setting covariates in glmer (fixed factors vs covariates)

#1
Hi everybody,
i am a complete beginner in R, so please forgive me if I ask a stupid question.
I ran an experiment on the effect of slowed speech (speech.rate; three speech rates) on syntactic structures (OS.SO.sentencetype; with a subject-first and an object-first (OS) structure) to see if slowed speech would facilitate comprehension, especially for the object-first structure (measured in accuracy, binary code). So these variables I defined as fixed factors and I definitely need that interaction as well.
All subjects were administered all 180 items. As random factors I chose subject and itemID.
I tested hearing impaired an non hearing impaired persons in different age groups. It is known that as well as older adults, hearing impaired adults also profit from slower speech rsp are more impaired in OS structures, so i defined two variables: hearing.state and age.group.
In SPSS (and how do I miss these times when SPSS was still fancy…) I would have been able to put hearing.state (binary, 2 values) and age.group (binary, 4 values) as covariates/predictors in generalizing estimating equations. Alternatively, I could run hearing and age as continous variables…
In R, I cannot figure out how to fit in these two variables as covariates. I don’t think that I can treat them with the same power as speech.rate and OS.SO.sentencetype
So far, I have tried several models, defining all of them as fixed factors. I was told that you have to set REML= FALSE if I compare 2 models with manipulation of fixed faxtors
gm1 = glmer(Accuracy ~ OS.SOSentenceType + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

gm1a = glmer(Accuracy ~ OS.SOSentenceType + speech.rate + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

gm1b = glmer(Accuracy ~ OS.SOSentenceType + speech.rate + hearing.state + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

gm1c = glmer(Accuracy ~ OS.SOSentenceType + speech.rate + hearing.state + age.group + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

gm2 = glmer(Accuracy ~ OS.SOSentenceType*speech.rate + hearing.state + age.group + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

gm2a = glmer(Accuracy ~ OS.SOSentenceType*speech.rate*hearing.state + age.group + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

gm2b = glmer(Accuracy ~ OS.SOSentenceType*speech.rate *hearing.state*age.group + (1|Subject) + (1|ItemID), data=dataC, family='binomial', REML = FALSE)

Atm, I feel like fishing in the dark without even a fishing rod.
Help would be very much appreciated!
Cheers, Gela

PS: I attached one part of my data