In functional connectivity is a statistical dependency like correlation or similar, it looks at the effect (or behavior), not at the mechanism (or cause). In this case a "ground truth" would only be such for a single mechanism, and not necessarily what FC should be after

FC methods don't necessarily recover all or any theoretical properties, yes. But evaluating a method against data with known theoretical properties will inform us about how to interpret the computed statistical dependencies.

I was hoping he meant neurophysiological measures of synaptic connectivity, but maybe he meant fMRI?