T2 hotteling for univariate time series

[borjaevo]

Hello,

I have to do a comparison between multiple subjects for a drug response variable over time. The variable to analyze is movement and is highly autocorrelated. Movement is measured one time per second.
Each subject should be compared to a control.
I read in some paper T2 hotteling could be used to do the comparison and is not affected by correlation but i don't understand why and how i have to perform the test. I don't know if i have to use time as a second variable in some way....
I really appreciate any help.

 

[Miner]

I have heard of T2 Hotteling, but have never used it. However, given your study you may want to consider a control chart for autocorrelated data. This would help you identify if a change occurs and when it occurs.

 

[borjaevo]

I have heard of T2 Hotteling, but have never used it. However, given your study you may want to consider a control chart for autocorrelated data. This would help you identify if a change occurs and when it occurs.

Thanks for the info. But i also need to statiscally conclude if a drug has an effect on the subject or not compared to the control. This is why i need something like a t-test for time series and hotteling T2 i think could be used for that. What i dont know is how i have to use the time in the formula.