Computations or processes - how do you think about the building blocks of the brain?

Returning to a book chapter by @knutson_brain that I appreciate (and endorse!). One take-home is that there are two different ways of thinking about the brain/mind: one that emphasizes "computations or algorithms" (like learning rules or building up object detectors for seeing) and the other that emphasizes "processes" (like motivation and aversion for studying feeling, emotion, attachment, and mood).
https://link.springer.com/chapter/10.1007/978-3-030-27473-3_7

Building on these ideas, I see how these differences lead to different ideas about the problem that needs to be solved. Both approaches consider levels of explanation, but think about them differently.

The approach that uses computations/algorithms emphasizes these as the bridge between the brain's biology and the brain's function. Documenting that the brain has face detectors does not tell you how to build one; you need to understand the brain's algorithm.

The approach that uses processes typically thinks about levels as levels of scale: genes>molecules>cells>circuits>behavior>self reports (as described by the NIHM RDoC project). But where are the processes? Those aren't levels; insofar as levels are rows, processes are columns (see attached picture). In other words, processes are NOT links between biology and behavior but instead the thing to be explained at different levels. This matrix has always bothered me - it's seems off. Now I see why:

@knutson_brain's proposal is that processes should be regarded as a link between behavior and biology. Absolutely! I think this implies that the RDoC matrix is misguided (or minimally we should not consider it as depicting "levels of explanation") but I'll let him comment 😉​.