Experimental Summary

December 1, 2008 - Leave a Response


Much research has been conducted on rats and mazes concerning scale learning, cognitive mapping and orientation (Morris, 1981; Morris, Garrud, Rawlins & O’Keefe, 1982; O’Keefe & Nadel, 1978; Olton 1979; Packard & McGraugh, 1996).  Our study concerns the orientation learning aspect of navigation in rats.  Work has been done to determine that rats utilize spatial memory, and there has even been physical evidence discovered that certain proteins aid in the consolidation of this type of memory (supported by the ability to suppress it) (McGauran et al, 2008).  Recent research shows that rats seem to use predictive cues to aid in spatial memory (Olton, 1979). 

While previous studies have shown that rats use cues to navigate mazes, our study seeks to eliminate external cues (extramaze and intramaze visual, olfactory, and other possible cues).  The remaining signals available to the rats to aid navigation are proprioceptive, or the internal signalling systems which help the rat determine their physicallocation (for example, signals from the nervous system in the muscles).  When an animal moves (locomotion) there is positive and negative feedback in the brain which leads to the musculoskeletal dynamics (use of proprioceptive feedback) (Smarandache, Daur, Hedrich & Stein, 2008).  Smarandache et. al, looked into muscle tendons in the nervous system of crabs.  They tested the effects of positive versus negative feedback on motor patterns through injection in order to excite or inhibit the muscle. Their results yield a change in the feedback from positive to negative or vice versa when the muscle tendon was excited or inhibited.  Based on these findings, we hypothesized that this proprioceptive feedback would signal to the rat that it is walking on an incline or a flat surface.  If the incline of that surface is changed, the rat’s proprioceptive feedback should signal the change to the rat’s brain and therefore, lead to a difference in behavior.


This study was intended to determine if proprioceptive memory in rats could be disrupted by changing the incline of a multiple T-maze used for a task.  The task was for the rats to run through the maze to receive a reward of food at the end.  During all 4 phases the maze was the same (although the incline changed), and an odor neutralizing spray was used after every trial to eliminate odor from urine trails.  The study was run in a small room with a closed door, helping to eliminate air pressure changes and reducing outside noises.  The inside of the maze was uniform to prevent intramaze visual cues.  The error rate of each run was recorded and averaged per day (an error being a time when a rat turns down a hall that does not lead to the finish box and places at least one paw into that hall) to demonstrate their rate of learning.  If the rats’ navigation was interrupted by the shift in proprioceptive feedback due to the switching of the maze incline, there would have been a change in their error rate between the phases following:

In Phase 1 we trained two rats on a flat maze and two on an maze raised at the highest point of 5.715, with a 10°angle incline (See figure 1). Each rat set had one rat familiar with other types of training and one rat naive to all training.  We marked all errors that the rat made within trials.  An error consisted of at least one paw in a turn that did not lead towards the finish.  All rats ran the maze on a daily basis (approximately 10 runs a day) until they made fewer than 3 errors per 10 runs (i.e. the maze had been learned once they reached this goal).

In Phase 2 we switched the incline of the maze: The rats on the flat maze now ran the inclined maze, while those trained on the inclined maze in Phase 1 switched to the flat maze.  No significant difference was seen in their performance between Phase 1 and Phase 2.

In Phase 3 all of the rats maintained training on their incline from Phase 2 while habituating to the Extramaze Visual Cue Reducer (VCR).  The VCR is a poster board covering the maze with slits cut throughout it.  The slits are just wide enough for the experimenter to observe the rat yet small enough that it would be difficult for the rats to see through while performing the task.  This helps to reduce the extramaze visual cues which may influence rat navigation.  The habituation results were not relevant to the study, but this was an important step as we did not want any confounds in Phase 4 due to the new apparatus.  At this stage a new rat was introduced to the study.  This rat was naive to all training.  She was trained only on the flat maze and was also introduced to the VCR.

In Phase 4 all 5 rats were switched to the maze raised at 15.24cm with a 15° angle incline.  Again, no significant difference was seen in the performance of the rats.


Because there was no change in performance for any of the rats between phases, we determined that switching the incline of the maze does not interrupt maze navigation in rats.  This suggests that there is another process being used by rats besides external cues.  Perhaps proprioceptive feedback is more complex than imagined, and therefore less likely to be disrupted by an incline change.  For example, there could be separate proprioceptive processes for determining surface incline and for determining a horizontal (right to left, left to right turn) change.  Also the memory used in recalling the “correct” direction of the maze (which we assume is spatial memory) may be at work.  Our results suggest that if this map is what’s being used during correct maze running, that this map is adaptable to the incline.  This is because the maze never changed, so the spatial memory map of the maze should not change either, which would explain the low error rates in sequential phases of our experiment with varying incline.


The following charts are created from the error rates of each rat in each phase of the experiment.  The different colored lines denote the different phases (1, 2 & 4).  The error rates for Phase 1 decrease relatively consistently for all rats, but the consecutive phases plateau, showing no significant difference in performance after incline changes.

rat 1 data

rat ran inclined maze in phase 1

flat maze in phase 1

rat ran flat maze in phase 1

To show Mrs. Brisby's Data from Phase 1,2 and 4

To show Mrs. Brisby

To show Dodger's Data from Phase 1,2 and 4

To show Dodger

naive rat introduced in phase 3

naive rat introduced in phase 3

 Figure 1:

To show how we measured our angle of incline.

To show how we measured our angle of incline.

Works Cited

McGauran, A. T., Moore, J. B., Madsen, D., Barry, D., O’Dea, S., & Commins, S. (2008).  A Possible Role for Protein Synthesis, Extracellular Signal- Regulated Kinase, and Brain-Derived NeurotrophicFactor in Long-Term Spatial Memory Retention in the Water Maze. Behavioral Neuroscience, 122-4, 805-815.

Morris, R. G. M. (1981). Spatial localization does not require the presence of local cue. Learning and Motivation, 12, 239–260.

Morris, R. G. M., Garrud, P., Rawlins, J. N. P., & O’Keefe, J. (1982). Place navigation impaired in rats with hippocampal lesions. Nature, 297, 681–  683.

O’Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map. Oxford, England: Clarendon Press.

Olton, D. S. (1979).  Mazes, maps, and memory, American Psychologist, 34(7); 583-596.

Packard, M. G., & McGaugh, J. L. (1996). Inactivation of hippocampus or caudate nucleus with lidocaine differentially affects expression of place and response   learning. Neurobiology of Learning and Memory, 65, 65–72.

Smarandache, C., Daur, N., Hedrich, U. & Stein W. (2008). Regulation of Motor Pattern Frequency by Reversals in Proprioceptive Feedback. European Journal of Neuroscience, 28 (3), 460.


Phase 4

November 24, 2008 - Leave a Response

We have begun the final phase of our experiment (phase 4).  Our experiment consists of the flat maze, the first incline and the final incline.  The first slope of the incline we used for each rat was .05715 with a highest point of 5.715cm.  Phase 4 consists of inclining the maze even further.  The slope of Phase 4 = .1524, with the highest point of 15.24cm.  We choose these heights due to the fact that they are significantly different in height.

Below is a picture to show an updated version of our apparatus and the incline that we are using for Phase 4:

Phase 4 Apparatus

Here are the graphs for Mrs. Brisby and Dodger comapring their data from Phase 1, 2 and 4:

Remember that in Phase 1, Mrs. Brisby was on the flat maze and Dodger was on the  slope of .05715 with a highest point of 5.715cm.  Then, in Phase 2, Mrs. Brisby was on the slope of .05715 with a highest point of 5.715cm, while Dodger was on the flat maze.  Finally, in Phase 4 both rats were placed on the slope of .1524, with the highest point of 15.24cm.

Mrs. Brisby and Dodger (1,2 &4)

Here is a quick video of Phase 4 with Mrs. Brisby and Dodger:

Here are charts comparing data from phase 1, 2 and 4 for Dinner and Eleanore, and a combined phase 1 and 3 (phase 3 being the habituation phase) with phase 4 for Jenny:

dinner data

nora datajenny data

Phase 3 Update

November 20, 2008 - Leave a Response

Phase 3 is complete!  All the rats are having an error rate less than 3 errors per 10 trials.  We were only concerned with the introduction of Jenny which is why we have extended our phase 3 until now.  However, her error rate is currently consistant with our guidelines.  We will begin Phase 4 (which is the incline of 55* for all rats) tomorrow.

Here is an updated graph for the error rates of Mrs. Brisby and Dodger:

Mrs. Brisby’s and Dodger’s Phase 3 Data

Phase 3

November 18, 2008 - Leave a Response

After we completed Phase 2 on November 14, we began Phase 3.  Basically, Phase 3 is habituation to the External Stimuli Reducer Maze Top (i.e. it’s phase 2 with poster board on top).  So, each rat is staying in their incline for phase 2 (Mrs. Brisby is 38* inclined, Dodger is flat, Eleanore is 38* inclined, and Dinner is flat).  Jenny was also introduced into the habituation process.

Here is a quick video showing some of the problems, progress and overall process of habituation to the maze using Mrs. Brisby and Dodger:

Here is a video of Jenny and Dinner with the poster board:

Here is a chart of Jenny, Dinner, and Eleanore’s error rates with the visual cue reducer:

Phase 3

Here is the graph for the end of Phase 2 compared to Phase 1 for both Mrs. Brisby and Dodger:


Here is the graph for Phase 3 (Habituation to the top of the maze) for both Mrs. Brisby and Dodger:


Continuing Phase 2

November 13, 2008 - Leave a Response

To further describe the progress of phase 2, here are lines showing data from both phase 1 and 2 for Eleanore and Dinner laid on the same chart to show the differences:

phase charts

Here is an updated version of Mrs. Brisby’s and Dodger’s Progress of Phase 2 compared to Phase 1:


Also, here is a video summarizing our progress so far, using Dinner and Eleanore as models:

Here is a quick video update on phase 2 for Mrs. Brisby and Dodger:

Beginning Phase 2

November 11, 2008 - Leave a Response

Phase 2 has begun.  In this stage we switched the degree of incline of the maze for each rat.  For example, Dinner was running the maze on an incline during phase 1 and now is running the maze while it is flat on the floor.  The following data represents no change in error rate for Dinner and Eleanore between phase 1 and 2:

Dinner (average number of errors):

Phase 2 Day 1: 0.266667

Phase 2 Day 2: 0.6

Phase 2 Day 3: 0

Eleanore (average number of errors):

Phase 2 Day 1: 0.066667

Phase 2 Day 2: 0.1

Phase 2 Day 3: 0.1

Error rates are still low, indicating no disruption of spatial learning base on changing the incline of the maze.

Based on this lack of results we have decided to extend our experiment with a third phase.  In this phase all rats will be switched to an even larger incline of the maze.  An apparatus has been designed to minimize the amount of visual cues used by the rats, and again we will be using an odor neutralizer spray between trials.  A fifth rat (Jenny) will also be run, first on the flat maze while the other rats are continuing phase 2, and then on the increased incline.  The purpose of phase 3 is to allow a more dramatic incline while avoiding the confounds of visual cues for which we had originally chosen a lower incline to the maze.  The apparatus is shown below:

Visual Cue ReducerThere are 1/4 in. slits cut through the posterboard in order to allow us to observe the rats while the posterboard itself helps block potential extramaze visual cues.

When an animal moves (locomotion) there is positive and negative feedback in the brain which leads to the musculoskeletal dynamics (use of proprioceptive feedback) (Smarandache, Daur, Hedrich & Stein, 2008).  Smarandache et. al, looked into muscle tendons in the nervous system of crabs.  They tested the effects of positive versus negative feedback on motor patterns through injection in order to excite or inhibit the muscle. Their results yield a change in the feedback from positive to negative or vice versa when the muscle tendon was excited or inhibited.

In our study ,we are trying to tap into the muscle feedback of an incline of a maze and how that effects the proprioceptive feedback to the brain.  In other words, will the incline (or decline) of the maze for phase 2 effect the feedback to the brain and produce more errors or will the spatial memory of learning the maze in phase one be utilized and a decrease in errors occur ?

Phase 2 for Mrs. Brisby and Dodger has been a little different.  Below, are two graphs showing the average errors rates per trial for phase 1 (blue lines) compared to phase 2 (red dotted lines):

Error Rates of Phase 1 Compared to Phase 2

According to the graph, Mrs. Brisby’s phase 2 error rate looks like it’s fluctuating a lot (as it did in phase 1).  However, her overall error rate for phase 2 is: .32, where as in phase 1 for her fifth trial her overall average was: 4.29.  So, Mrs. Brisby’s data is consistant with Dinner and Eleanore.  However, Dodger’s phase 2 data seems to show an increase in error rate as we move along trials.  Although her overall phase 2 average for the 5 trial days is: 3.995, in comparison to the first 5 days of phase 1: 19.16 it still seems like it’s cutting it close.  In light of Dodger’s difference between rats, is another reason why we have decided to introduce Jenny into our experiment.

Here is a current video of Mrs. Brisby’s and Dodger’s Progress:

Phase 1 Complete

November 5, 2008 - Leave a Response

Mrs. Brisby/Dodger:

Phase 1 of our Spatial/Motor Memory Maze Experiment is finally completed.  Although, most of our rats consistently progressed through phase one; it was not all smooth sailing.

The progress of Mrs. Brisby has stabilized to a consistent average of 1 error per 10 trials.  Mrs. Brisby has remained stable since October 27, 2008.

Dodger on the other hand, showed a consistent progression within trials until she reached the finish box.  When she reached the box, she would turn around and move towards the start multiple times.  I decided to move the food just before the finish box during trials of October 30th.  Dodger’s error rate significantly decreased from 15.26 to 4.80.  I then began shaping Dodger in phases of 2 centimeters (back towards the finish corner).  I also spent an evening trying to habituate Dodger to the finish box before moving the food all the way into the box (last phase of shaping).  Currently, Dodger has 3.67 errors on average per 12 trials which we believe to be successful enough to move to phase 2.

Here are charts of time and wrong turns for Mrs. Brisby and Dodger:

Mrs. Brisby’s and Dodger’s Data of Phase 1

Here is a chart comparing wrong turns of Mrs. Brisby to Dodger:

Mrs. Brisby’s and Dodger’s Wrong Turns Compared

Here is a video of Mrs. Brisby’s current state as well as Dodger’s shaping and current state (Note: The volume must be turned up high! — my mic was acting up again!) :

Here is a chart of error rates for Dinner in Phase 1:

Dinner Phase 1 Data

And the  chart of error rates for Eleanore in Phase 1 as well:

Eleanore Phase 1 Data

Video and information of the beginning of Phase 2 will be posted shortly!

Experiment progress

October 30, 2008 - Leave a Response

While running the experiment we have come across some problems:

Dinner and Eleanore:

Initially both Dinner and Eleanore, the two rats I have been running through the maze, spent most of the trial times going back and forth through the maze but not entering the “finish” box.  In order to help motivate the rats (as I figure this may have been the reason they were not looking for the food) I lowered their diet to 6-8 g. of food per day.  Since then, both rats have spent much less time in the maze and quickly run to the finish box.  The reduction of food also may be why several other behaviors stopped or at least were minimized.  Both rats spent a significant amount of time chewing on the matchsticks which are glued to the bottom of the maze and also poking the maze cover with their nose.  Either they habituated to the maze, reducing these behaviors, or they now have a stronger drive to find the food.

Following is a video taken today, after the food reduction and habituation, where both Dinner and Eleanore move quickly through the maze:


Mrs. Brisby and Dodger:

In the beginning, both Mrs. Brisby and Dodger had minimal trials in the maze.  This was because if I tried to run more trials, the trials would go over 15minutes.  I began making the pellets at the end of the maze much smaller as well as providing water.  Afterwards, an increase in the number of trials for both rats occured. 

Also, in the beginning of phase 1, Mrs. Brisby was running the maze slower than Dodger with significantly lower errors.  Mrs. Brisby had a tendency to groom and sniff the box/air during much of the trials.  I began not feeding her immediately after trials. I also decreased the interval inbetween trials.  As you can see in the video, Mrs. Brisby currently makes on average .3 errors and also runs through the maze in about 12.09 seconds in 10 trials.

Dodger, on the other hand has had significant “problems” in phase 1.  Many of her trials in the beginning were very slow (over 15 minutes) and included errors exceeding 45.  Dodger also had the tendency to recieve the food at the end of the maze and IMMEDIATELY run towards the start of the maze.  To help increase the likelihood of decreasing her errors, I decided to do many different things. Firstly, I decreased her overall daily intake from 20.0 grams to 16.5grams.  I also decreased her pellet size (even smaller than Mrs. Brisbys).  Finally, after she recieved the food at the end of the maze, I would open the top and pick her up before she could run back.  This took many trials to master, but she is defintely progressing.  Currently, Dodger’s average time is 73.81 seconds with 9.27 errors over an 11 trial period.

Here is a video showing some of the problems and the progression of Maze Training in Phase 1 of our experiment (for Mrs. Brisby and Dodger): Please turn the volume very high – Sorry, I had issues with my microphone!

Spatial Memory and the Altered Incline of a Maze

October 22, 2008 - Leave a Response

This study is based on research demonstrating that muscle skills are learned (Cohen & Nicolelis, 2004; Kargo & Nitz, 2003; Kargo & Nitz, 2004; Kleim et al., 2004;  Wolpaw & Chen, 2006). Spatial memory (as described by Olton, 1979) involves these learned muscle skills, as certain physical action are required for completion of the maze.  For example, if a person is blindfolded and asked to walk to the corner of the street, it is likely they will succeed based on physical familiarity with the feel of walking to the corner.  However, what would happen if the street were changed from a flat walk to a walk up a hill?

Research has shown that learning tasks in rats results in alteration of brain formation after 5 days of training (Kargo & Nitz, 2003; Kargo & Nitz, 2004).  Based on this research we are questioning whether rats, having learned a multiple T-maze (shown below) over 5 days and incorporated the physical moves required to complete the maze (turns, methods of walking, etc.) have a disrupted learning pattern once the maze is placed on an incline.

Multiple T-Maze

Multiple T-Maze

For this experiment we are running four rats through the maze.  After each run (from start to finish) we are spraying the maze with an odor neutralizer to minimize/eliminate urine trails left by the rats.  In stage 1 two rats (one familiar with unrelated learning tasks and one naive to these tasks) will learn the maze as it is laid flat while two (also one familiar and one naive) will learn the maze on an 35 degree incline.  The purpose of the 35 degree incline is that it is low enough that extramaze cues (primarily visual cues) would not change significantly.  A rate of errors made (wrong turns) will be taken.  in stage 2, after 5 days of training, the incline of the maze will be changed to the alternate incline.  The error rate will be taken at this point as well, and the rats will be run for a number of trials equal to the number of trials given during the stage 1.  A comparison will be made between the error rates of stage 1 and stage 2.  If the spatial memory of the rats is disrupted due to the incline of the maze there will be an error rate in stage 2 similar to that of stage 1, but if there is no disruption the error rate of stage 2 will be significantly lower than in stage 1 if not absent.

Following is a video describing the initial stages of the study:

Here is another video of some of the problems we encounter during trials as well as an example of a successful trial at the end:

Works Cited

Cohen, D. & M. A. L. Nicolelis.  Reduction of single-neuron firing uncertainty by cortical ensembles during motor skill learning, The Journal of neuroscience, April 7, 2004. 24(14):3574-3582.

Kargo, W. J. & D. A. Nitz. Improvements in the signal-to-noise ratio of motor cortex cells distinguish early versus late phases of motor skill learning, The Journal of Neuroscience: June 16, 2004, 24(24): 5560-5569.

Kargo, W. J. & D. A. Nitz. Early skill learning is expressed through selection and tuning of cortically represented muscle synergies, The Journal of Neuroscience: December 3, 2003, 23(35): 11255-11269.

Kleim, J. A., T. M. Hogg, P. M. VanderBerg, N. R. Cooper, R. Bruneau, & M. Remple.  Cortical synaptogenesis and motor map reorganization occur during late, but not early, phase of motor skill learning, The Journal of Neuroscience: January 21, 2004. 24(3):628-633.

D. S. Olton.  Mazes, maps, and memory, American Psychology. July 1979, 34(7):583-596.

J. R. Wolpaw & X. Y. Chen. The cerebellum in maintenance of a motor skill: A hierarchy of brain and spinal cord plasticity underlies H-reflex conditioning, Learning & Memory, 2006.  13: 208-215.

Experiment Begins!

October 22, 2008 - Leave a Response

So after I built the multiple T-maze over the weekend (see pictures below!) – we decided to begin our experiment on Tuesday. After discussing what consists of a wrong turn (at least one paw in the wrong direction), how many rats our experiment will consist of (4 rats, 2 with flat version 1st, 2 with incline version first) and how large of an angle we will use (35°  angle) – we began our trials.

 Mrs. Brisby had 6th successful trials with the flat maze.  I taped all of her trials and will put up a video of it soon.  For the most part each trial had a degression in time of at least 20 seconds as well as a degression in wrong turns of at least 2.  Her 5th and 6th trials her time went up as well as wrong turns, but she was also fiddling with the top of our maze (which consists of a window screen) — once she stopped pushing on the screen, she immediately ran to the food.  This made me believe, if on task, she could do the maze faster than her last two trials show.  Her slowest time was : 3:56.63 and her fastest time: 51.92.  Her most number of errors: 21, her least number of errors: 2.

As for my other rat, Eleanor her trials were a bit different.  We only completed three trials.  This was because I believe she was not hungry (due to the fact that she had some food left over in her cage before “working in the maze” and her showing no interest in the food at the end).  Her times were significantly slower than Mrs. Brisby’s as well as her wrong turns significantly larger.  She had a tendency of going half way through the maze, turning around and running back to the start.  Once on task, she completed the maze faster than her times show.  Her slowest time was 7:50.95 and her fastest time was 2:33.85.  Her most number of errors: 43, her least number of errors: 15.

Although our overall goal is to complete the maze without errors, it is interesting to compare the rate of learning through time as well.