Totally agree and the point is well taken. On a bayes account, although a negative result doesn’t prove ‘unnecessary’ it does lower the posterior p that C is necessary (depends on prior, n, etc. etc.)

Not easy (circuit out of a brain can’t behave). But wouldn’t up/do
wn modulation of a specific circuit (e.g. w/ opto) tightly controlling a behaviour (e.g. www.science.org/doi/epdf/10....) show ‘production sufficiency’ (to borrow from the TINS paper)?

Wouldn't demonstrating joint necessity/sufficiency for a set of circuit satisfy "causal production" as in TINS, below? Perhaps I misunderstood, but this seems like it fits that definition of causality (production) in distributed systems. Happy to be corrected if I’m missing anything!

The necessity/sufficiency distinction seems important here. A lesion in C with no effect tells you C is not necessary for b, but nothing about sufficiency (although brain stim in C could clarify). Likewise, C2 might be unnecessary when lesioned alone, but the set {C, C2} could be jointly necessary.

In my view there will be many such circuits, but if you don’t know how to look for them it will seem like everything is everywhere. But this view is not very fashionable at the moment. It will return. And then it will be unfashionable again :)

Prob a small effect, but the FAQ mentions they excluded multi-disciplinary journals from the sample. Some of these journals (esp. PNAS, Nat comms, etc.) tend to publish imaging > EEG afaik, it may partially reflect a shift in where imaging gets published…

Oh man. Science Neural Circuits would be my new favorite journal.