Interesting! I don’t think this necessarily negates the modular view of the brain though (except in the simplistic “x brain structure does x things” sense). I do think that it suggests that neuroscience may still be pre-paradigmatic as a science. I think there are a couple of things going on here. One principle that’s relevant is that biology conserves energy - therefore physical structures in the brain are always, to the extent possible, multi-use (why waste energy building when you can repurpose). This means the “thing” we’re trying to study is usually at a level of abstraction that means it’s more of a “patterning” than anything physical. The elements of the patterning are sometimes operating at wildly different times scales (400 hz to <1 hz), which in some cases work somewhat like radio frequencies, allowing stacking of signals with minimal interference. Another issue is that, it’s very hard, even in rodents, to isolate a particular brain pattern - the brain is active even in sedation, and that activity is often using the same structures and even neurons as the ones you’re testing. All of this makes instrumentation and measurement very, very hard, as you noted. So the actual analysis often looks more like ecology (unbounded but patterned system, sometimes observed pre and post intervention) than traditional biological research (fairly close to a closed experimental system). I’m working on an essay related to all of this that explores a “Parliamentary Model” of the self and agents. The basic idea is that the brain has shifting coalitions and competitions at multiple levels of scale (neurons v structures v networks v modes), all of which can have impacts up and down the complex adaptive systemic chain, which is why humans are both patterned and nebulous, often teetering on the edge of systemic criticality.