, both rats had higher percentages in the last phase of our experiment. This confirms our hypothesis that by giving a rat an activity to do while waiting, it will make the DRL schedule better.
Below is a list of references used for our study.
Killeen, P.R. & Fetterman, J.G. (1988). A behavioral theory of timing. Psychological Review, 95, 274-295.
Today was the first day of our test phase of the experiment. The requirements for this phase are the same as when we originally tested the rats on the 6 second DRL schedule. As a reminder, the rats must wait at least 6 seconds before pressing the bar, but this time they have another task they can, but are not required to do. We are curious to see if giving the rats a filler task while they wait will make the timing easier for them.
Today, even though it was only the first day, both rats showed huge improvement on their ability to wait the 6 seconds. After a few minutes with one of the rats, we were able to increase the amount of required waiting time to 10 seconds and still the rats did much better. Although we are going to take a few more days of data, we are optimistic about the results of our study!
Below is a graph comparing the two phases of DRL with Sophie. Sophie only had two phases, where as Lisa had three because Sophie was not being used in the initial first trial (Brooklynn was). By the time Sophie was needed in the experiment we had already figured out the benefits of the light acting as a cue to press the bar. As you can see in the graph below, the box with the light has already shown a much higher percentage of correct responses, compare to just the light. In addition, we also were able to get this high percentage with a longer time interval. Very exciting!
Below is a graph comparing the three phases of DRL with Lisa. In the first phase, Lisa was required to wait 6 seconds before pushing the bar, but there was no light to cue her when the 6 seconds was up. In the second phase she was still required to wait 6 seconds, but the light was added as a cue and there was no filler activity. Finally, the third phase involved both the light as the cue and the box to stand on as a filler activity, but the rat was required to wait 10 seconds before pushing the bar again. As you can see from the graph, even with the increase in the amount of required wait time, Lisa still did much better when given the filler activity.
Today was the last day we double clicked when the rat has waited the 4 seconds on the box. Tomorrow we will begin fading the double click to test to see if giving a rat something to do between bar presses helps with the waiting process. The new criteria go back to the 6 seconds between each bar press, and we will continue to use the light as a cue for when the 6 seconds are up. Below are the graphs of the data for both Sophie and Lisa throughout this phase of our experiment:
After further discussion with our class, we have decided to add a new step in our DRL schedule. We are currently double-clicking when the rat has waited a certain amount of time on the box, and clicking again as soon as they press the bar to receive food. We feel that phasing out that initial double-click will put our rats on a true DRL schedule. If we can successfully do this, then we have taught our rats to go to the box after they eat and wait for the light to be presented, which serves as a cue to press the bar.
Standing on a box while waiting has some theoretical implications about DRL schedules. Specifically, it can help facilitate the amount of time between bar presses. Think about generalizing this to human behavior. If we want to cook something that requires boiling water, that takes a couple minutes to do. So what do we do to help time pass by more quickly? We go watch TV, we read a magazine, we go stand by the window and look outside…we do anything except watch endlessly as the water prepares to boil. This is essentially what we are doing with our rats – we are giving them something to do while waiting for the light to come on. Sure, waiting on the box is still asking the rat to wait, but the time it takes for the rat to eat the morsel, go to the box, and stand there for a brief period is a significant amount. The only thing standing between where we are now and a successful DRL schedule is phasing out the double-click, and we are fast-approaching that final step.
Below are upgraded graphs on the rats process done by date and each second increase step:
Below is a video of what we have been doing during training sessions for the past nine days:
Below are graphs of Lisa and Sophie’s progress since we have added the block to the box. A correct trial means the rat has waited the allotted time on the box then presses the bar to receive food. An incorrect bar means that the rat has pushed the bar without going on the box, while an incorrect box means that the rat did get on the box, but did not wait the allotted amount of time.
Here is a graph of Lisa’s progress by date:
Below is a graph of Lisa’s Progress by step (second progression):
below is Sophie’s data for the first week of Box training wait times:
As discussed in class last week (11/13), we have decided to shift away from putting our rats on a DRL schedule and instead look at why it is difficult for rats to stay on a DRL schedule. We hypothesize that by removing access to a reinforcer (i.e. the bar) the rat should be able to wait in between bar presses and not be “tempted” to press the bar again. We will do this by giving the rat a simple task to do, specifically standing on a box after receiving food.
We have begun shaping our rats to stand on a box after they receive food. Lisa has already been taught to stand on a box from being trick-trained by Maggie. To shape Sophie, we reinforced her when she had one foot on the box, then two feet, then three, and so on. Once Sophie caught up with learning to stand on a box, we shaped them to wait on the box. The rats had to wait one second on the box (all four feet remain on top of the box for one second), then we double-clicked. As soon as the double-click goes off, the light is presented and still serves as a signal for the rat to press the bar. When the rat presses the bar, she receives food. We continued this and increased the waiting time to 2 seconds, 3 seconds, all the way to 6 seconds.
Below is a video of Lisa and Sophie’s progress thus far:
Today in class we discussed possible directions for our project because of limitations on time. Instead of continuing with our original research question of whether or not a friend rat versus a stranger rat affects another rat’s DRL schedule, we have decided to look at why putting a rat on a DRL schedule is so difficult. So far, both rats have ignored timing and continually press the bar. We are curious to see if another task will possibly help the rat wait the allotted amount of time. In order to do this, we are adding a block that the rat must stand on before pressing the bar. Once the rat has pushed the bar, they will go wait on the block instead of standing right next to the bar. If they are able to wait on the box, we can predict that maybe by removing temptation and accessibility, the rat isn’t as concerned with reinforcement on a delayed schedule.
Yesterday, we started the rats on the Differential Reinforcements of Low rates (DRL) training schedule. This means that each rat will have to wait 10 seconds between each time they press the bar, including eating time. In order to do this, we took the average time it took the rats to press the bar, eat the food, and press the bar again, then increased this average slightly. The time chosen as the starting point was 6 seconds for each rat. If the rats do not wait 6 seconds before pressing the bar again, the time starts over (from the incorrect bar press). The rats are only reinforced if they have waited the allotted amount of time and we are still using the light as an aide for the rats. Below is a video showing the procedure used:
Below are graphs from data taken on the first two days of DRL training for each rat:
Today was the final day of Light Training with the rats. For the past week(s) (one week with Sophie, and two weeks with Lisa) we have be spending 40 minutes every day running trials. These trials consisted of reinforcing the rat when they pressed the bar when the light was on (3 minute time interval), and not reinforcing them when they pressed it when the light was off (1 minute time interval). This phase was needed so that the rats would begin to associate the light, as the signal that they were able to receive food. This information will be needed as they move into the next phase of our experiment, which is the DRL schedule.
Below is a bar graph representing the amount of bar presses Sophie did during all trials. The blue bar represents the amount of bar presses when the light was on, while the red represents the amount of presses when the light was off. The amount of bar presses when the light was off has significantly decrease from the first days of training, while the amount of presses with the light on has stayed more or less constant. Ideally we want the amount of bar presses when the light is off to be zero, but because of time we have to move onto Phase II of our experiment (DRL schedule)
The graph below shows the data collected from the last day of light training with Sophie. As has been seen in the other sessions Sophie seems to continue to press the bar even when the light is off in the beginning. But after four trials during the training session she begins to cease pressing the bar when the light is off. A trial is defined as 3 minutes with the light on, followed by 1 minute of the light off.
1. Pressing the bar when the light is on
2. The removal of the light, followed by no bar pressing
3. The re-introduction of the light, followed by bar pressing.
Below is a graph depicting Lisa’s overall progress with the light and bar pressing:
As was the case with Sophie, Lisa also begins the session by still pressing the bar when the light is off. As the session moves on, the amount of pressing while the light is off decreases. Below is a graph showing Lisa’s data from the last light training session:
The next step of our experiment will begin tomorrow and will be training both Sophie and Lisa to press the bar on a DRL schedule, with the light as a prompt.
Currently, we are still working on getting the rats to only push the bar when the light is on. Lisa is making progress with the light training and below is a graph showing how many times she pushes the bar when the light is on and how many times when it is off. As a reminder, the light is on for 3 minutes (after the first bar press) and then off for 1 minute.
As you can see, Lisa is consistently pushing the bar more times when the light is on. It is important to remember that the light is on for 3 minutes and only off for 1 minute. On average, the sessions are around 40 minutes with approximately 7 periods where the light is on then off. The extra time is the time the rats spent between turning the light on and the first time they push the bar during that period.
So far, in this step, the biggest issue we have had to overcome was Brooklynn’s non-response to the light. After she was trained to push the bar, we added the light and Brooklynn would no longer push the bar. After multiple attempts, we decided to start training Sophie to push the bar and use Jenny as the “stranger” rat. Sophie is now pushing the bar and has started training with the light. Below is a graph of her progress with the light using the same 3 minute on, 1 minute off schedule Lisa is on:
As you can see from the graph Sophie is associated that when the light is on, and she pressed the bar, she will receive food. I would say that about 90-95% of the time that the light is on, Sophie is sitting by the bar and consistently pressing after she consumes each piece of food. Occasionally she will press the bar with the light on, but this is decreasing as well. Now when I turn the light off she walks away from the bar, sometimes after waiting a little she will try to press the bar again, but when no food comes she gives up. Bar pressing with alternating the light on and off will continue for the next few days. Hopefully by Monday we are planning to get both Lisa and Sophie started on their DRL schedules (which involves waiting a specific amount of time between each bar press to receive food). Until then we are hoping to see the amount of presses with the light on increase, while having the number of bar pressing without the light decrease, until they are no longer pressing the bar at all when the light is off.