John Lighton, Ph.D.
Creating behavior transition probability matrices with data extracted from Promethion metabolic and behavioral phenotyping systems.
VIDEO NOT LOADING? Viewing videos requires acceptance of all cookies. To change consent settings, open the MANAGE CONSENT tab at the bottom right of this window and click ACCEPT.
[Music]
In the third and last video, we’re going to show you how to create behavior transition probability matrices and Markov chain visualizations from your behavior lists.
Just to refresh your memory, here’s a behavior list. And as you can see, each behavior will transition to another behavior, and what we’re interested in is the probability of that transition. How do we visualize that? Here is one way and that’s using a behavior transition probability matrix. Each row sums to 100% and each column corresponds to a particular behavior. So as an example, after eating food, it’s about 42% likely the animal is going to drink water, and it will not go on the running wheel. After drinking water, it also will not get on the running wheel, about a quarter of the time it will get into its habitat, and about half the time it will just lounge around the cage. After lounging around the cage for between 5 and 60 seconds, what’s the most likely behavior? It’s running on the running wheel. So, we gain some amazing insights into the mind of the mouse using this array.
Now how do we actually visualize this data? One way of doing that is with a behavior transition probability matrix visualized by Markov chain analysis. Each of these nodes corresponds to a particular behavior and each of these behaviors has a certain percentage probability of giving rise to the next behavior. So, here’s a blow-up, and this gives you some uniquely detailed insight into the mind of the mouse. Which would allow you for example to distinguish in an objective numerical fashion between control and experimental groups, between wild-type and knockout, and what have you. Again, these are totally objective numbers that you can take to the bank and use with confidence. Thanks for listening!
[Music]
