2. what is the point of defining constant term as your eof, it's just a
scalar: mean signal for the whole session, so there is no variance there at
we are talking about the first-level here. Hence, the constant is actually estimated for each time-point and therefore should give me what I like to think of as spontaneous activity of a particular region after controlling for everything else. Correlating the spontaneous activation signal between two regions then gives me a measure of functional connectivity.