To just cover one little gap, antibonding orbitals lie at specific energy levels, just like bonding orbitals. There's a n=1 sigma bonding, n=1 sigma s antibonding, n=2 sigma s bonding, n=2 sigma s antibonding, etc. The n=1 sigma s antibonding falls before the n=2 sigma s bonding, etc. So antibonding orbitals do get filled in nonexcited molecules and can even be filled in the valence level. The relative numbers of bonding and antibonding orbitals is what determines how many bonds will be present in that molecule (bonding orbital electrons - antibonding orbital electrons)/2. So, by exciting electrons from bonding to antibonding orbitals, you can cause bonds to break, like BRT was saying.
I don't remember which of this is MCAT level and not.