Changed the fabric around within the maze, situating where wing 1’s fabric used to now wing 3
For our experiment we wanted to study the role that urine marking plays in cognitive mapping using a radial maze. Our research indicates that rats urine mark for directions as an indication for food (citation). We have also found that rats will also use other rats urine marking track to find their own direction (citation).
In our study we are using three Sprague Dawley rats, Winnifred, Minoo, and Noelle, who all have been previously handled, clicker trained, and trick trained. We have split our experiment into three stages.
Stage One: Serane wrap the entire radial maze. Then cover the whole maze with a cotton fabric using masking tape on the underside.
Place a pebble of food at the end of the designated wing officially labeled, Wing 1
The rats are placed in the center of the radial maze randomly facing different directions. The rats travel around the radial maze habituating to the new environment. A trail was considered from the time the rat started off in the middle of the radial maze to the time the rat ate the food in Wing 1 and returned back to the center at which point more food would be placed in Wing 1 and another trial would start. The location of Wing 1 remained the same throughout stage 1. As the experimenters we were constantly moving around the maze to various positions, moving both in a clockwise and counter clock wise direction, and fulfilled all three various roles of putting the food in, handling the rats, and recording information. Stage one of our experiment lasted for 10 days at which point all three rats met the criteria of going to wing one 75% (+) of the time to eat the food [Video of successful stage 1 trials].
1) Each rat is different in the time they take to complete the trials.
2) The rats have most visibly urinated on the center of the maze rather then on the wings. The most visible concentration of the urine stains are not in front of Wing 1 but rather between Wing 6 and 7.
3) The percentage of times the rats when down the correct wing, Wing 1, compared to other wings they went down is charted on the graph below. Our criteria was that they would go down Wing 1 75% of the time before the rats could move onto stage two. [Graph of Stage one data]
Problems encountered during Stage 1:
1) During habituation the rats constantly would nibble on the tape and/or fabric getting distracted from the task. As a result we had to re-do the fabric covering by pulling the fabric tighter and fixing the tape so the rats could not nibble on it. [Video of the rats eating the tape and fabric]
2) In the beginning after a couple of days, we realized that the rats had no indication to let them know that food was available in wing 1. Therefore we decided to use the clicker method on the radial maze where the clicker had the transferred stimulus control of food. In order to clicker train the rats on the maze we initially started by clicking once when the rat would approach Wing 1 and then again when the rats would eat the food. We slowly started weaning out the second use of the clicker so that we would only click once and the rat would go down Wing 1 and eat the food. [Video of the rats clicker training on the radial maze]
3) One of the rats, Noelle would not respond to the clicker after it was used. She would often just sit at the end of a wing for several minutes at a time not completing the task at hand of getting food which is located down Wing 1. [Video of Noelle being distracted]
Stage Two: For stage two we will now identify and label eight different positions around the radial maze, “A”-“H”. In stage one, Wing 1, where food was always placed, was located at position “A” for every trial. Now we will change the location of Wing 1 to various positions in accordance with a random number generator. We are testing to see if the rats go down position A or do they choose to go down Wing 1 regardless of what position it may be at. Wing 1’s position is changed after every trial. A trial starts when the rat is placed in the middle of the radial maze and ends when the rat eventually goes down Wing 1 and eats the food. At this point the rats are picked up, the maze is turned and another trail begins. The rats are placed on the center of the radial maze facing various wings which is also determined through a random number generator. During this stage we will be testing to see how effective the external visual cues were, that were in place during stage one.
[Video below shows our initial trials during stage 2]
1) During the first trials, two of the three rats did not even go down Position A where Wing 1 used to be during Stage one but rather went down Wing one in a different position. After continuing the trials this same pattern continued of going down the correct position of where Wing 1 was located. The results can be found in the graph below.
[Graph of Stage 2]
Stage 1 video:
For the past two weeks we have been working on designing our experiments and teaching our rat a trick. The trick I chose for my rat is to jump through a hoop. In order to teach the trick I used operant condition and shaping techniques. Operant conditioning involves behavior modification using the principles of reinforcement and punishment. Shaping entails gradually molding behavior by reinforcing closer and closer approximations of the target behavior. The clicker is used as a bridge to help make reinforcement immediate and allow us to reinforce the appropriate behavior. The target behavior for my rat is jumping through a hoop. I have not defined the height of the hoop as am not sure what height she is capable of jumping.
I found that Noelle responded best to the shaping if I did a few clicker-training trials before beginning shaping for the trick (with no hoop in the box). I also allowed Noelle approximately two minutes to habituate to environment of the box and another one minute to habituate to the new stimulus of the hoop.
There were a lot of behaviors with the hoop when we first started. This worked well as it gave me a variety of behaviors to reinforce. To begin the shaping process, I placed the bottom of the hoop on the floor of the box. Every time Noelle came near the hoop (as long as she was not eating it) I would reinforce her with a click and then food. At one point I tried putting a large piece of food on the other side of the hoop to entice her to walk through it and this seemed to help. I also tried to have her follow my finger through the hoop. I gradually increased the criterion for reinforcement until I had her walking through the hoop consistently. Next I began to raise the height of the hoop. It is still very close to the ground.
I am currently working on raising the hoop higher in the air in hopes of shaping an actual jump. Based on the experiences of other people in our class, it seems rats can jump very high.
This week we worked on designing our experiments. The following is the experiment we hope to begin when we get back from fall break.
The participants in this experiment will be three female Sprague-Dawley rats. These rats will be deprived of food for twelve hours prior to training and experimentation. The rats have previously been clicker-trained. This experiment will take place in a psychology lab room at the University of Mary Washington. The lab contains many doors, windows, tables and chairs. For this experiment we will use a radial maze that has been wrapped in saran wrap and covered in fabric (see diagram). The maze will be set __ feet off the ground on top of a wooden stool. Rat urine from the three rats will be collected and stored in eyedroppers. The food pellets used during experimentation and training will average .05 grams.
The independent variable for phase two is wing location. The independent variable for phase three is urine string location. The dependent variable for phase two and three is the number of mistakes made. Mistake is defined as the rat traveling to the end of the wing and its whiskers reaching the crevice (see diagram). A trial is defined as when the rat finds the food.
Phase 1 – A food pellet will be placed at the end of wing one. The rats will be trained individually to respond to the sound of a clicker by traveling to the end of wing one. When the rat is responding like this consistently is being done consistently (90% over two training sessions) we will move on to phase two. The placement of the maze during phase one will remain the same.
Phase 2- We will place five drops of pee on the fabric of wing one before beginning each training session. The maze will be rotated a random number of degrees (determined from a random number generator). One rat will be put in center of the maze. We will record number of mistakes made until the rat finds the food. This process will be repeated and data record for ten to fifteen trials.
Phase 3 – The placement of the maze will be returned to that of phase one. The saran wrap and the fabric on the maze will be replaced by clean fabric and saran wrap. The fabric previously covering of wing one will be transferred to a randomly chosen wing (determined form a random number generator). One rat will be put in center of the maze. We will record number of mistakes made until the rat finds the food. This process will be repeated and data record for ten to fifteen trials. This is to ensure that the rat is not choosing wings based on visual differences in the maze.
* This post was written 10/08/08 and posted 10/16/08