Maybe there's a scheme to be found analogous to procedural, episodic, etc. in memory. Something that integrates behavioural and physiological findings at a medium level of complexity
I definitely see some papers that I think over-define terms, especially e.g. coming up with terms that over-specify the concept of emotion vs mood vs feeling etc. But then often I think a general sense of what emotion is and what it's for is kind of lacking.
What isn't?
We probably need something to show for, like big muscles. Maybe if wearable EEG tech came back in a way that actually works (but I think that may be unlikely), then an optimization-obsessed culture with youtube stars would grow up around it.
From a #neuroscience perspective, or not: What is an emotion?
Though I have to read further of the claims that these slow wave components are task-relevant. Could be underlying mechanism, but I can imagine some ways where you can squeeze a little bit of predictive performance from what is primarily artifacts