How do I run a post hoc test for a mixed effects model in Rstudio I am new to R and pretty average with SPSS so I meed a little advice. 
I have 34 participants and I am looking at how muscle contractile properties vary amoung 12 different test sites along the back. 
So each person was tested at the 12 sites. 6 different vertebral levels (L1, L2, L3, L4, L5, MT) on the left side of the spine and 6 on the right side.
I want to see 1) if contractile properties vary between vertebral levels, but i want to seperate the left and right sides so I know exactly where the difference occurs.
And 2) whether there are any differences between sides but same vertebral level, for example L1 on Left vs L1 on right. 3) if there are any Gender differences between contractile properties at each of the 12 sites.
I was told to do a lmer but not sure exactly how and what the best post-hoc test is. I also imported my data table from SPSS. And dont know if I have to/how to convert it to R text becasue when I try to run the test it has an error. 
I have SPSS too if that is easier to explain. I am worried I am selecting the wrong settings in SPSS.
Also for a mixed effects model what is the most appropriate way of presnting the stats? What values would you need to provide in a journal article and what setting do you need to mention in the methods?
Any help would be greatly appreciated!
 A: Something like 
lmer(contract ~ gender*side*vert +
  (1|subject) + (1|subject:vert) + (1|subject:side), data = your_data)

Since this is a split-plot design (every side/vert combination is measured for every subject), you can in principle estimate (side*vert|subject) instead of all of the random effects listed above, but it will in practice be way too complex to fit with 34 subjects. The version here estimates variance among subjects, among subject-vert combinations, and among subject-side combinations (since as I understand it there is one measurement per subject-side-vert combination, this variance is equivalent to the residual variance which is automatically included in the model).
You haven't said much about how your data are formatted, but it should be in "long format", i.e. look something like this:
 contract gender side vert subject
 1.2      M      L    L1    1
 1.7      M      R    L1    1
 1.2      M      L    L2    1
 ...

(best to convert your subject variable to a factor explicitly).
There are several packages that can do post-hoc tests but emmeans may be best/easiest (multcomp is another choice).
