Just out at Nature Neuroscience, our multi-lab collaboration, "Advancing functional connectivity research from association to causation”. Unifies functional/effective conn. under goal of inferring causal interactions rdcu.be/bUfIw (a thread) 
Cognition emerges from network interactions, such that causal interactions should be central to studying brain function. Yet the research closest to this – functional connectivity – focuses primarily on association (correlation/coherence) rather than causation.
We suggest that a way forward is to shift the theoretical target of functional connectivity research to causation. This gives us a yard stick for validating and improving functional connectivity as a scientific approach.
Despite major limitations of correlation (the current most popular measure) for making causal inferences, it still provides *some* causal information (see figure). Thus, current methods can be reinterpreted in a causal framework.
And yet we can do much better than correlation (and others like coherence) right now, using measures like partial correlation to reduce confounds, or effective/causal connectivity methods such as fGES or DCM.
Without the goal of making causal inferences, "functional connectivity" is problematic as a scientific concept, since it confuses methodological properties (means of measurement) with theoretical properties of interest.
We present a framework for improving functional connectivity inferences. This has big implications for some of the most popular uses of functional connectivity, such as classification of clinical conditions, resting-state correlations, and dynamic functional connectivity.
We describe a way forward, improving functional connectivity measures via more validation and better method reporting. Causal inference is the common goal and also a common ontology for making cumulative progress across functional connectivity methods.
We provide examples of improving functional connectivity involving fMRI, EEG/MEG, and multi-unit recording. We expect these innovations will lead to much deeper insights into neural mechanisms and the neural bases of cognition. rdcu.be/bUfIw (colelab.org/#publications)
