Mouse Conditioning Apparatus that requires some simple EE work.

Thread Starter

kave325

Joined Oct 5, 2006
4
I know this a lot to read, but it is a fairly simple ME/EE project that I could definitely use even the least bit of guidance in.

I am trying to build a self-automated apparatus that can condition mice to “learn a certain image.”

There are two compartments in a box. In one compartment, the mice rest and there is a door that can allow only one mouse at a time to enter the other compartment, where the conditioning will occur.

In the conditioning compartment, there are three computer screens that display 3 different image patterns. One of these is the “positive” image that we will try to teach the mouse. This is done like so:

All three screens have a food dispenser below them. However all screens also have a “shock” platform. Of the 3 screens at any given time, one will be a positive image in which the mouse will be dispensed food. However, if the mouse chooses to get food from the other 2 food dispensers beneath the “negative” images, then punishment will occur in the form of a minor shock.

Unfortunately, the position of the “positive” image cannot always be in one position and most appear random to the mouse so that the mouse doesn’t always, let’s say, go left because it has learned there is no shock there. That means the 2 places that shock occurs must correlate to the position of the “negative” image and no current must be flowing to the “positive” image shock mechanism.

So my question to you guys is how can I build this thing so that upon a stimulus (which will be the opening/closing of the door between the compartments) the positions of the shock change so that only 2 of the 3 are on at a given time and how can I make this alternate? They need to be in some order, maybe 1 1 0, 1 0 1, and 0 1 1—that’s fine with me, as long as the positive image is not in the same position all the time. Is there a simple way to make a circuit board sequence through these positions and provide current in this pattern?

ANY help is appreciated, sorry for the long post guys. Feel free to ask questions.
 

beenthere

Joined Apr 20, 2004
15,819
Hi,

It would be easier to use a microcontroller to do all that stuff. It's more flexible when it comes to making fune-tuning changes to suit the experimental needs.

That said, you will still need a certain amount of custom hardware built and programming skills to do the apparatus.

Got any resource on campus?
 

Thread Starter

kave325

Joined Oct 5, 2006
4
I've talked to an EE professor about it for his input. He was unable to give me anything detailed of course, but he mentioned using a 3-bit counter. Even this may be a bit complicated.

Alternate idea:

Is there anyway I can take prexisting hardware from a toy that already sends signals in sequences? Perhaps a toy that alternates between LED colors which is caused by a stimulus that i can attach to the door in my apparatus?

Thanks again. ANY suggestions welcome again.
 

beenthere

Joined Apr 20, 2004
15,819
Hi,

You might be able to use something like that as a basis, but you will need enough power to change the state of the door. For proper messing with the rodents' minds, a microcontroller is probably best (plus supporting hardware).
 

Chris Wright

Joined Jul 26, 2006
62
...... So my question to you guys is how can I build this thing so that upon a stimulus (which will be the opening/closing of the door between the compartments) the positions of the shock change so that only 2 of the 3 are on at a given time and how can I make this alternate? They need to be in some order, maybe 1 1 0, 1 0 1, and 0 1 1—that’s fine with me, as long as the positive image is not in the same position all the time. Is there a simple way to make a circuit board sequence through these positions and provide current in this pattern?
You say that there are three computer screens for the images? How are you going to synchronize the images with the shock grids? What is driving these screens?

I would think a single computer with three video boards might work. Get a cheap or broken keyboard and use one of the key contacts as input for your door. This key stroke starts a basic program that runs a randomizer to select one of the screens to display the "good" image, the other two get the other images and signals are output to close relays to energize the shock grids under the "bad" images.
Here is a circuit for controlling relays from a parallel printer port:

http://ourworld.compuserve.com/homepages/Bill_Bowden/page6.htm#PPRIC

I know, parallel ports are old, but this won't take much computing power and there must be an old computer in a closet some where.
 

Thread Starter

kave325

Joined Oct 5, 2006
4
Hey Chris,

I was actually just going to try and split up one large screen into three different sections with 3 images. I think that would be easier than having 3 monitors. I would probably just use a template or something to cover the monitor and make it look like 3 separate images are present.

Any input on how i can control the shock grids to be 'randomized', yet still follow some coherent pattern that i can sequence the position of the images in so that the position of the positive image and "no shock" grid correlate? They don't have to directly communicate with each other, but the positive image position and the no shock just have to match up whenever the sequence is changed.

Thanks again guys.
 

Chris Wright

Joined Jul 26, 2006
62
Hey Chris, .......................Any input on how i can control the shock grids to be 'randomized', yet still follow some coherent pattern that i can sequence the position of the images in so that the position of the positive image and "no shock" grid correlate? They don't have to directly communicate with each other, but the positive image position and the no shock just have to match up whenever the sequence is changed.
If you had a "pattern" that allowed you to synchronize the grids and images, how do you know that the mice are not learning this pattern and instead of the images? Wouldn't this invalidate the data?

But, as I said above, you could control the whole thing from the one computer. No need to have a separate circuit to control the shock grids. The link above shows a simple circuit using a transistor and a relay* that can be controlled by the parallel port of the same computer that is displaying the images. A parallel port has 8 data lines, you only need three. You build three of the circuits and hook them up to three of the data lines to control the three shock grids. You write a program that randomly selects a "screen" to display the good image on and then the program sends a data bit to the parallel port relays of the shock grids of the other two screens to turn them on.

The only question is do you have a computer with a parallel port. Or perhaps a USB to parallel port converter box might work.

*(It actually shows three circuits. One using a solid state relay, that you probably won't use. The other two have regular relays but differ in whether they are controlled by a data "1" or "0". If you were displaying the "good" image in the center screen, you would send a binary 101 or a 010 depending on which circuit you used.)
 

Thread Starter

kave325

Joined Oct 5, 2006
4
Hey Chris,

I have a computer with a parallel port so that shouldn't be a problem. I really like this idea--it sounds the least complicated (still somewhat complex though) of the suggestions I've been hearing.

About the mice being able to "learn" the sequence of the "random" patterns...there will be multiple mice in the resting compartment and one will be allowed to enter the conditioning compartment at a time due to the nature of the door that I will build (only fits one mouse to go through). The exiting of the mouse from the conditioning area back to the resting area will trigger the change in shock and image position. Since there are multiple mice in the resting area, odds are that a different mouse will enter the conditioning area next (since the other has already satisfied his appetite) and thus be exposed to a different/random pattern. In other words, the same mouse won't be exposed to a definite pattern in image position.

Some questions though, if you or anyone else could answers though:

1. Let's say i get a program written that can go through a sequence of 101, 110, and 011, with 0 being no shock, how can i get this program to correlate then to the image positions? will the program have to be able to display these images as well or communicate to another program on the computer to change it's positions?

2. How can i get a stimulus sent from the closing of the door between the compartments to signal the computer to change the image positions and shock grid on/off?

Appreciate your input.
 

Søren

Joined Sep 2, 2006
472
Hi,

any more ideas anyone?
Since your screen will probably have an unused area (above or below the images), you could make each picture (part) taller and fill the remaining part with eg. black on the "good" picture and white on the rest.
Then you just mount 3 Light Dependant Resistors (LDR), plus a simple transistor circuit, to see the black/white areas and trigger each plates zapper circuit. If the LDR sees a light screen, it turns the zapper on. Then you just randomize the pictures you send to the screen and the zappers will behave in sync.
The part with either black or white should be shielded from the mouses sight of course and a metal screen might be the best way to keepn the wires from the rodents teeth.
 
Top