[Thanks to Bryan Krause's advice, I edited the question.]
Definition: Let two functional units be two sets of neurons, A and B, such that the probability of being connected (synaptically or by neurohormons) for A-neurons among each other is very high, the same for B-B-connections, but for A-B-connections, the probability is very low.
Assume there are two functional units whose occupied brain regions (i.e. whose voxels) strongly overlap.
By fMRI (with its given spatio-temporal resolution) those two units could not be distinguished and it would be hard to observe their functional connectivity or non-connectivity directly (because fMRI scans would not show, if only one or both of them are "active"). Nevertheless some experimental set-ups might reveal their functional connectivity or non-connectivity indirectly.
My question is: Has this ever been observed? Is it known, whether there are two functional units that occupy essentially the same brain region and have been shown indirectly to be functionally connected or not (as defined here)? Or is it very unlikely that this will occur.