>>12320501imo it really should be called something else... comp neuro nowadays encompasses a bunch of different things that don't really have anything to do with treating the brain as a computer. Sure there's lots of work on making spiking NNs with biologically inspired constraints that act like part of the brain, but it's a lot more than that, and imo, the hardcore functionalist old school style comp neuro is out of style.
Lots of the work that falls under this label has to do with building new data-analytic tools that help us make sense of high dimensional neural data, inferring latent low-dimensional neural dynamics that tell us something about how the brain does X, and other things that obviously need to be done that aren't what you'd probably think of when you hear "computational neuroscience."
>is computational neuroscience a memeI mean there's lots of shitty work being done because it's such a new science, but I'm convinced that we can't use qualitative biology-style methods to understand how networks of neurons accomplish the things that they do. For that, we need to quantitative models that are able to describe measurable patterns of neural activity while linking these patterns to whatever it is we're trying to describe.
There's no reason that general brain function should be open to qualitative inspection in the way that, say, simple visual cortex cells are (or were thought to be immediately after Hubbel/Wiesel at least–I think there's debate about this now)
