sorry but you mean you want to know if a mouse is trapped ? what you mean by home-cage activity? so i can help you

Elie, thanks for your response. When I say "home-cage activity" I simply mean the 24hr surveillance of a mouse's activity (movement) within their home-cage, not in a testing apparatus. The "activity" that I would like to measure is movement vs. no movement, particularly during the dark cycle hours. My goal is to monitor the duration of physical activity within the mouse's homecage during all hours of the day using the rpi.

Ok so it’s not for zapping them ;)

For multiple cages a camera solution is not practical (especially at night) I recommend PIR detectors (Passive Infrared), there is some complete modules you can buy (for like 10$) on eBay or AdaFruit

https://learn.adafruit.com/pir-passive-infrared-proximity-motion-sensor

These modules goes high when there is motion (and low when there aren’t any ;) ) so managing them on RPI GPIO is a piece of cake (and this cake is not a lie!)

NB1: RPI does not have a hardware clock, it gets the time from an NTP Server (so you need to be online) and Unplug your Raspberry Pi completely, pull out the network cable and start the Raspberry Pi up again. You will see that the date and time are incorrect. The solution is simple and cheap (like 4$) is by adding an RTC (Real Time Clock) module with a battery, these modules have I2C line

https://thepihut.com/blogs/raspberry-pi-tutorials/17209332-adding-a-real-time-clock-to-your-raspberry-pi

NB2: if your cages aren’t near each other’s and you don’t want wires everywhere (be careful long wires = loss of signal and bad communications) I might be tempted by wireless solution. An idea is using cheap RF 433 MHz modules (many for few $) with HT12E encoders. On receiving you can decode using HT12D (same address) this will give you 16 channels (4bit). However if more than one emits it won’t be pretty as none will be received!!! So think of limiting emission time to notification (like 3 burst elapsed by 1 sec … you’ll think of something) … or use multiple emitter-receiver-decoder (different address) … etc.

Good luck and say hello to Mikey ;)

Very helpful, thank you! From what I've read I won't be able to run more than a maximum of 2 camera modules on a single rpi if i'm using USB input, which is why I've been leaning toward the PIR detectors as you noted. How many PIR modules do you think the RPI3 can handle simultaneously? I'd like to get 4-6 cages (i.e. individual detectors) per rpi3 using a GPIO, to minimize cost., but I'm afraid this may be a bit too much for the little rpi s to handle. Will obviously figure this out with trial and error but I'd prefer not to spend the money if its a lost cause. 

It's a simple io read . Cost nothing so you can use as many io you have (17 I guess ) you read the state periodically.  Even so you can go for much more by multiplexing or using i2c expansion. ... 

The camera solution ( ir camera with illumination) is heavy as you must use image processing . Comparing to read an io is like planting a nail with a jackhammer 😂

Awesome. Thank you again for your help, Elie. 

Jacob

Jacob, It looks like you wanted to do the same thing I now want to do. Could you provide any information as to what you ended up choosing? I'm hoping to monitor (x,y) or (x,y,z) activity in up to 16 cages.

Thanks,

 Matthew

Matthew, 

I eventually determined that the cost was too much for my current interests. The way I approached the situation granted ~3 PIR motion sensors per Rpi. This was based on what the GPIO had to offer. However, this approach did not have the level of sensitivity as I would have liked. Also, given the number of animals we test at the same time, 3 mice/Rpi is not practical. I then moved to infrared beams, in which movement would be tracked by beam breaks (estimated 3 per cage). But, I have not been able to secure these out of the reach of my mice (they'll chew them up). Also, this is only possible with 1 Rpi per cage, which is far too expensive for the implications of the data. 

If you make any advancements on a cheaper or more efficient method, please let me know. 

Best of luck,

Jacob, 

 Thanks for the quick reply.  This is what I'm currently looking into: http://www.colinst.com/products/animal-activity-meter-multi-varimex.  I've been quoted a cost of about $20k for 32 IR curtains, which I think will give me X, Y, Z positions for 16 cages.  Not cheap. :(

Thanks,

 Matthew

Oh wow, that is expensive. The IR curtains are likely far more accurate though. The Rpi IRs wont give you all three axes. The PIRs will measure active vs. not-active, but wont work with software to measure distance and movement pathway in the homecage, which I believe most of these commercial products will model for you. One thing you could consider doing (which is what I have considered) is only running a subset of animals from each group, i.e. n=4 or 5. Depending on how you stagger your experiment replications, you could cut the total cost by quite a lot.

A wild guess guys. Can you put a marker on the mice (A colour dot or micro or led) and then do imagery processing using cam monitoring the cage. (Blob detection and tracking) (it may work without marking but lot easier with )