Hello,
I am designing an experiment to test changes (from baseline) to a measured physiological parameter in response to two different treatments (A & 
used on their own or combined. Resources are limited so I am attempting to perform all treatment combinations in each subject (animal), in randomized order.
The runs would look something like this:
| Subject | Time0 (baseline) | Time1 | Time2 | Time3 | Time4 |
| 1 | null | A | null | A+B | B |
| 2 | null | A+B | B | A | null |
| 3 | null | null | A | B | A+B |
| 4 | null | B | A+B | null | A |
| ... | ... | ... | ... | ... | ... |
Carry-over effects are a large concern and the way that I am hoping to combat this issue is to wait for effects to wear off and return to time0 (baseline) values prior to starting the next randomized treatment instance in time. I am sure this is not a bulletproof method, but I have to work within my means.
I am curious if others have opinions on how I would go about analyzing this type of data (MANOVA, GLMM, etc.) or if there is a better way of arranging my runs to ultimately quantify significance.
Thank you
