Repeated Measures with a baseline Covariate

I am in the final stages (i.e., manuscript is due on July 21) of my dissertation and I am stuck on what seemed to be a minor stats issue. This issue has come up before but not been completely answered as of yet. SO,

I have a 3 way interaction (2 engaged or not, 2 ostracized or not, 2 Time of measurement of handgrip strength) in a repeated measures design with a baseline continuous covariate (the baseline is the initial handgrip strength measure before any manipulation is introduced).

I have scoured books and the internet trying to find a way to decompose these effects. While I am trying to find a way to do it in SPSS I have NO problem doing it by hand at this point. Someone suggested that you might use the mean square error as a means of comparing by hand but was not sure if that still worked when a covariate and a repeated measure is in place.

Any help at all in this would be great- I am trying the bits and pieces of suggestions taken from other posts but if anyone has anything more specific...